Part 1 (of 3): What are major aspects of evolutionary theory?
an...@sci.sci
summarizes the major aspects of (modern mainstream) (biological)
evolutionary theory? For example, the following should be listed:
- There are traits which are passed down the generations in chemical form.
- These traits are coded in DNA (except in RNA viruses).
- Mistakes are sometimes made in copying the DNA.
- These copying-mistakes (mutations) are essentially random.
- Sometimes there are alternate versions of the same gene, i.e. allelles.
- Sometimes one allelle results in better success than another, causing
the proportion of one to increase at the expense at the other.
- In the absense of such a difference, neutral drift occurs at random.
- As a result of random mutations combined with either selection or
drift, inherited traits in a population change their allelle
proportions.
- When we see a geographic area in which one apparent species appears
in the fossil record, and then later in approximately the same
geographic area we see a different but very similar apparent species,
in most cases the latter descended from the former. (43 words, sorry.)
- This descent with modification we see in the fossil record is mostly
the result of random mutations combined with selection and/or drift.
- When we see these descent-with-modification paths that converge in
the distant path, thus seeming to diverge from a common ancestor when
looking forwardly in time, that's indeed what happened, a common
ancestor followed by one or more splits of one species into more than
one species. (Oops, that's 46 words, can anybody make it shorter?)
- Some parts of eukaryotic cells derive from freeliving prokaryotes
which have been trapped in the cells and co-evolving for a long time.
- Often two different species, or two sexes of a single species,
compete with each other over a long period of time, whereby each
evolves an attack against the other whereupon the other evolves a
defense to that attack whereupon the former evolves a better attack and
the latter a better defense etc. back and forth nearly endlessly.
(Ouch! That's 57 words. I really need help making it more concise.)
- Speciation splits combined with non-splitting lines of descent, form
large "trees" (clades) whereby a large number of species descend from a
single ancestral species. The total number of such maximal clades is
very small (less than twenty, perhaps only one) compared to the total
number of species (several tens of millions). (Ouch again: 51 words!)
That's probably only about half the major aspects of modern
evolutionary theory. Where is there a complete list in a Web-accessible
site? And under 25 words each if possible??)
(Part 2 will be to ask anyone who disagrees with any of those aspects
of evolutionary theory to state precisely which of them he/she is
disputing.)
(Part 3 will be to ask those disagreeing with any particular aspect of
evolutionary theory to state what precise alternative they propose.)
I was originally going to start with part 2, just ask
anti-evolutionists to state what aspect they disagree with, but I've
seen such a huge shitload of strawman, gross misunderstandings of
evolution, already, that I figured if I did that I'd get another batch
of strawmen, and never anything serious, so I decided I'd better list
the aspects of mainstream evolutionary theory first, and *after* that
is agreed upon *then* the anti-evolutionists can state which of those
already-specified aspects they dispute.
.
chris.li...@gmail.com
an...@sci.sci wrote:
> Is there a Web site which concisely (like 25 words or less each)
> summarizes the major aspects of (modern mainstream) (biological)
> evolutionary theory?
Some things cannot be described in 25 words or less. That is what lots
of people want, but sometimes it just ain't gonna happen (tell me how a
TV works in 25 words or less- and I know almost nothing about
electricity).
But you might start with the FAQ at
Chris
r norman
<chris.li...@gmail.com> wrote:
>
>an...@sci.sci wrote:
>> Is there a Web site which concisely (like 25 words or less each)
>> summarizes the major aspects of (modern mainstream) (biological)
>> evolutionary theory?
>
>Some things cannot be described in 25 words or less. That is what lots
>of people want, but sometimes it just ain't gonna happen (tell me how a
>TV works in 25 words or less- and I know almost nothing about
>electricity).
>
>But you might start with the FAQ at
>
>www.talkorigins.org
>
Short cartoon versions of "evolution, the highlights" are bound to be
incorrect or misleading in many details and hence not useful for your
purposes. The simple fact is that evolution is a fairly complex
subject with innumerable details. Any good introductory biology book
(the giant ones intended for biology "majors") tend to have three to
six or more chapters on evolution. And the real McCoy requires even
more intensive and detailed coverage. Perhaps the best introduction
is Futuyma's shorter introductory text, "Evolution" by Sinauer
Associates, 2005
http://www.sinauer.com/detail.php?id=1872
shane
> an...@sci.sci wrote:
>
>>Is there a Web site which concisely (like 25 words or less each)
>>summarizes the major aspects of (modern mainstream) (biological)
>>evolutionary theory?
>
>
> Some things cannot be described in 25 words or less. That is what lots
> of people want, but sometimes it just ain't gonna happen (tell me how a
> TV works in 25 words or less- and I know almost nothing about
> electricity).
>
> But you might start with the FAQ at
>
> www.talkorigins.org
>
> Chris
>
Careful, doesn't that mean that creationism is more likely correct, by
occams razor? "Goddidit" is the one word explanation of everything.
Shane
flon...@longship.net
this does not mean at the cost of the rest of the the science. It is
within scientific methodolgy's frame of reference which Occam's razor
must apply, not some contrived straw man frame of reference. So one
must be careful about naive applications of Occam's razor. I do not
think that William of Okham wanted his razor to be used to slice things
so thin that all's that left is "God did it."
"Do not multiply entities unnecessarily"
William of Okham.
Steven J.
"shane" <remarcs...@ozemail.com.au> wrote in message
news:Onvhf.65$kQ2....@nnrp1.ozemail.com.au...
least.
Note that Occam's razor does not refer to the length of an explanation, but
to the number of entities or causes employed in that explanation for whose
existence one can't provide independent evidence (i.e. that simply have to
be assumed to make the explanation work).
>
> Shane
>
-- Steven J.
Dale
news:1132883775.4...@f14g2000cwb.googlegroups.com...
> an...@sci.sci wrote:
> > Is there a Web site which concisely (like 25 words or less each)
> > summarizes the major aspects of (modern mainstream) (biological)
> > evolutionary theory?
>
> Some things cannot be described in 25 words or less. That is what lots
> of people want, but sometimes it just ain't gonna happen (tell me how a
> TV works in 25 words or less- and I know almost nothing about
> electricity).
>
> But you might start with the FAQ at
>
> www.talkorigins.org
You know, this brings up an important point. The talkorigins.org website is
in severe need of reorganization and redesign.
To a beginner it isn't at all obvious where to start. People here might
think it's obvious, start at the FAQ! But do joe...@web.tv or
jan...@aol.com know that? Maybe people figure since it's based on a Usenet
newsgroup it doesn't have to be user friendly, but in my opinion it would be
much more useful if it were, and it wouldn't be that hard to do.
Also, the front page format is so late 90s! What's up with the white on
black, why does a huge graphic take over the entire upper half of the page,
and why is the front page formatted into a single 600 pixel wide column?
Of course, it's a good effort for a site that nobody gets paid to create,
but I never feel comfortable giving someone a link to the main page. I start
them off with a link to something more friendly, like Five Major
Misconceptions About Evolution or 29+ Evidences for Macroevolution, though
the latter is a bit daunting still.
Actually, I know there are better sites out there for beginners, but I never
can remember their urls!
Mark Isaak
<NotMyRealEmail@_comcast.net> wrote:
>On 24 Nov 2005 17:56:15 -0800, "chris.li...@gmail.com"
><chris.li...@gmail.com> wrote:
>
>>
>>an...@sci.sci wrote:
>>> Is there a Web site which concisely (like 25 words or less each)
>>> summarizes the major aspects of (modern mainstream) (biological)
>>> evolutionary theory?
>>
>>Some things cannot be described in 25 words or less. That is what lots
>>of people want, but sometimes it just ain't gonna happen (tell me how a
>>TV works in 25 words or less- and I know almost nothing about
>>electricity).
>>
>>But you might start with the FAQ at
>>
>>www.talkorigins.org
>>
>
>Short cartoon versions of "evolution, the highlights" are bound to be
>incorrect or misleading in many details and hence not useful for your
>purposes. The simple fact is that evolution is a fairly complex
>subject with innumerable details.
I disagree. Evolution is a fairly simple concept whose consequences
are extremely complex.
An analogy is that gravity in our solar system can be described
completely as F = G*m1*m2/r^2. Some consequences of this simple
formula are Kepler's laws, which are a little more complex but still
simple. Further consequences can be seen in Saturn's rings, e.g.
racing moons; they are complex enough that I won't describe them here,
but still simple enough that they can be described. Further
consequences are the chaos you get into whenever you have three or
more bodies.
Similarly, evolution can be described (although not so completely) as
imperfect replication plus selection. From there, you can build up
more an more consequences, getting more and more complex in the
process. I think a lot of creationism stems from people looking at
all the complexity and giving up, never considering that simple
principles can lead to complex results. Teaching the simple
principles, and starting to teach some of the not-quite-so-simple
results might counter some of this.
> Any good introductory biology book
>(the giant ones intended for biology "majors") tend to have three to
>six or more chapters on evolution. And the real McCoy requires even
>more intensive and detailed coverage. Perhaps the best introduction
>is Futuyma's shorter introductory text, "Evolution" by Sinauer
>Associates, 2005
> http://www.sinauer.com/detail.php?id=1872
>
--
Mark Isaak eciton (at) earthlink (dot) net
"Voice or no voice, the people can always be brought to the bidding of
the leaders. That is easy. All you have to do is tell them they are
being attacked, and denounce the pacifists for lack of patriotism and
exposing the country to danger." -- Hermann Goering
Mark Isaak
>Occam's razor says that the simplest theory generally wins.
No, it does not. As you quote below, it says the theory with the
fewest unnecessary entities generally wins. That does not mean
simplest, since one unnecessary entity can take the place of a
near-infinite number of necessary details.
>"Do not multiply entities unnecessarily"
>William of Okham.
--
r norman
The evidence clearly indicates that "imperfect replication plus
selection" convinces nobody. One critically important factor you
overlook is that it is the genome sequence that is imperfectly
replicated but that selection acts on the individual (ordinarily;
there is a lot of discussion even on this point). The characteristics
of the individual (phenotype) differ enormously from those of the
genotype and the connection is very circuitous. Furthermore, there
are a lot of details to be learned about just what "imperfect"
replication means. For instance, it does NOT at all mean changing a T
to a C. Chromosomal mutations, recombination involved with sexual
reproduction, .... There are a lot of cell biology and molecular
biology factors that really have to be known.
Simple "imperfect replication plus selection" produces the incredulity
exhibited as people look at the truly complex organization of a modern
cell or organism: the eye, the blood clotting system, even the
bacterial flagellum are thrown out as features demanding to be
explained. The complexity you describe must be part of the original
package.
Then there is the problem of evidence that people constantly complain
is "missing". The fact is that evidence for evolution covers the
entire gamut of biology: genome sequencing, molecular biology, cell
biology, biochemistry, physiology, development, ecology, biogeography,
the fossil record.
Then there is the problem of differentiating adaptive changes within a
population from the origin of the major taxonomic groups. There is
an enormous amount of biology here and a failure to understand these
issues produces a refusal to accept evolution.
You are right that a massive amount of education is needed. But that
means teaching biology, not encapsulating evolution to a few pithy
phrases and expecting people to be bowled over in shock and awe.
matts2
shane wrote:
> chris.li...@gmail.com wrote:
>
> > an...@sci.sci wrote:
> >
> >>Is there a Web site which concisely (like 25 words or less each)
> >>summarizes the major aspects of (modern mainstream) (biological)
> >>evolutionary theory?
> >
> >
> > Some things cannot be described in 25 words or less. That is what lots
> > of people want, but sometimes it just ain't gonna happen (tell me how a
> > TV works in 25 words or less- and I know almost nothing about
> > electricity).
> > But you might start with the FAQ at
> Careful, doesn't that mean that creationism is more likely correct, by
> occams razor? "Goddidit" is the one word explanation of everything.
But it is a different explanation each time. This is a subtle but
important point. Let us take two situations: A) I drop a rock and it
falls down and B) I drop a second rock and it falls down. I can posit a
single explanatory framework to explain both A and B: roughly speaking
momentum and gravity. Or I can explain A by saying God did A and B by
saying God did B. Notice that I have replaced the equivical "it" with a
specific A or B. If I add rocks C-Z I can still use the same framework
or I can posit new actions by God causing those specific rocks to act.
Since "God did X" does not contain any law-like predictive statement I
can't apply it to any other observation. Not only does the Razor not
lead us to God did X, Ockham used it specifically to separate
philosophy (and later science) from theology.
Mark Isaak
<NotMyRealEmail@_comcast.net> wrote:
I still think the creative power of evolution can be taught without
introducing the full complexity of biology. For example:
Suppose, after you graduate from high school, your rich uncle offers
you a job designing cars for his automobile company. Your specific
task is to design cars that can better withstand head-on collisions.
Suppose further that you have learned nothing in high school that
might apply to designing cars. (Remember, this is a hypothetical
situation. Don't try this at home!) Your uncle hires you anyway
because he is your uncle, but can you get anything done?
Fortunately, your uncle supplies you with unlimited time and
resources. So here's what you do: Make several copies of the
blueprints, make some slight modifications to each set, and have cars
built according the different plans. Put those cars through crash
tests, and keep the blueprints only of the car that does the best in
those tests. Now make more copies, more random changes, and more
tests, and again keep only the best plans. Keep up this practice for
a long time. Gradually, you will build a car that performs very well
in crash tests, and you never have to learn the first thing about
cars.
That is basically how evolution works. Yes, there are differences
between manufacturing cars and breeding pigeons, but the principle of
trial and error is analagous. You can and will get improvements with
no knowledge whatsoever about how the organism works. You can
probably see already some problems that might arise, though. For
example, if you are selecting only for surviving head-on collisions,
you are likely to make a car that performs poorly in other areas, like
mileage and handling. [Segue into discussion of evolutionary
trade-offs, competition, etc.]
A valid criticism of this lesson is that it does not illustrate the
origin of a new "feature," but it is off the top of my head, and I
think something leading to a new feature could be introduced with the
proper choice of item and selection criterion.
As a side effect, you can tell the fundamentalist parents that you are
teaching intelligent design :-).
r norman
<eci...@earthlinkNOSPAM.next> wrote:
The "valid criticism" you cite will be the major feature thrown back
in your face. That is the crux of the complaints -- the seeming
absence of explanation for "new features", for "new kinds", for
macroevolution. Arguing microevolution is already a done deal to most
people.
Noelie S. Alito
> On Sat, 26 Nov 2005 04:44:07 GMT, Mark Isaak wrote:
>>On Fri, 25 Nov 2005 18:03:10 -0500, r norman wrote:
>>>On Fri, 25 Nov 2005 22:04:57 GMT, Mark Isaak
<snip trial-and-error car design evo analogy>
>>
>>A valid criticism of this lesson is that it does not illustrate the
>>origin of a new "feature," but it is off the top of my head, and I
>>think something leading to a new feature could be introduced with the
>>proper choice of item and selection criterion.
>>
>>As a side effect, you can tell the fundamentalist parents that you are
>>teaching intelligent design :-).
>
> The "valid criticism" you cite will be the major feature thrown back
> in your face. That is the crux of the complaints -- the seeming
> absence of explanation for "new features", for "new kinds", for
> macroevolution. Arguing microevolution is already a done deal to
> most people.
In that case, don't forget to ask for a "macroevolution" difference
between chimps and humans.
Noelie
--
When I was a child, I acted as a child...and pretty much stayed that way.
Von R. Smith
Mark Isaak wrote:
> On 24 Nov 2005 20:47:18 -0800, flon...@longship.net wrote:
>
> >Occam's razor says that the simplest theory generally wins.
>
> No, it does not. As you quote below, it says the theory with the
> fewest unnecessary entities generally wins. That does not mean
> simplest, since one unnecessary entity can take the place of a
> near-infinite number of necessary details.
Indeed, one virtus dormitiva can obviate all sorts of chemical
reactions.
John Wilkins
next to me fell asleep.
--
John S. Wilkins, Postdoctoral Research Fellow, Biohumanities Project
University of Queensland - Blog: evolvethought.blogspot.com
Nihil tam absurdum quod non quidam Philosophi dixerit - Cicero
TomS
<HtCdnYaNPeYeihTe...@giganews.com>, Noelie S. Alito stated..."
I'd like to hear about any feature of "irreducible complexity"
or "complex specified information" or whatever which distinguishes
the human body from that of most *mammals*.
Among the standard examples, the most restrictive seems to be
the blood-clotting system, which is general among extant mammals.
This would seem to tell us that the purposes of the intelligent
designer(s) did not include any ends for us which were different
from what they intended for most mammals (let alone monkeys and
apes).
--
---Tom S. <http://talkreason.org/articles/chickegg.cfm>
"It is not too much to say that every indication of Design in the Kosmos is so
much evidence against the Omnipotence of the Designer. ... The evidences ... of
Natural Theology distinctly imply that the author of the Kosmos worked under
limitations..." John Stuart Mill, "Theism", Part II
Robert J. Kolker
>
> I went to a lecture on virtus dormativa, but the hypochondriac in the seat
> next to me fell asleep.
It is a good thing that he and you were not attending the lecture on
flatulensa egregiosa.
Bob Kolker
>
Nic
an...@sci.sci wrote:
> Is there a Web site which concisely (like 25 words or less each)
> summarizes the major aspects of (modern mainstream) (biological)
> evolutionary theory? For example, the following should be listed:
> - There are traits which are passed down the generations in chemical form.
> - These traits are coded in DNA (except in RNA viruses).
> - Mistakes are sometimes made in copying the DNA.
> - These copying-mistakes (mutations) are essentially random.
> - Sometimes there are alternate versions of the same gene, i.e. allelles.
> - Sometimes one allelle results in better success than another, causing
> the proportion of one to increase at the expense at the other.
> - In the absense of such a difference, neutral drift occurs at random.
Are there not two different forms of drift? 'Content' drift due to
random mutation, and allele population drift (amongst mutually equally
beneficial alleles) in the absence of mutation - eventually leading to
fixity of one of them.
> - As a result of random mutations combined with either selection or
> drift, inherited traits in a population change their allelle
> proportions.
> - When we see a geographic area in which one apparent species appears
> in the fossil record, and then later in approximately the same
> geographic area we see a different but very similar apparent species,
> in most cases the latter descended from the former. (43 words, sorry.)
> - This descent with modification we see in the fossil record is mostly
> the result of random mutations combined with selection and/or drift.
> - When we see these descent-with-modification paths that converge in
> the distant path, thus seeming to diverge from a common ancestor when
> looking forwardly in time, that's indeed what happened, a common
> ancestor followed by one or more splits of one species into more than
> one species. (Oops, that's 46 words, can anybody make it shorter?)
> - Some parts of eukaryotic cells derive from freeliving prokaryotes
> which have been trapped in the cells and co-evolving for a long time.
Has that to do with evolution per se? How about a general point on
symbiosis, plus the point that it can end in complete assimilation?
> - Often two different species, or two sexes of a single species,
> compete with each other over a long period of time, whereby each
> evolves an attack against the other whereupon the other evolves a
> defense to that attack whereupon the former evolves a better attack and
> the latter a better defense etc. back and forth nearly endlessly.
> (Ouch! That's 57 words. I really need help making it more concise.)
They don't even need to be different sexes. Competition creates
hurdles for no reason. Trees don't *need* to be that tall.
> - Speciation splits combined with non-splitting lines of descent, form
> large "trees" (clades) whereby a large number of species descend from a
> single ancestral species. The total number of such maximal clades is
> very small (less than twenty, perhaps only one) compared to the total
> number of species (several tens of millions). (Ouch again: 51 words!)
I don't understand. Do you mean there could be between two and twenty
clades with no common ancestor? Or between one and twenty clades which
diverged by some other means than the norm? Or that there are between
one and twenty morphological character states that are qualitatively
different from all the rest?
> That's probably only about half the major aspects of modern
> evolutionary theory. Where is there a complete list in a Web-accessible
> site? And under 25 words each if possible??)
I notice you didn't say anything about vestigial organs. I'm not
saying you should have done, but they are controversial.
> (Part 2 will be to ask anyone who disagrees with any of those aspects
> of evolutionary theory to state precisely which of them he/she is
> disputing.)
>
> (Part 3 will be to ask those disagreeing with any particular aspect of
> evolutionary theory to state what precise alternative they propose.)
>
> I was originally going to start with part 2, just ask
> anti-evolutionists to state what aspect they disagree with, but I've
> seen such a huge shitload of strawman, gross misunderstandings of
> evolution, already, that I figured if I did that I'd get another batch
> of strawmen, and never anything serious, so I decided I'd better list
> the aspects of mainstream evolutionary theory first, and *after* that
> is agreed upon *then* the anti-evolutionists can state which of those
> already-specified aspects they dispute.
>
Good luck with it anyway,
Nic
an...@sci.sci
> lots of people want, but sometimes it just ain't gonna happen
What I'm looking for are labels that identify the various topics, then
links can point to additional explanation. Ideally a single word would
suffice to clearly identify the sub-topic, but most of the jargon
regarding evolution is ambiguous and misunderstood so a single word
wouldn't suffice and some phrase is needed instead.
I just want a check-list. To Joe-anti-evolutionist, which of the
following aspects do you dispute ... Joe can then respond with a list
of which he considers to be mistaken. For example, if Joe disagrees
that layers of sediment can be dated by ratios of isotopes, such as
uranium and lead, that would be a specific point of dispute where we
can then discuss why Joe disagrees with published reports of such
datings. If Joe says he disagrees with evolution, but he agrees with
every item we list, then obviously he agrees with evolution but
disagrees with some strawman idea of evolution he heard from elsewhere.
> (tell me how a TV works in 25 words or less- and I know almost
> nothing about electricity).
Here's a partial check-list for somebody who doesn't believe TV is possible:
- Using lens to create a real 2-d image of a view of a 3-d scene.
- Using an image-to-data device, such as a charge-coupled chip.
- Copying such data from I-D device to main transmitter.
- Encoding data in standard NTSC or PAL format.
- Modulating carrier to yield RF signal, and feeding to antenna.
- EM waves traveling from transmitting antenna to receiving antenna.
- Selecting particular band of RF signal and amplifying it.
- Demodulating RF signal to yield NTSC or PAL signal.
- Decoding NTSC or PAL format to actual scanlines of brightness values.
- Scanning across a raster such as CRT line-by-line to build up 2-d image.
For closed-circuit systems, omit modulating & EM & selecting & demodulating.
> But you might start with the FAQ at
> www.talkorigins.org
I assume you're referring to:
<http://www.faqs.org/faqs/talk-origins/archive/part1/>
That is full of stupid argumentative/rhetorical questions from
creationists, and rebuttals to them. Not one of the questions is of the
form "what are the major parts of evolutionary theory", nor have I been
able to find such a toplevel checklist embedded within the answers to
any of the included questions.
I guess I'll have to compile my own list (starting from what I posted
earlier), and if others don't think my list is complete, tough shit,
because you can't expect me to get it all correct myself without any
help. I'm thinking of dividing my list into two sections, direct
observations such as fossil record, and underlying theory as to what
caused stuff to appear that way.
Side remark: One of the answers in the FAQ had a link to:
<http://www.talkorigins.org/faqs/evolution-definition.html>
The ontogeny of an individual is not considered
evolution; individual organisms do not evolve. The changes in
populations that are considered evolutionary are those that are
inheritable via the genetic material from one generation to the
next.
Circular definition: Genetic material is defined as whatever is passed
from one generation to the next and down the line many generations, yet
here that definition is used backwards creating a definitional cycle.
It is important to note that biological evolution refers to
populations and not to individuals and that the changes must be passed
on to the next generation. In practice this means that,
What is a "population"? In sexually mating species, it can be defined
as the whole species, or as some part of the species that is
reproductively isolated from the rest of the same species. But for life
that doesn't engage in mating, every individual is on its own. There
are clades at the individual level, that is an exact tree of
relationships with each node being an individual, but no such thing as
a species except if you arbitrarily pick some clade going back so many
million years and thereby include every individual in that clade. For
non-mating clades, reproductive isolation is useless to define a
"population" since every individual is reproductively isolated from
every other individual. Even if you somehow define a particular size
clade as a "population", allelle changes within that population are
meaningless.
Also, what is "generation" in any species that mates at arbitrary times
and mates more than once in a lifetime? "Generation" has precise
meaning only for species that do synchronous once-per-lifetime mating,
whereby each generation gives birth to next generation all about the
same time, then dies out shortly after the next generation achieves
birth or maturity or other standard point in life cycle. For example
most "annual" flowering plants have well-defined generations.
.
Noone Inparticular
an...@sci.sci wrote:
> > Some things cannot be described in 25 words or less. That is what
> > lots of people want, but sometimes it just ain't gonna happen
>
> What I'm looking for are labels that identify the various topics, then
> links can point to additional explanation. Ideally a single word would
> suffice to clearly identify the sub-topic, but most of the jargon
> regarding evolution is ambiguous and misunderstood so a single word
> wouldn't suffice and some phrase is needed instead.
>
> I just want a check-list. To Joe-anti-evolutionist, which of the
> following aspects do you dispute ... Joe can then respond with a list
> of which he considers to be mistaken. For example, if Joe disagrees
> that layers of sediment can be dated by ratios of isotopes, such as
> uranium and lead, that would be a specific point of dispute where we
> can then discuss why Joe disagrees with published reports of such
> datings. If Joe says he disagrees with evolution, but he agrees with
> every item we list, then obviously he agrees with evolution but
> disagrees with some strawman idea of evolution he heard from elsewhere.
>
> > (tell me how a TV works in 25 words or less- and I know almost
> > nothing about electricity).
>
> Here's a partial check-list for somebody who doesn't believe TV is possible:
> - Using lens to create a real 2-d image of a view of a 3-d scene.
> - Using an image-to-data device, such as a charge-coupled chip.
What's an "image-to-data device"? What's a "charge-coupled chip"?
> - Copying such data from I-D device to main transmitter.
I-D device? Transmitter? What are they? How do they work?
> - Encoding data in standard NTSC or PAL format.
What is NTSC? PAL? How is data encoded in these formats?
> - Modulating carrier to yield RF signal, and feeding to antenna.
What is a "carrier" and how is it modulated? What is RF?
> - EM waves traveling from transmitting antenna to receiving antenna.
EM? What is it about EM that makes it ammenable? Why do you say they
exist as waves?
> - Selecting particular band of RF signal and amplifying it.
How many "bands" are there? What are they? How do you select them? How
are they amplified? By what?
> - Demodulating RF signal to yield NTSC or PAL signal.
What does demodulating mean?
> - Decoding NTSC or PAL format to actual scanlines of brightness values.
What does "Decoding NTSC or PAL fromat" mean? How is this done? What is
an "actual scanline"?
> - Scanning across a raster such as CRT line-by-line to build up 2-d image.
What is a "raster"? CRT? How is a 2-d image built up by this process?
> For closed-circuit systems, omit modulating & EM & selecting & demodulating.
Not so easy to do in a short list, is it?
<snip rest>
an...@sci.sci
How am I supposed to answer a question that is asked in the negative?
> 'Content' drift due to random mutation, and allele population drift
> (amongst mutually equally beneficial alleles) in the absence of
> mutation
I specify random mutation, and *then* either selection or drift.
Accordingly, per my way of phrasing it, only the latter of the two you
stated fits within what I call drift.
It is true that a very slight new-allelle drift occurs every time a
single mutation happens in a single individual, but such a singleton
event is infinitesimal among the statistics of a large population. For
example, if there are a million individuals, and one of them gets a
mutation, that's one in a million, below the level of measurement
error, but when the frequency drifts up to one thousand out of a
million, then the fraction gets large enough somebody might be able to
actually measure it to a single digit of accuracy.
Now in an extremely small population, a single mutation event might be
significant, and if the population is totally known (such as last
survivors of an endangered species in zoos) it might be possible to
directly measure the genome of *every* individual and thereby compute
exactly what fraction of them have the mutation (1/n where n is the
total number of haploid individuals, or 1/2n if it's just one
chromosomes among two in a diploid pair). However such cases are quite
rare, and in explaining evolution to novices such cases can be omitted.
> - eventually leading to fixity of one of them.
Not necessarily. With more than two allelles initially, one allelle
might go extinct, reducing the total number by one, but not fixing any
particular one of the remaining allelles. And then even with only two
allelles for a while, neither might go extinct, hence neither become
fixed, especially in a large population, before a new mutation
increases the number back to three or more. It's possible there are
loci which have survived for half a billion years with two or more
allelles at all times throughout that long time span, even if no more
than one allelle has survived for more than a hundred million years or
so.
> > - Some parts of eukaryotic cells derive from freeliving prokaryotes
> > which have been trapped in the cells and co-evolving for a long time.
> Has that to do with evolution per se?
Yes. Endosymbiosis and subsequent co-evolution is one type of change in
genome of indivduals hence change of statistics of genomes across
populations, whereby two separate populations (because they were of
different species) evolve to become a single population (now regarded
as a single new species), containing the merge of the two original
genomes. This doesn't happen all at one moment. Just like species-split
("speciation") events, whereby a single species gradually becomes two
species, less and less able to cross-breed, except in reverse time:
endosymbiosis involves gradual increase of dependency whereby each
original species is less capable of living without the other, until
eventually there is not a single remaining individual of either old
species that comes via escape from recent individuals of the new
endosymbiotic species. Both original species may very well continue to
survive somewhere, but co-evolving descendents from those original
species are no longer viable individuals, but are now merely parts of
individuals of the new endosymbiotic species.
> How about a general point on symbiosis, plus the point that it can
> end in complete assimilation?
Perhaps. I'm thinking that in addition to evidence of evolution
(fossils) and theory of evolution (mutation plus drift/selection), I
might also include prerequisites (radioactive decay, plate tektonics,
sedimentation, etc.), where symbiosis in general might appear in that
'prerequisites' section. But for my purpose I'm not sure it's worth
listing all the prerequisites together. It might be better to list only
the specially needed prerequisites associated with each particular type
of evidence. For example, to date fossils, you need to date crystals
within the sediment, by radioactive decay, and you need to know when
the layer of sediment can be dated as a whole by those crystals, and
you need to know when to expect that a particular fossil is associated
with a particular layer of sediment rather than some sort of intrusion
from another time (such as coal-mining equipment found in abandoned
coal mines, or tree roots extending several meters down into soft
sediment, or ancient fossils eroded from ancient walls of river and
washed downstream and deposited in modern sediment). So when the
anti-evolutionist claims a problem with dating of fossils, *then* that
person is asked whether he/she accepts various theories of radioactive
dating and sedimentation etc.
> > - Often two different species, or two sexes of a single species,
> > compete with each other over a long period of time, whereby each
> > evolves an attack against the other whereupon the other evolves a
> > defense to that attack whereupon the former evolves a better attack and
> > the latter a better defense etc. back and forth nearly endlessly.
> They don't even need to be different sexes. Competition creates
> hurdles for no reason. Trees don't *need* to be that tall.
Righto. I need to clean up the wording of that item to make it fully
general to include such cases. Do you know a single word which
encompasses that general topic and is well understood, not ambiguous or
misleading? If not, a short phrase that is clear and unambiguous?
> Do you mean there could be between two and twenty clades with no
> common ancestor?
Yes. The concept of clades at the species level work fine for sexually
mating species, whether they do meiosis, or they form cliches that
undergo horizontal gene transfer (HGT) within a cliche but not outside
that cliche. However during the very early days HGT might have involved
very fuzzy clades, to where all life whatsoever was a single "ring
species" with respect to HGT. It wouldn't be at all meaningful to
ignore HGT and consider only whole-cell cladistics, especially if more
than 50% of the genome of any individual came from HGT. But it wouldn't
be quite honest to claim Universal Common Ancestry merely because all
early life formed a single "ring species". So it may be best to cease
using whole-cell cladistic analysis earlier than some ancient epoch,
whenever we determine that HGT dominated over cellular reproduction
earlier than that time. So we'd do our cladistic analysis back to that
epoch, and thereby have several disjoint clades from that epoch forward
to the present, but before that time we'd organize life using a more
appropriate system instead of whole-cell cladistics. So the species
that were distinct at that transition epoch would be common ancestors
of their respective clades to the present, and that number might be as
many as twenty by my complete guestimate. So the claim of whole-cell
common ancestry, which is part of evolutionary theory, would go back
only to that transition epoch.
Earlier than the HGT/cladistics transitional epoch, we might track each
gene/haplotype individually, instead of tracking any whole-cell genomes
as a single unit. We might then discover that most genes (at the
transition epoch) evolved from a small set of primordial genes (at the
RNA/DNA takeover epoch), which in turn evolved from a small set of RNA
sequences that originally didn't have any genetic function but were
used for some other purpose by an earlier genome such as clay crystal
dislocations. So from the RNA/DNA transitional epoch to the
HGT/wholeCell transitional epocy, we'd be doing single-gene cladistics.
And before the RNA/DNA transitional epoch, back as far as the first
biosynthesis of RNA sequences for some other purpose, as the Chinese
would say, "interesting times", we'd need a totally differerent way of
tracking RNA over time, which I can't envision at this time because the
theory of RNA-world hasn't yet been defined well enough to say what the
original purpose of RNA really was.
Note: The above is based on my present assumption that there was some
sort of original replicator, which eventually fabricated RNA for some
non-genomic purpose then a takeover happened to the RNA-genome world,
then DNA was fabricated as a more reliable storage medium for
remembering RNA sequences and another takeover happended. But there may
never have been a RNA world. Instead, something other than RNA or DNA
might have been the mechanism for genome, then directly fabricated DNA.
In that case, "interesting times" all the way to the DNA takeover.
> I notice you didn't say anything about vestigial organs. I'm not
> saying you should have done, but they are controversial.
The talk.origins FAQ has, as its primary purpose, to respond to
frequently asked anti-evolution questions, mostly from Creationists and
IDiots. That's not my purpose here. I'm trying to organize/list the
various topics of evolutionary theory so that we can present a
checklist to anti-evolutionists: which parts of the evidence or theory
do you not accept? First there's evidence that needs explaining, with
the obvious surface explanation that the various fossils are remains of
formerly living critters, and that there's some biological relationship
between the various critters at various times within the same
geographic area. But once that's established, what's the deep
explanation as to what sort of biological relation we're observing, and
what mechanisms caused the succession of forms to appear. I don't
regard vestigal organs as the primary evidence that needs explaining,
so I think I'll exclude it from my checklist of evidence for evolution.
I'm thinking that the evidence that needs explaining by theory are:
- Apparent fossils (formerly living) buried over the past half billion years.
- Succession of form of fossils in each geographic area or uniformly worldwide.
- Apparent branchings of those lines of succession of forms.
- Nested hierarchy (NH) of many features closely matching fossil branchings.
- (so-far very limited) DNA cladograms closely match fossil branchings & NH.
- Complicated well-adapted features of biochemistry and mechanism.
- In-our-time new variation such as dog/horse/plant breeds and pathogens.
And the anti-evidence also needs explaining, basically every
mythological or sci-fi chimera, why it would make perfect sense for
such a being to exist, yet we never see any such in nature, only in our
fiction/fantasy/mythology, or in artificially-grafted plants such as
apple trees bearing grafted pear fruit or transplanted baboon hearts or
pig-skin grafts etc. And why a whole bunch of different marsupials are
*only* in Australia, and North/South poles with very different fauna
despite similar habitat. Why all those restrictions, beyond the basic
environmental requirements, on where a particular type of critter might
naturally live? And why those geographic restrictions aren't as
strictly applied to microbial life, yet even with such life there is
some correlation between geographic location and specific DNA
sequences? And why two different apparent clades use very different
methods of accomplishing the same goal, one better than the other,
instead of either using the other's solution so that both are using the
best known solution? Why does any apparent clade use an inferior
solution when a better solution has already been known for hundreds of
millions of years?
.
an...@sci.sci
> complex.
I agree. But see later below...
> evolution can be described (although not so completely) as
> imperfect replication plus selection.
You omitted neutral drift, caused by sampling error.
That's an important omission you need to correct.
But even if you can get a Creationist to admit that evolution sometimes
happens, you need to also show that evolution is the cause of the tens
of millions of species we currently observe, as well as the variation
within each species, as well as the complicated biochemistry and body
forms/functions. If the other person agrees that evolution happens, but
says that all of life as we currently know it was caused by God, not by
evolution, you have accomplished almost nothing.
For just the minimal theory of evolution, we need to ask the other
person to agree-with or dispute each of the following allegations:
- The characteristics (form and function) of any individual are partly
inherited from parent(s) and passed to descendents, which we call
"genetic phenotype", and partly acquired individually during life and
*not* passed to descendents, which we call "acquired characteristics".
- It's not the genetic phenotype which is directly inherited, but
rather the genome itself, which codes for the phenotype. Some parts of
the genome don't code for any phenotype at all. Some parts have
alternatives which code for exactly the same phenotype. But some
alternatives for the genotype really do code for different phenotype.
- In a sexually-reproducing species, such as humans, each individual's
genome is half from the mother and half from the father.
- In the process of replicating half+half of those genomes, occasional
mistakes in replication occur, causing new variations of the genome not
present in either parent, and some of these cause variations in the
phenotype.
- Each parent has two not-necessarily-identical copies of each part of
the genome.
- The half+half combination is approximately random, whereby any part
of either parent's genome has equal chance of one or the other
near-copy being included in the final combination, except that there's
linkage whereby in some sections of the genome one large chunk of one
alternative or the other is included or excluded as a unit.
- As a result of such random inclusion/exclusion, the total number of
copies of any particular part of the genome needn't remain constant
over the generations. By random chance, one alternative could happen to
be selected more often than the other, so its proportion would increase
at the expense of the other. But each generation, such random choice is
independent of any previous generation, so instead of a consistent
forced trend what happens is a "random walk", called "genetic drift".
- But sometimes one alternative section of genome results in a
phenotype significantly better adapted to circumstances than another
alternative for that same section of genome, so then individuals with
one alternative survive or reproduce better than individuals with the
other alternative, and so each generation the proportion of one
alternative increases while the other descreases, this consistent
change accumulating over many generations, until the less-adapted
phenotype is totally eliminated from the population, or circumstances
change to as to eliminate this difference in adaption thus stop the
consistent accumulation of one alternative at the espense of the other.
This forced change in proportion of alternatives is called "natural
selection".
- With asexually-reproducing life, the situation is somewhat simpler,
whereby each daughter gets *all* of the one parent's genome, but still
there are occasional mistakes in copying and hence new variation
between one daughter's clade (which got the mistake-copied genome) and
the other daughter's clade (which got the exactly-correctly-copied
genome). There's no genetic drift due to random sampling of half+half
of parents. The only genetic drift is due to random circumstances
causing one clade to reproduce more despite having no better genome
than the other clade, such as Bambi and her equally-good sister romping
in the meadow and Godzilla happens to step on one but not the other.
But then even after convincing the anti-evolutionist that all the above
theory is correct, still we must show how that explains the very
complicated evidence, such as apparently forking fossil record, and
nested hierarchy of form that closely matches it, and DNA cladograms
that seem to match both of those.
How would you propose to best organize all this theory and evidence to
make it easy to ask the anti-evolutionist to check which points he/she
agrees with and which disagrees with so as to clarify the debate? In
particular, how best to start with just a few basic points of possible
dispute, then expand to something about the size I've been relating in
this thread, then expand to more detailed aspects of each disputed
point? How can you condense the nine bullet points (509 words total) I
listed just above into something that expresses all that content in 25
words or less as a starter statement? Your four-word phrase "imperfect
replication plus selection", or my improvement upon it to include
drift, probably six words total, may suffice as a very preliminary
statement, a pre-starter or teaser perhaps, but I think something
between 6 words and 509 words is needed to bridge the gap between
those. What's the missing link we need here? Perhaps three or four
bullet points of 5-10 words each? I don't have any good ideas here.
.
Larry Moran
an...@sci.sci <an...@sci.sci> wrote:
[snip]
>> evolution can be described (although not so completely) as
>> imperfect replication plus selection.
>
> You omitted neutral drift, caused by sampling error.
> That's an important omission you need to correct.
The mechanism is called random genetic drift. The affected alleles
may be beneficial, neutral, or detrimental with respect to natural
selection. The frequencies of all three types of allele can be
influenced by random genetic drift so it's not appropriate to
refer to the mechanism as "neutral drift."
That's an important error you need to correct.
Larry Moran
an...@sci.sci
> may be beneficial, neutral, or detrimental with respect to natural
> selection. The frequencies of all three types of allele can be
> influenced by random genetic drift so it's not appropriate to
> refer to the mechanism as "neutral drift."
Ah, yes, I was thinking of which factor dominates. If there's no
selection, drift dominates, and it's truly neutral, whereas if there's
lots of selection, selection dominates, so you can mostly ignore drift.
But to handle the between cases, where there's selection but there's
also drift of comparable amount, so you must not ignore either, your
point is well taken.
By the way, it seems to me that random genetic drift due to sampling
error happens only during meiosis, so it doesn't apply to asexually
reproducing cells. Is that basically correct?
So in a nutshell (well a rather *large* "nut", such as coconut, shell):
Replication, random mutations, sampling-of-genome error in meiosis only
(random genetic drift), sampling-of-environment error in normal life
situations, and selection bias due to differences of "fitness".
No other known significant factors in evolution process, correct??
(No goddiddit, no space aliens tinkering with earthers, etc., AFAWK)
The most interesting factor turns out to be one specific type of
mutation, namely largescale duplications, which thereby create two
copies of each gene within the duplicated segment, allowing each copy
to specialize to a different subset of what the original gene did.
Everything else about evolution is pretty much what Darwin and Wallace
independently figured out, not knowing about DNA but still their
unknown genetic means would have similar consequences, and what Malthus
might have figured out earlier if he'd thought about the issue.
But largescale duplication events present a major conceptual advance
beyond what Darwin and Wallace had in mind. With no idea about DNA,
they probably would have thought that copying the genes to make legs
would result in two sets of legs, a disaster, hence not worth
considering in evolutionary theory. But now we know duplication merely
makes some biochemical pathways more productive than necessary, which
then get down-regulated to restore balance, but then can diverge to
make new biochemical pathways, yielding novel capability without losing
essential ancestral capability. I think this is the key improvement to
Darwin/Wallace theory that solves the "irreducible complexity" problem.
A new capability can just pop out of thin air if it's biochemically
similar enough to some old capability that duplication followed by
divergence can yield it. And it doesn't actually have to be similar, if
a slight change in amino acids can cause one polypeptide to fold in a
very different way thereby drastically changing its enzymatic behavior.
Does anybody know how commonly drastic changes in folding, due to
mutation, generate new useful function apparently unrelated to the
original function?
By the way, per my original purpose in this thread: I'm currently
leaning toward two stipulate-or-dispute checklists:
- Evidence, which emphasizes three "trees" which all agree with each other:
-- Apparent evolutionary tree as shown by fossil record.
-- Nested heirarchy in characteristics, especially development of embryo.
-- Cladograms computed from DNA sequences.
- Theory basically as I coconutshelled it above, which is claimed to be
consistent with the evidence, and not consistent with the non-evidence
(such as chimeras that violate time/geographical constraints of evolution).
("Stipulate" in the legal sense. If you agree to a point raised by the
opposing lawyer, without needing any further evidence, you just accept
the allegation as true, and the judge/jury then treats it as if true
beyond reasonable doubt.)
.
an...@sci.sci
> > - Using lens to create a real 2-d image of a view of a 3-d scene.
You didn't dispute that item, so I presume you're willing to stipulate
that you believe it's possible?
> > - Using an image-to-data device, such as a charge-coupled chip.
> What's an "image-to-data device"?
Anything that converts a visual image into a data representation of
that image. Your local Kinko's probably has one of them. You lay your
original on a glass surface, like a copying machine, and you insert a
diskette, and you close the lid over the original, and press the
take-image button, and the device scans your image with a bright light
and records the reflection brightness values and generates a data file,
which is then written onto your diskette. Deep inside that large
machine is the particular image-to-data device which does the key work
of converting visual image to data.
> What's a "charge-coupled chip"?
One particular kind of I-D device. Check Sky&Telescope for ads about
charge-coupled-device (CCD) chips, then go to Fry's electronics to
purchase one of your very own.
Most of the rest of your questions you can answer yourself by doing a
Google search on each term, such as "modulate" or "RF" (Radio
Frequency), or "EM" (Electromagnetic Radiation), etc. Apparently you
dispute quite a number of those parts of TV theory/practice, which
means you have quite a lot of studying to do before you will understand
all the parts of the theory well enough to put them together to
understand the overall theory of how TV works.
> Not so easy to do in a short list, is it?
What, my list wasn't short enough? It had only ten items. How short do
you require the list to be? Do I need to satisfy military requirements
that each level of hierarchy must have no fewer than three and no more
than seven items?
.
Larry Moran
>> The mechanism is called random genetic drift. The affected alleles
>> may be beneficial, neutral, or detrimental with respect to natural
>> selection. The frequencies of all three types of allele can be
>> influenced by random genetic drift so it's not appropriate to
>> refer to the mechanism as "neutral drift."
>
> Ah, yes, I was thinking of which factor dominates. If there's no
> selection, drift dominates, and it's truly neutral, whereas if there's
> lots of selection, selection dominates, so you can mostly ignore drift.
> But to handle the between cases, where there's selection but there's
> also drift of comparable amount, so you must not ignore either, your
> point is well taken.
>
> By the way, it seems to me that random genetic drift due to sampling
> error happens only during meiosis, so it doesn't apply to asexually
> reproducing cells. Is that basically correct?
No, ... not even close.
Imagine a population of one million single-cell organisms that divide
by binary fission. This could be mitosis if the cells are eukaryotic.
The population size doesn't change from generation to generation so
half of the daughter cells die before reproducing. Sometimes both
daughter cells die and the lineage comes to an end. Sometimes both
survive and the allele frequency of that lineage increases.
Allele frequencies in the population change over time due to the
random survival of the daughter cells. That's random genetic drift.
The founder effect is another example of random genetic drift.
The same effect applies to sexually reproducing organisms. Some
die by chance and some survive by chance. The classic textbook case
is flowers that are killed in a mudslide. Other good examples are
squirrels that are run over by a car and innocent Middle Eastern
citizens who are blown up by suicide bombers.
Larry Moran
Noone Inparticular
<blink> <blink>
Try to remember long ago in the thread in which these posts reside. A
poster, a quite erudite and highly respected one, wrote in response to
your request for a 25 word explaination of the major features of
evolution;
"Some things cannot be described in 25 words or less. That is what lots
of people want, but sometimes it just ain't gonna happen (tell me how a
TV works in 25 words or less- and I know almost nothing about
electricity)."
In response to this you wrote the stufff above, even though what you
wrote was far more than 25 words.
Now this is where it gets tricky. Try to keep up.
I asked you some additional questions to illustrate to you what Chris
tried (vainly, it appears) to get across to you; it is nigh on
impossible to explain complicated things in very brief ways (this
phenomena is the chief reason the politics of the US is in such a
terrible way these days). In trying to explain the major features of
evolution in 25 words or less you wind up with an explanation that
explains little if anything. Much like your TV analogy.
John Wilkins
>>The mechanism is called random genetic drift. The affected alleles
>>may be beneficial, neutral, or detrimental with respect to natural
>>selection. The frequencies of all three types of allele can be
>>influenced by random genetic drift so it's not appropriate to
>>refer to the mechanism as "neutral drift."
>
>
> Ah, yes, I was thinking of which factor dominates. If there's no
> selection, drift dominates, and it's truly neutral, whereas if there's
> lots of selection, selection dominates, so you can mostly ignore drift.
> But to handle the between cases, where there's selection but there's
> also drift of comparable amount, so you must not ignore either, your
> point is well taken.
And this is why I prefer Gavrilets these days - selection maintains organisms
at a fitness level. Drift applies within the fitness ridges that percolate
through the fitness landscape. Drift can apply when the trait sits low on the
fitness landscape, or when it sits high.
...
Not having a go at you, anon1, but noting how this has changed the way I
conceive of this.
--
John S. Wilkins, Postdoctoral Research Fellow, Biohumanities Project
University of Queensland - Blog: evolvethought.blogspot.com
Nihil tam absurdum quod non quidam Philosophi dixerit - adapted from Cicero
an...@sci.sci
> > error happens only during meiosis, so it doesn't apply to asexually
> > reproducing cells. Is that basically correct?
genome, which is inherent in the reproduction process. I treat
differential survival due to sampling of environment as a different
factor elsewhere.)
> No, ... not even close.
> Imagine a population of one million single-cell organisms that divide
> by binary fission.
One million clones (identical genomes), or one million of a clade, but
with accumulated differences in genome due to mutations that have
happened since the LCA of the clade?
> This could be mitosis if the cells are eukaryotic.
> The population size doesn't change from generation to generation so
> half of the daughter cells die before reproducing. Sometimes both
> daughter cells die and the lineage comes to an end. Sometimes both
> survive and the allele frequency of that lineage increases.
If they are identical clones, this makes no difference whatsoever in
the resultant allelle frequencies.
If they have diverged due to mutation from LCA, their differential
survival could be due to either differences in phenotype which make one
pair of daughters more fit than another pair, or due to differences in
sampling of environmental factors, for example one pair happened to be
in a resource-depleted region so both starved while other pair happened
to be in a resource-abundant region so both fed well and survived.
I don't know which situation (ident, diff/fitter, diff/random) you're
quibbling over. diff/fitter is already covered under selection bias,
and diff/random is already covered under random sampling of environment.
Neither needs to be included within random sampling of parent's genome.
> Allele frequencies in the population change over time due to the
> random survival of the daughter cells. That's random genetic drift.
If you want to use a single term that conflates two different concepts,
intrinsic random sampling of parent's genome in meiosis, and extrinsic
variation in survival due to life's circumstances which have a chaotic
element, that's your choice. But I think such conflation would confuse
novices trying to understand evolutionary theory. I think it's better
to make it clear that replication itself introduces both mutations
(mistakes in copying, in all forms of replication) and sampling error
(in meiosis only), while survival to maturity and to make babies
depends on both intrinisic fitness and extrinsic circumstances that
happen.
> The founder effect is another example of random genetic drift.
Well the chance event that Joe instead of Sue happens to wander across
the river during the dry season and find lush rainforest and thereafter
have much opportunity to grow exponentially, while everyone else
remains on the original territory fighting for resources and limited in
number to a kind of statis, is indeed somewhat random. But that's more
of a "saltation" or "catastrophe" than a drift. Suddenly that
individual can reproduce exponentially like never before. I really
don't think "drift" is at all the right word to describe such sudden
radiation-in-number events which later yield radiation-in-variety from
that one founder genome.
> The same effect applies to sexually reproducing organisms. Some
> die by chance and some survive by chance. The classic textbook case
> is flowers that are killed in a mudslide. Other good examples are
> squirrels that are run over by a car and innocent Middle Eastern
> citizens who are blown up by suicide bombers.
That's all covreed under random sampling of environment, which is a
factor in differential survival. I think you may be responding to an
early version of my list of aspects of evolutionary theory, before I
clarified it well by separating intrinsic reproductive factors from
later environmental/fitness factors. Too bad there's no way I can set
up an anonymous Web site where I can maintain the latest version of my
ideas, so if I make an improvement you can see it immediately the next
time you look at it. As it is now, any improvement I make is hidden
deep inside some discussion thread, and even if you're reading
elsewhere within the same thread you wouldn't see that latest version.
.
John Harshman
How about this as a definition of genetic drift?: differential
reproductive success not correlated with genotype.
r norman
<jharshman....@pacbell.net> wrote:
>How about this as a definition of genetic drift?: differential
>reproductive success not correlated with genotype.
This suggests that it is correlated with something else. The whole
point of drift is that it is not correlated with anything.
Genetic drift is a change in the genetic composition of a population
due to stochastic processes in the transmission of genetic information
from generation to generation and not correlated with any biological
or environmental feature.
an...@sci.sci
> features of evolution;
No, I didn't make that request. You've built a strawman by misquoting me.
> "Some things cannot be described in 25 words or less."
And that's another strawman, made by the other person, not me.
> I asked you some additional questions to illustrate to you what Chris
> tried (vainly, it appears) to get across to you; it is nigh on
> impossible to explain complicated things in very brief ways ...
And now *your* strawman again. I never asked for anyone to explan
anything complicated in 25 words or less.
Here's my original request verbatum:
Is there a Web site which concisely (like 25 words or less each)
summarizes the major aspects of (modern mainstream) (biological)
evolutionary theory?
Note that the 25-word limit is per item, and the whole list *summarizes*,
does not explain or provide explanation, nor describe, merely summarizes!
You need to learn the difference between:
- Describe - Tell all the different characteristics of something.
- Explain - Tell in detail how that thing works or how it was manufactured.
- Summarize - Just give a quick synopsis of either of the above.
Ideally there'd just be a list of ten or twenty technical terms, each
naming one aspect of evolutionary theory, each term a single word or a
short phrase, and together all these items would *summarize* what we
mean by evolutionary theory. But if there isn't any simple word or
brief phrase that tells a novice what the heck we're talking about,
then a slightly longer but still brief text clarifying the topic, no
longer than 25 words, would be acceptable.
I was hoping there'd be a Web page that already had such very brief
tags on the various aspects to evolutionary theory, where the whole Web
page would fit on a half page of paper (one old-style 24-line CRT
terminal screen), with links to additional information, but visually
the front page would be easily seen in such very concise format.
But so-far nobody has pointed me to anything like that, so I've been
forced to draft my own from scratch.
For example, the table of contents of the Wikipedia page on evolution
is not at all what I'm looking for. The portion of that table of
contents dealing with basic mechanisms of evolution:
+ 6.1 Genetic variation
o 6.1.1 Gene transfer
o 6.1.2 Mutation
o 6.1.3 Gene flow
+ 6.2 Natural selection
+ 6.3 Genetic drift
is not quite what I'm looking for, not at all really. Likewise the
section of that table of contents on evidence of evolution:
+ 4.1 Morphological evidence
+ 4.2 Genetic sequence evidence
is not at all what I am looking for, both because it makes no mention
of any nested hierarchy, and because it lists only two types of
evidence instead of all three. The point is that evidence isn't
additive. You don't have morphological evidence by itself, then you
discard all that and have genetic sequence evidence by itself, and then
you add the quantities of the two kinds of evidence and pass some
threshold. The point is that if you look only at currently living
forms, looking carefully at the way the embryo develops, you see a
clear nested hierarchy, no doubt about it, no chimeras, pure and simple
nesting in a Platonic classification system. Then if you look only at
fossils over time and space, you see a clear pattern of variation over
time, including splitting events, forming an evolutionary tree, and
amazingly that tree matches the nested hierarchy you got before. Then
if you look only at DNA cladograms, you get another tree, and amazingly
that tree matches the trees you got before. Why should there be any
nested hierarchy at all? Why not like a relational database about makes
and models of automobiles where each company borrows at will various
inventions from other companies? Why should fossil patterns follow a
tree shape either? And why should DNA sequences be easy to put into
stable cladograms? Why shouldn't nearest-neighbor or parsimony graphs
show a haywire of connections as you draw a map of all possible roads
between cities in the USA? Why do they seem to nicely follow unrooted
trees (which can be presumed to match rooted trees with some unknown
point being the root)? No explanation for the three trees that appear
from such analysis, and no explanation for matching between those three
differently-generated trees, except some kind of actual evolutionary
process that has occurred over time.
> In trying to explain the major features of evolution in 25 words or
> less you wind up with an explanation that explains little if anything.
I've never tried any such thing. You are attacking a strawman. You
obviously didn't read my original request carefully enough to
understand what I was requesting.
> Much like your TV analogy.
My TV anology was a good example of what I originally requested.
It was a simple checklist of topics, which together explain how a TV
works, but the checklist of topics merely gave a brief one-line summary
of each topic, not trying to explain any of them, just identifying what
each topic is about, so if you already know that topic you can skip it
but if you don't know that topic you can ask for more info about it.
You obviously completely misunderstood the point I was making, just as
you misunderstood my original request. Go back to the start of this
thread and try reading for better comprehension.
.
John Harshman
> On Fri, 02 Dec 2005 20:57:11 GMT, John Harshman
> <jharshman....@pacbell.net> wrote:
>
>
>
>>How about this as a definition of genetic drift?: differential
>>reproductive success not correlated with genotype.
>
> This suggests that it is correlated with something else. The whole
> point of drift is that it is not correlated with anything.
What? Not being correlated with X suggests a correlation with Y? I have
no idea where that would come from.
Anyway, drift can be correlated with something, as long as it's not
genetic. It could be correlated with stepping in front of a bus, for
example. Raup's "field of bullets" is drift, but survival is definitely
correlated, in that scenario, with not being in the path of a bullet.
> Genetic drift is a change in the genetic composition of a population
> due to stochastic processes in the transmission of genetic information
> from generation to generation and not correlated with any biological
> or environmental feature.
No environmental feature? Suppose, for example, that predators for some
reason become twice as common in valley A as in valley B, and, as a
result, whatever genotypes are prevalent in valley B increase in
frequency. That's not drift? I would certainly say that it was.
Stochastic changes in genotype frequency must have some kind of physical
mechanism. As long as that mechanism is uncorrelated with genotype
(which is what "stochastic" or "random" commonly means to evolutionary
biologists), it would seem to count.
an...@sci.sci
No. Selection either kills an organism or fails to kill it. Selection
never modifies an organism nor prevents an organism from being modified
by a(n early-development) mutation. Selection never maintains any
individual organism at any particular fitness level.
Selection doesn't even maintain a line of (asexual) descent to remain
at its original fitness level. It merely kills off any lines of descent
that deviate toward bad fitness due to mutation along the way, while
not killing off lines of descent that stay at the original fitness
level. No individual line of descent is maintained at a fitness level
by any action of selection.
In the case of sexual reproduction (meiosis), there's no such thing as
lines of descent, every replication event merges half+half of parents,
so there's nothing from the previous generation that even can be
maintained in the next generation much less multi-generation lines of
descent.
Selection maintains the statistics of a population, nothing else, not
individuals and not lines of descent. Only when a population has a lot
of very fit individuals, plus a few mutations from that fitness that
are inferior, selection eliminates the individuals with that bad
mutation, thereby restoring the original statistics that had been
present before the mutations happened, therefore in total preventing
statistics from drifting far from 100% best-fit among all genomes
occurring in the population.
> Drift applies within the fitness ridges that percolate through the
> fitness landscape.
Yes, but two caveats/nitpicks here:
The so-called "ridge" is really a multi-dimensional descrete-choice set
in most cases. For example, if there are n different SNPs active in the
population, each neutral with respect to fitness, each on a different
haplotype block, then there are 2**n different combinations possible,
and all those 2**n are explored as meiosis mixes and matches the SNPs
from generation to generation. It's a major geometrical mistake to
picture it as a linear ridge with travel continuously in a linear
manner along that ridge.
If you talk about populations instead of individuals, then you are not
talking about traveling along a ridge, rather you are talking about
quantum jumps from one point to another within a descrete space,
whereby the populations at each such point change over time, so in the
abstract space of population statistics there is "drift". But this is
entirely different from drift along a ridge of fitness landscape.
Now in the rare case where you have several SNPs that are additive for
some phenotype character which does not affect fitness, i.e. the
phenotype is proportional to the number of the SNPs which are in one
state compared to the other, then you can approximate the descrete
binomial distribution as if it were a continous line. For example, with
4 SNPs that are additive, with each occurring half the time, you have
equilibrium populations of (1 4 6 4 1)/16 along that linear scale, and
drift due to environment sampling can skew that distribution away from
equilibrium. But that's so rare as to be not worth considering.
> Drift can apply when the trait sits low on the fitness landscape, or
> when it sits high.
Yes, but selection tends to eliminate all genomes on the lower level,
allowing genomes on the higher level to increase in number, so after
selection has run its course all individuals at the lower-level are
gone and there's drift of population statistics only along the highest
level that was present, the *only* level that remains in use by the
surviving population.
Now in virtually every case there's a **potential** mutation, such as a
complete re-write of the genome, about as common as all the electrons
in the Universe jumping at one time, which would result in better
fitness than anything currently anywhere on Earth. So if you consider
that pie-in-sky **potential** genome as "high" and anything that
actually exists in the population as "low", you are correct, but that's
not a reasonble way of thinking about it. Among genomes actually
present in a population at a particular time, selection eventually gets
rid of all but the highest (most fit), making the drift-at-low-levels
irrelevant to longterm statistics. (And in a sexually-reproducing
population, where haplotype blocks are mixed&matched in all
combinations, all but the fittest allelle of each haplotype block are
eventually eliminated from the population, except where they aren't
additive in fitness, such as where a mix is better than either extreme,
where some sort of mix is maintained, complicating our analysis.)
.
Nic
> How about this as a definition of genetic drift?: differential
> reproductive success not correlated with genotype.
Maybe, but I think there is an enormous amount of drift even if all
misadventure could be eliminated say, by having a guardian angel
assigned to each organism.
Differential reproductive success due to what you've got at one locus
is random success for what you've got at all the other loci. Good
allele A gets killed because time is up for bad allele B at some other
locus. Probably happens a lot more often than mudslides.
Nic
John Harshman
True if you consider only one locus at a time. But we realize that
selection happens to entire organisms, and fitness is a term that really
applies to entire genomes. In one-locus models, the implicit assumption
is that the genetic background is otherwise uniform in the population.
Multi-locus models define the fitnesses of genotypes, even if they are
only the sum of allelic fitnesses. That's not drift.
Nic
It's not drift, or it's no drift?
What I mean by that question is, where there is selection, you can say
there is no drift, but I think that is just a way of saying the drift
component averages out to zero. I thought the point under discussion
here is whether the averages taken by nature (over population, and over
species life time) are such that we can consider that drift averages
out to zero in all but the (near enough) completely unselected cases.
I think it is valid to ask whether drift in typical population sizes,
over typical species life times, is capable of overcoming weak to
moderate selective pressure.
Larry Moran:
> The mechanism is called random genetic drift. The affected alleles
> may be beneficial, neutral, or detrimental with respect to natural
> selection. The frequencies of all three types of allele can be
> influenced by random genetic drift so it's not appropriate to
> refer to the mechanism as "neutral drift."
The point in my post was however a different one: that the 'engine' of
random drift is there without any need to bring in misadventures like
mudslides. Indeed, who is to say that the mudslide isn't selective
pressure for an allele that we haven't discovered yet? It causes just
the same irrelevant increases and decreases in the frequencies of
alleles we *have* discovered.
Nic
John Harshman
No, not true. Selection and drift are two different processes that can
operate simultaneously. Of course any drift component always averages
out to zero, since there is no force pushing it in any direction. But of
course, also, no single case is average, and an allele subject only to
drift has only two possible fates: extinction or fixation, with the
relative probabilitie of each at any moment equal to 1-p and p,
respectively, where p is the current frequency.
> I thought the point under discussion
> here is whether the averages taken by nature (over population, and over
> species life time) are such that we can consider that drift averages
> out to zero in all but the (near enough) completely unselected cases.
I can make little sense of this. Drift and selection operate
simultaneously to affect the probability of fixation or extinction of
alleles. Whether drift or selection is the most important parameter in
any given case depends on both the selective value of the alleles and
the size of the population.
> I think it is valid to ask whether drift in typical population sizes,
> over typical species life times, is capable of overcoming weak to
> moderate selective pressure.
Yes, it is, depending on your particular values of population size and
"weak to moderate". A text on population genetics would provide you with
some nice equations that would tell you the exact relationships among
these parameters.
Nic
> >>>Differential reproductive success due to what you've got at one locus
> >>>is random success for what you've got at all the other loci. Good
> >>>allele A gets killed because time is up for bad allele B at some other
> >>>locus. Probably happens a lot more often than mudslides.
> >>True if you consider only one locus at a time. But we realize that
> >>selection happens to entire organisms, and fitness is a term that really
> >>applies to entire genomes. In one-locus models, the implicit assumption
> >>is that the genetic background is otherwise uniform in the population.
> >>Multi-locus models define the fitnesses of genotypes, even if they are
> >>only the sum of allelic fitnesses. That's not drift.
> >
> >
> > It's not drift, or it's no drift?
> > What I mean by that question is, where there is selection, you can say
> > there is no drift, but I think that is just a way of saying the drift
> > component averages out to zero.
>
> No, not true. Selection and drift are two different processes that can
> operate simultaneously.
understood what you meant. That's the big picture, and it isn't drift
in anybody's book - it's natural selection. The small picture is of
allele A's evolution suffering a set back, even though selection for it
will prevail in the long run. I was trying to make the point that one
*could* call that a case of drift in opposition to selective pressure.
Further below, you agree it's OK to talk like that.
> Of course any drift component always averages
> out to zero, since there is no force pushing it in any direction. But of
> course, also, no single case is average, and an allele subject only to
> drift has only two possible fates: extinction or fixation, with the
> relative probabilitie of each at any moment equal to 1-p and p,
> respectively, where p is the current frequency.
>
> > I thought the point under discussion
> > here is whether the averages taken by nature (over population, and over
> > species life time) are such that we can consider that drift averages
> > out to zero in all but the (near enough) completely unselected cases.
>
> I can make little sense of this.
more than what I put in my next paragraph.
> Drift and selection operate
> simultaneously to affect the probability of fixation or extinction of
> alleles. Whether drift or selection is the most important parameter in
> any given case depends on both the selective value of the alleles and
> the size of the population.
>
> > I think it is valid to ask whether drift in typical population sizes,
> > over typical species life times, is capable of overcoming weak to
> > moderate selective pressure.
>
> Yes, it is, depending on your particular values of population size and
> "weak to moderate". A text on population genetics would provide you with
> some nice equations that would tell you the exact relationships among
> these parameters.
What if your drift definition was changed from: differential
reproductive success not correlated with genotype.
To: differential allelic reproductive success not correlated with that
allele?
How prevalent are neutral polymorphisms which are currently drifting?
I had always assumed they would be quite rare and short lived simply
because the so called molecular clock seems to work, and that wouldn't
be much use if things were taking 10s of millions of years to become
fixed or go extinct (or were taking a length of time which is a strong
function of population). I now think my reasoning may be flawed in
this.
Nic
John Harshman
No, I don't think I do. You are talking here about opposing selection
pressures. It's not drift. Nobody says its drift, or treats it as drift.
Then you would have a definition of drift that everyone else would
consider selection, and you would need different equations, which you
could also find in a population genetics text. Are the loci unlinked?
What is the nature of their interaction?
> How prevalent are neutral polymorphisms which are currently drifting?
Most sites in the human genome are evolving neutrally.
> I had always assumed they would be quite rare and short lived simply
> because the so called molecular clock seems to work, and that wouldn't
> be much use if things were taking 10s of millions of years to become
> fixed or go extinct (or were taking a length of time which is a strong
> function of population). I now think my reasoning may be flawed in
> this.
Indeed it is. If there's a molecular clock, it works precisely because
of drift. However, it has a huge variance.
r norman
<jharshman....@pacbell.net> wrote:
<snip>
> Selection and drift are two different processes that can
>operate simultaneously. Of course any drift component always averages
>out to zero, since there is no force pushing it in any direction. But of
>course, also, no single case is average, and an allele subject only to
>drift has only two possible fates: extinction or fixation, with the
>relative probabilitie of each at any moment equal to 1-p and p,
>respectively, where p is the current frequency.
>
I disagree. The transmission of alleles from one generation to the
next is inherently a stochastic process based on probability
distributions. Whether an individual survives or not, whether it
reproduces or not, whether the offspring survive or not depends both
on phenotype (and genotype) and on plain dumb luck. Which of the
parental alleles get passed on to a particular gamete and which gamete
succeeds in fertilization depends on plain dumb luck (possibly with
some contribution of the genotype). The point is that the
probabilities are not uniform -- some alleles end up with a higher
probability of appearing in the next generation than others. You can
separate drift from selection conceptually by saying that the
deviation of probabilities from a uniform distribution is called
"selection" while the fact that there are probabilities present at all
is called "drift". But the biological processes at work are all
stochastic, probabilistic, with non-uniform probabilities.
So I amend my previous definition of drift where you misunderstand the
meaning of correlation and environment. How is this?
The transmission of alleles from generation to generation is a
stochastic process, depending on random factors expressed through
probability distributions. Therefore allele frequencies will behave
as a sort of "random walk" process, changing over generations. The
probability distribution need not be uniform, all alleles equally
likely to appear in the next generation. The deterministic change in
the frequency distribution (mean, variance, etc.) from generation to
generation due to the unequal probabilities (unequal "fitness") is
called "selection". The remaining variation in frequency
distribution, including that occurring even when the probability
distribution is uniform, is called "drift".
John Harshman
> On Mon, 05 Dec 2005 02:16:16 GMT, John Harshman
> <jharshman....@pacbell.net> wrote:
>
> <snip>
>
>>Selection and drift are two different processes that can
>>operate simultaneously. Of course any drift component always averages
>>out to zero, since there is no force pushing it in any direction. But of
>>course, also, no single case is average, and an allele subject only to
>>drift has only two possible fates: extinction or fixation, with the
>>relative probabilitie of each at any moment equal to 1-p and p,
>>respectively, where p is the current frequency.
>>
>
>
> I disagree.
With what?
> The transmission of alleles from one generation to the
> next is inherently a stochastic process based on probability
> distributions. Whether an individual survives or not, whether it
> reproduces or not, whether the offspring survive or not depends both
> on phenotype (and genotype) and on plain dumb luck. Which of the
> parental alleles get passed on to a particular gamete and which gamete
> succeeds in fertilization depends on plain dumb luck (possibly with
> some contribution of the genotype). The point is that the
> probabilities are not uniform -- some alleles end up with a higher
> probability of appearing in the next generation than others. You can
> separate drift from selection conceptually by saying that the
> deviation of probabilities from a uniform distribution is called
> "selection" while the fact that there are probabilities present at all
> is called "drift". But the biological processes at work are all
> stochastic, probabilistic, with non-uniform probabilities.
I wouldn't put it this way. Selection is not stochastic. The stochastic
component in evolution is drift. (There are several potential
non-stochastic components that aren't selection, e.g. meiotic drive).
That's certainly the way it's treated mathematically, and selection is
after all a mathematical concept, an abstraction from the actual
behavior of alleles. I think it's useful to separate selection and drift
in this way, too. Obviously, there is no real system in which pure
selection exists without drift, i.e. without stochastic variation. As
you learn in elementary population genetics, only an infinite, panmictic
population has no drift.
> So I amend my previous definition of drift where you misunderstand the
> meaning of correlation and environment. How is this?
>
> The transmission of alleles from generation to generation is a
> stochastic process, depending on random factors expressed through
> probability distributions. Therefore allele frequencies will behave
> as a sort of "random walk" process, changing over generations. The
> probability distribution need not be uniform, all alleles equally
> likely to appear in the next generation. The deterministic change in
> the frequency distribution (mean, variance, etc.) from generation to
> generation due to the unequal probabilities (unequal "fitness") is
> called "selection". The remaining variation in frequency
> distribution, including that occurring even when the probability
> distribution is uniform, is called "drift".
I don't find it to be any different from mine, except that you define
selection as encompassing all deterministic factors, whereas I and
population geneticists do not.
r norman
<jharshman....@pacbell.net> wrote:
>>
>> <snip>
>>
>>>Selection and drift are two different processes that can
>>>operate simultaneously.
<snip>
>> I disagree.
>
>With what?
With what I didn't snip out. I claim that selection and drift can be
separated conceptually (and mathematically) but cannot be separated in
terms of mechanism.
We certainly agree on the important results but disagree on
interpretation and perspective.
The mechanism of selection, that is the survival and reproduction of
individuals depending on genotype (as expressed through phenotype) is
inherently stochastic. That is, you cannot separate out the drift.
Look at the genes that "produce" deleterious effects like breast
cancer or high blood pressure or whatever. These are not expressed
absolutely, only in a certain percentage of cases. That is, the
chance of getting breast cancer or high blood pressure or whatever is
significantly higher if you carry certain alleles. No, these are not
the best examples -- the real harm of these occur after the major
reproductive age is past. Still, there are a lot of genes that only
produce a tendency to be expressed. Even factors that are definitely
expressed may play a role in enhancing or limiting reproduction only
in exceptional circumstances -- enough to produce a statistically
significant effect but not to affect any specific single individual.
If "fitness" is only expressed in probabilistic terms, how can you say
that selection is not stochastic?
You are right, though -- I should not attempt to define "drift"
because I am focusing on only one aspect of things. I'll leave the
evolutionary biologists to the task. Still, I insist (OK, that is
pretty strong -- I strongly suggest) that all the underlying
mechanisms at work are inherently stochastic in nature. There are
some instances of deterministic selection -- lethal mutations for
example, or governmental restrictions on reproduction -- but these are
exceptions.
John Harshman
> On Mon, 05 Dec 2005 18:40:20 GMT, John Harshman
> <jharshman....@pacbell.net> wrote:
>
>
>>> <snip>
>>>
>>>>Selection and drift are two different processes that can
>>>>operate simultaneously.
>
>
> <snip>
>
>>>I disagree.
>>
>>With what?
>
>
> With what I didn't snip out. I claim that selection and drift can be
> separated conceptually (and mathematically) but cannot be separated in
> terms of mechanism.
If by mechanism you just mean differential reproduction, then I would
agree. Selection is a theoretical abstraction in the same sense that a
mean is a theoretical abstraction. Shit happens, and averaged out over a
large number of events, we call that selection or drift, or some of each.
Whoa, that's a whole nother matter; what you're talking about (in
general, not in the specific examples you give) is incomplete
penetrance. That's neither drift nor selection. I suppose you could
consider it another part of the stochastic component of evolution -- it
increases variance of result.
> That is, you cannot separate out the drift.
> Look at the genes that "produce" deleterious effects like breast
> cancer or high blood pressure or whatever. These are not expressed
> absolutely, only in a certain percentage of cases. That is, the
> chance of getting breast cancer or high blood pressure or whatever is
> significantly higher if you carry certain alleles. No, these are not
> the best examples -- the real harm of these occur after the major
> reproductive age is past. Still, there are a lot of genes that only
> produce a tendency to be expressed. Even factors that are definitely
> expressed may play a role in enhancing or limiting reproduction only
> in exceptional circumstances -- enough to produce a statistically
> significant effect but not to affect any specific single individual.
> If "fitness" is only expressed in probabilistic terms, how can you say
> that selection is not stochastic?
Because selection is commonly understood as the non-stochastic component
(and only that component) of an inherently mixed process. I think this
same sort of thing is true of any process with a stochastic component.
We commonly separate all sorts of results into "true value", or "mean"
or whatever and "error" or "variance" when actually there is only a
bunch of results with a mixture of causes. Why? Because it's highly useful.
> You are right, though -- I should not attempt to define "drift"
> because I am focusing on only one aspect of things. I'll leave the
> evolutionary biologists to the task. Still, I insist (OK, that is
> pretty strong -- I strongly suggest) that all the underlying
> mechanisms at work are inherently stochastic in nature.
Sure they are. Selection is not, oddly, an underlying mechanism. It's
merely the sum of all manner of events with their own causations. These
events have stochastic components out the wazoo. But we find it very,
very useful to separate this sum of events into stochastic and
non-stochastic causes, one of the latter being selection.
> There are
> some instances of deterministic selection -- lethal mutations for
> example, or governmental restrictions on reproduction -- but these are
> exceptions.
There's plenty of room for weaseling even with those. That's all
irrelevant, though.
r norman
<jharshman....@pacbell.net> wrote:
OK. I feel better now. Generally, selection is taught as one of the
underlying mechanisms of evolution, distinct from drift. Given that
both are abstractions of the way biology works, that should be a
perfectly way of describing things.
Nic
> > Do you mean there could be between two and twenty clades with no
> > common ancestor?
>
> Yes. The concept of clades at the species level work fine for sexually
> mating species, whether they do meiosis, or they form cliches that
> undergo horizontal gene transfer (HGT) within a cliche but not outside
> that cliche. However during the very early days HGT might have involved
> very fuzzy clades, to where all life whatsoever was a single "ring
> species" with respect to HGT. It wouldn't be at all meaningful to
> ignore HGT and consider only whole-cell cladistics, especially if more
> than 50% of the genome of any individual came from HGT. But it wouldn't
> be quite honest to claim Universal Common Ancestry merely because all
> early life formed a single "ring species". So it may be best to cease
> using whole-cell cladistic analysis earlier than some ancient epoch,
> whenever we determine that HGT dominated over cellular reproduction
> earlier than that time. So we'd do our cladistic analysis back to that
> epoch, and thereby have several disjoint clades from that epoch forward
> to the present, but before that time we'd organize life using a more
> appropriate system instead of whole-cell cladistics. So the species
> that were distinct at that transition epoch would be common ancestors
> of their respective clades to the present, and that number might be as
> many as twenty by my complete guestimate. So the claim of whole-cell
> common ancestry, which is part of evolutionary theory, would go back
> only to that transition epoch.
I see what you mean. But what if cell membranes had a single common
ancestor (which they probably don't)? Wouldn't that be a good
candidate for a single species bottleneck? I mean, any HGT mechanisms
that worked prior to the first membrane probably wouldn't work for the
new membraned life forms. Instead they would need a modified version
of their own which could get through membranes. This would in effect
create an exclusive gene pool.
Nic
an...@sci.sci
Now that the haplotype map is out, I have a question: What are the
statistics of haplotype blocks in regard to neutral vs. non-neutral
evolution? I.e. which haplotype blocks contain at least one selected
gene within it, forcing the entire block to evolve in accord with that
gene, and which haplotype blocks don't contain any selected gene at
all, allowing the entire block to drift randomly without any
constraint? What are the statistics of these two kinds of blocks? For
example, are the lengths of blocks statistically the same between the
two kinds of blocks, and then in total 80% of the blocks are devoid of
any active selection site while all the active selection sites are
concentrated within the remaining 20%? So then 80% of sites are within
those 80% neutral blocks while 20% of sites are within the 20% actively
selected blocks?
.
an...@sci.sci
> next is inherently a stochastic process based on probability
> distributions. Whether an individual survives or not, whether it
> reproduces or not, whether the offspring survive or not depends both
> on phenotype (and genotype) and on plain dumb luck. Which of the
> parental alleles get passed on to a particular gamete and which gamete
> succeeds in fertilization depends on plain dumb luck (possibly with
> some contribution of the genotype). The point is that the
> probabilities are not uniform -- some alleles end up with a higher
> probability of appearing in the next generation than others. You can
> separate drift from selection conceptually by saying that the
> deviation of probabilities from a uniform distribution is called
> "selection" while the fact that there are probabilities present at all
> is called "drift". But the biological processes at work are all
> stochastic, probabilistic, with non-uniform probabilities.
I agree, except for the wording of the math. Let me try to fix the math
wording: Given a starting distribution, which is something like a
normal distribution, and letting the population evolve for a while,
there is now an ending distribution, which is also something like a
normal distribution. Subtracting the two, we have the change in
distribution, which again is something like a normal distribution. The
mean of this change distribution is the amount of forced selection,
while the variance of this change distribution is the amount of random
drift. But that's not quite right, because both variance of the
original distribution, and variance of the actual per-individual
change, are conflated by that calculation. You don't know whether
perhaps the change is entirely deterministic, and all variance you see
in the final result was entirely due to variance already in the
starting distribution. You need to somehow factor out the original
distribution, to see how with a given initial state it may change over
time with both change-in-mean and additional variance. I suppose you'd
need to do factor analysis. That is you do linear regression
(least-squares fit to a line) to compute how much the final result
depends on the initial state, as one factor, and subtract that mean
from each individual's final state, then you have two remaining
parameters in the residual of the final state, mean change not
accounted for by input state, and "noise" which is all remaining
unexplained variance not accounted for by either input state or mean
change. Any stat experts here: Does that sound like the right math?
> The transmission of alleles from generation to generation is a
> stochastic process, depending on random factors expressed through
> probability distributions. Therefore allele frequencies will behave
> as a sort of "random walk" process, changing over generations. The
> probability distribution need not be uniform, all alleles equally
> likely to appear in the next generation. The deterministic change in
> the frequency distribution (mean, variance, etc.) from generation to
> generation due to the unequal probabilities (unequal "fitness") is
> called "selection". The remaining variation in frequency
> distribution, including that occurring even when the probability
> distribution is uniform, is called "drift".
Yes, again with the math a bit sloppy, but not quite as bad as before,
so usable as a preliminary explanation, to be followed by the better
math explanation.
Does anybody feel like running a simulation, with hypothetical
selection mechanism? To keep it simple, assume every individual cell
divides to produce two daughter cells, but then there's less than
certainty of each daughter surviving to next replication point, with
probability of survival based on genotype and resource saturation, and
random numbers used to make each actual live/die decision. Use a
totally abstract life/die probability function of the genome. To avoid
the chance of false result due to *all* individuals dying due to lack
of food despite the death of others freeing up food for each, use a
randomized sequence of individuals facing life/death choice, and
compute remaining food for remaining individuals *after* earlier
individuals have already been decided whether they live or die.
Or instead of a totally abstract live/die probability function, use an
actual cellular-automaton simulation, with individuals wandering around
by various genotype-specified strategies collecting food as they go, or
not getting food if no new food has appeared since the last visitor to
that cell, and dying if not getting enough food within required time,
and replicating whenever sufficient food has been stockpiled by eating.
In either case, remember the distribution of genotypes at the start,
and keep track of the entire clade from each individual, and remember
the distribution of genotypes after various time periods, and compute
statistics from start to each other epoch as I stated earlier above.
Anybody feel like doing that? (Or has it already been done and somebody
knows where the report is on the Web?)
.
an...@sci.sci
> mathematically) but cannot be separated in terms of mechanism.
I agree. See for example what I posted a few minutes ago, where factor
analysis is used to pull out the portion of final genome distribution
accounted-for by initial genome distribution, then whatever remains is
a mean (for which "selection pressure" is assigned credit) and an
unaccounted-for variance (for which "random drift" is assigned credit).
Anyway, per my original start-of-thread, are we all agreeing that
mathematically we can define two observable factors of Darwinian
evolution, one that adapts to circumstances to approach optimal fitness
as best as possible given the available variation in the genotype pool,
and one what wanders all over like a random walk, and that some sites
seem to have no fitness adaption pressure whatsoever, purely random
walk, while others have fitness adaption pressure, which tends to rein
in the random walk? (I invented that term "fitness adaption pressure"
just now, to replace the misnomer "selection pressure" that has been in
use until now. Anybody like the replacement better than the original?)
So anyway, after we outline the evidence (three kinds of trees:
(e1) descriptive classification hierarchy, including embryology studies
and biochemical pathway studies,
(e2) fossil chains for extinct species and actual live geneologies for
current species,
(e3) DNA cladograms; which all agree with each other far beyond what
would be random chance coincidencal agreement);
when we then state the underlying theory to explain such amazing
agreement of evidence, do we agree on the specific parts of the theory,
namely:
(t1) replication,
(t2) mutation (new variation),
(t3) stochastic mechanism for selecting who begats and who dies without
begatting which can be mathematically factored into:
(t3a) "fitness adaption pressure",
(t3b) "random drift"?
.
John Harshman
>>Most sites in the human genome are evolving neutrally.
>
> Now that the haplotype map is out, I have a question: What are the
> statistics of haplotype blocks in regard to neutral vs. non-neutral
> evolution?
I don't know what you mean by "haplotype block".
> I.e. which haplotype blocks contain at least one selected
> gene within it, forcing the entire block to evolve in accord with that
> gene, and which haplotype blocks don't contain any selected gene at
> all, allowing the entire block to drift randomly without any
> constraint?
Ah, here you are apparently talking about frequency of recombination. Is
"haplotype block" some kind of synonym for "linkage group"? I really
don't know about recombination frequency through the human genome. But
the presence of one site under selection does not alter the question of
whether even an adjacent site is evolving neutrally. It just means that
if there is a selective sweep, variation at those neutral sites will be
reduced.
> What are the statistics of these two kinds of blocks? For
> example, are the lengths of blocks statistically the same between the
> two kinds of blocks, and then in total 80% of the blocks are devoid of
> any active selection site while all the active selection sites are
> concentrated within the remaining 20%? So then 80% of sites are within
> those 80% neutral blocks while 20% of sites are within the 20% actively
> selected blocks?
There is, to my knowledge, no such block structure as you imply here.
There are regions in which recombination is more or less frequent, but I
wouldn't say that talking about "blocks" does more than confuse what
structure there is.
John Harshman
>>I claim that selection and drift can be separated conceptually (and
>>mathematically) but cannot be separated in terms of mechanism.
>
>
> I agree. See for example what I posted a few minutes ago, where factor
> analysis is used to pull out the portion of final genome distribution
> accounted-for by initial genome distribution, then whatever remains is
> a mean (for which "selection pressure" is assigned credit) and an
> unaccounted-for variance (for which "random drift" is assigned credit).
>
> Anyway, per my original start-of-thread, are we all agreeing that
> mathematically we can define two observable factors of Darwinian
> evolution, one that adapts to circumstances to approach optimal fitness
> as best as possible given the available variation in the genotype pool,
> and one what wanders all over like a random walk, and that some sites
> seem to have no fitness adaption pressure whatsoever, purely random
> walk, while others have fitness adaption pressure, which tends to rein
> in the random walk?
Yes.
> (I invented that term "fitness adaption pressure"
> just now, to replace the misnomer "selection pressure" that has been in
> use until now. Anybody like the replacement better than the original?)
No.
> So anyway, after we outline the evidence
Evidence for what, precisely? This seems like evidence for common descent.
> (three kinds of trees:
> (e1) descriptive classification hierarchy, including embryology studies
> and biochemical pathway studies,
> (e2) fossil chains for extinct species and actual live geneologies for
> current species,
> (e3) DNA cladograms; which all agree with each other far beyond what
> would be random chance coincidencal agreement);
I think this is confused. Particularly, what do you mean by "fossil
chains"? In reality, what we have here is a single nested hierarchy,
formed by data from various different sources. That's all.
> when we then state the underlying theory to explain such amazing
> agreement of evidence, do we agree on the specific parts of the theory,
> namely:
> (t1) replication,
> (t2) mutation (new variation),
> (t3) stochastic mechanism for selecting who begats and who dies without
> begatting which can be mathematically factored into:
> (t3a) "fitness adaption pressure",
> (t3b) "random drift"?
> .
Those are, more or less, some of the parts of a particular theory of
evolution. I think the first two are necessary for any evolutionary
theory. The third is specific to some subset of theories. None of these
seem to have all that much to do with e1-e3.
r norman
You completely misunderstand the probability distribution I refer to.
I am speaking of the probability that an individual possessing allele
An (n=1, 2, ... N, the number of alternatives) will contribute a copy
of that allele to the next generation. You can make more complex
models with gene combinations and interactions and whatever you like.
If the probability is uniformly distributed, that is, does not depend
on An, then all change in gene frequencies must be due to drift. When
the distribution is not uniform -- some alleles contribute more to the
next generation than others, they have greater fitness -- then part of
the change in gene frequencies is attributed to the non-uniformity of
the probabilities and part do to the stochastic nature of the process.
I know that you can separate the two factors, selection and drift,
conceptually. I am simply arguing that the underlying mechanism,
only changes probabilities which inherently confound the two notions.
<snip more of a similar nature>
>Does anybody feel like running a simulation, with hypothetical
>selection mechanism? To keep it simple, assume
<snip details of model>
There are many people who run simulations with all sorts of different
underlying assumptions. Often, some of the critically important
assumptions go unnoticed by the person designing the simulation but
are implicit in the details of the implementation. The problem is
that, unless you manage the biology correctly, the results are not at
all useful.
Computer simulations can be extremely useful. But they are not easy to
do correctly.
an...@sci.sci
> But what if cell membranes had a single common
> ancestor (which they probably don't)? Wouldn't that be a good
> candidate for a single species bottleneck? I mean, any HGT mechanisms
> that worked prior to the first membrane probably wouldn't work for the
> new membraned life forms. Instead they would need a modified version
> of their own which could get through membranes. This would in effect
> create an exclusive gene pool.
AFAIK we have no idea how the original cell membrane/wall came to be.
For all we know, there might have been an arms race, where at first
there was no membrane/wall and HGT was totally promiscuous, then as
some clumps of replicators developed a wall/membrane to protect
themselves from invasion by foreign genome, the invasive genome
developed ways to penetrate the defense, provoking another round of
stronger defense, etc. back and forth, to where nowadays we continue to
have such an arms race in progress in prokaryotes, whereby some strains
of bateria work hard to prevent all invasion while others allow
invasion by appropriate suitors.
Like most things in biological evolution, there was no "first" of
anything, there was a gradual building up of almost-something until at
some point we can say it finally had achieved close enough to what we
know today that we call it "modern", as with "modern Homo sapiens"
which was a gradual transition from "archaic Homo sapiens" which was a
gradual transition from Home erectus, no such thing as "first humans"
as in Adam/Eve mythology. There was no first cell wall. There wasn't
any first life, there was a gradual transition from non-life through
semi-life until at various points various people would say it's fully
life.
Back to the HGT/membrane/wall arms race: It's possible that at some
point one strain of bacteria developed a really good HGT mechanism,
whereby it could infect all the cells in the whole ocean, causing death
to all cells with incompatible genomes, but unifying all cells that had
compatible genomes, whereby there'd be a single strain of bacteria
worldwide for a while. Or such a "superpower" strain of bacteria might
have been limited to a range of temperature and salinity etc., so that
it couldn't invade the hyperthermophiles etc., so that the extreme*
remained separately evolving without any contamination from the
"superpower".
.
an...@sci.sci
Any single contiguous segment of DNA that appears to have evolved as a
single unit for hundreds of thousands of years according to the
preliminary haplotype-map summary report that I saw in _Science_ a few
weeks ago. Meiosis seems to have mixed-and-matched among the different
blocks, yet seems to have kept each single block intact for all that
time. At least that's the impression I got from reading the report.
I had always thought that DNA can break just about anywhere, that
some places it breaks slightly more often than other places, but in
general there's only a small variation between hot spots and cool spots
in regard to breaking and crossing during meiosis. But the report seems
to say that old idea was wrong, that there's such a drastic difference
between hot and cool spots that to a good approximation the chromosomes
are strictly divided into fixed blocks that act as units during meiosis.
SNPs can be tracked through the generations, to detect ancestry along
mixed male/female lines of descent, but because any one SNP has only
four possible states, the same state recurs at random many times, so if
you see the same exact rare SNP in two distant populations you can't be
sure whether there was recent gene flow between them or the same SNP
recurred by chance in both places.
But any single haplotype block has *many* different states, and each
evolutionary line accumulates occasional mutations which don't recur in
the same exact pattern along different evolutionary paths, so seeing
the same identical haplotype block in two remote populations definitely
shows recent gene flow between them, more recent than the last mutation
in that particular block within either population.
So if that's true, haplotype blocks offer a really powerful way to
detect tiny bits of gene flow between distant populations, such as if a
single explorer wandered from one to the other and left one or two
descendents who then left more descendents such that after hundreds or
thoudsands of generations a few of the haplotype blocks from that
invidiual explorer might remain in the population.
(And if two very similar haplotype blocks are found, it indicates
semi-recent gene flow, more recent than the bulk of mutations, but
older than the very latest mutations after the gene-flow occurred.)
The haplotype tool is much more powerful than the mitochondrial or Y
method, because a lot of different haplotype blocks are passed from a
single individual, instead of just one set of intact mitochondria or
one Y chromosome, and then these blocks are split from each other as
they are passed to subsequent generations, so an entire population with
a single common ancestor is likely to each have a few blocks from that
one ancestor mixed in with blocks from the many other ancestors equally
far back in generations. In theory, if we knew the genomes of *every*
individual at a certain time, then by matching haplotype blocks many
generations later we can determine exactly which modern individuals are
descended from exactly which ancestors. But more likely the best we can
do is measure the current genome of all living individuals in each
local population that is tightly inbred, and extrapolate back to the
ancestors who had the same set of haplotype blocks (modulo mutations
since then), and then detect any discrepancies from this neat model
which would indicate any gene flow from one such tight group to/from
another. If we could get DNA from several Neandertals, and thereby
deduce what haplotype blocks they had, we might in fact be able to
detect whether any modern humans have even one such
distinctly-Neandertal haplotype block.
> Ah, here you are apparently talking about frequency of recombination.
> Is "haplotype block" some kind of synonym for "linkage group"?
Checking online definitions:
<http://dictionary.reference.com/search?q=linkage%20group&db=*>
A pair or set of genes on a chromosome that tend to be
transmitted together.
Yes, that's part of what I mean, except I mean they are transmitted
together virtually without exception for tens or hundreds of thousands
of years, not just that they tend to be transmitted together slightly
more often than at randon over a single generation.
: a set of genes at different loci on the same chromosome that except
for crossing-over tend to act as a single pair of genes in meiosis
instead of undergoing independent assortment
No, I mean that *in*spite* of crossing-over they *nevertheless* remain
linked, i.e. crossing-over simply doesn't "ever" happen (not in a
hundred thousand years) between them.
n : any pair of genes that tend to be transmitted together; "the genes
of Drosophila fall into four linkage groups" [syn: linked genes]
No, not at all. That definition simply means they are on the same
chromosome (Drosophila has four chromosomes, as we all know from
high-school biology), so there is *some* (possibly very weak) linkage
between them, that they are not *totally* independent of each other.
<http://www.biochem.northwestern.edu/holmgren/Glossary/Definitions/Def-L/linkage_group.html>
A group of gene loci known to be linked; a chromosome. There are as
many linkage groups as there are homologous pairs of chromosomes.
No! That's most definitely not what I mean here. (Note that "syntenic"
means the same thing, but is unambiguous, applies to *any* two loci on
the same chromosome, regardless of whether the linkage is strong enough
to detect by standard tests, presumably Mendel-type breeding
experiments.)
<http://opbs.okstate.edu/~melcher/MG/MGW1/MG111122.html>
Ah, a fine Web site with lots of quotes to mine, such as:
* Markers that have measurable recombination frequencies are said to
be linked.
* Markers related through a chain of linkage constitute a linkage
group.
Ah, so a "linkage group" is a maximal connected set of markers (loci)
via the pairwise "linked" property (presumably meaning not easily
separable by recombination, which is easier for me to understand than
the verbage in the definition quoted). In most cases that would imply
that such a maximal connected set is a whole chromosome. It's roughly
analagous to a "ring species" which is a maximal connected set via
pairwise ability-to-mate, or maximal-connected-graph in any other model
that has binary links between nodes.
* As more and more markers are studied, linkage groups become larger
(and the number of groups smaller) until the number of linkage
groups equals the number of chromosomes.
So technically, "linkage group" is not globally defined for a species,
rather it's defined contingent on a given set of markers under study.
(SIde remark: When two ancestral chimp/human chromosomes joined
end-to-end to yield a single human chromosome, two "linkage groups"
were merged to form just one. In particular, genes that were near the
joining ends of the original chromosomes, suddenly changed from being
totally unlinked to being very closely linked. I wonder how such new
linkage affected subsequent evolution of those loci? In particular, I
wonder whether those two joined ends are now parts of a single
haplotype block or parts of two adjacent haplotype blocks?)
<http://66.102.7.104/search?q=cache:APrIlzHRp6wJ:hal.weihenstephan.de/genglos/asp/genreq.asp%3Fnr%3D519+linkage+disequilibrium&hl=en&ie=UTF-8>
Nice definition:
Linkage - An association in inheritance between characters such that the
parental character combinations appear among the progeny more often
than the non-parental.
Note that any linkage, no matter how weak, qualifies for that definition.
Connected in that way is not at all what I mean by "haplotype block"
wherein all characters (or loci or markers) are virtually 100% linked
over many generations.
And for fun, slightly side topic, see this attempt to clear confusion
regarding "linkage" vs. "linkage disequilibrium":
<http://linkage.rockefeller.edu/wli/lld.html>
Due to the ambiguity of definition, and the fact that the centroid of
the various definitions simply defines linkage group as the contents of
a chromosome, which is *not* what I mean, I think I'll avoid using
that standard term when I mean the more specific kind of longterm
linkage discovered by the haplotype-map project, where there are
several different such blocks on each single chromosome.
I'm going to cut my reply here, and respond to later remarks in another
posting.
.
an...@sci.sci
> genome.
Did you see the article in _Science_ summarizing the results from the
human haplotype-map project (based on Oct. 27 full report in _Nature_)?
I'm not sure when exactly the _Science_ summary appeared, but probably
close to the _Nature_ report.
<http://groups.google.com/group/net-gold/msg/65477f6396258f93>
It has been recently realized that in humans, most such swapping
occurs primarily at a limited number of "hotspots" in the genome.
By analyzing the HapMap data, the researchers have produced a genome-wide
inventory of where recombination takes place.
> But the presence of one site under selection does not alter the
> question of whether even an adjacent site is evolving neutrally.
I disagree. For any new mutation, it occurs on one haplotype-block
allelle or another, among the several variants (allelles) of that
particular haplotype-block. From that point onward, it's linked to all
the other SNPs within that block-allelle, separate from any SNPs within
other allelles of the same block. If natural selection is still working
on that particular block, whereby different allelles will be
retained/killed differently, because of any single selectable gene
elsewhere on the block, then even if the new mutation is itself
neutral, it'll be dragged with its mates on the same block, toward
extinction or fixation, rather than drifting neutrally. Only if an
entire haplotype block contains not a single non-neutral site, do all
allelles of that block compete on a neutral basis, drifting neutrally
as a unit, not dragged in one direction or another.
Furthermore if there are two or more allelles of a block, with *every*
site on *every* allelle that block totally neutral, so that the block
is drifting neutrally, but then a new mutation in one allelle causes a
selectable variation between one allelle and the others, all the
neutral sites on that one block-allelle will be dragged one way or the
other depending on whether the mutation is beneficial or harmful (and
the other allelles will go the other way, dragging all their neutral
SNPs that other way). Only when that one mutated allelle has gone
extinct, or it's become fixed by driving all the other allelles to
extinction, will that block revert to neutral drift again.
And even while any such block is drifting neutrally (because no allelle
of it has anything selectable), all the neutral SNPs on each allelle
are drifting as a single unit, not drifting individually.
So in any case there's a fundamental difference between a neutral site
drifting all by itself, and a neutral site linked to other sites in a
block so they drift or select as a unit.
> It just means that if there is a selective sweep, variation at those
> neutral sites will be reduced.
If you look only at each site by itself, this is what you observe. But
that's not the whole picture, of linked neutral drift whenever there
isn't linked selection. Looking simultaneously at two sites on the same
block, you see a very different picture from looking at two sites on
different blocks. An analogy from quantum mechanics may be of value.
When two spins are entangled, each spin by itself seems truly random,
but if you observe the two spins together you see the entanglement,
which is decidely non-random. You see ud or du, each with 50%
probability. but you never see uu or dd. If they weren't entangled,
then all four combinations would be equally (25%) likely. Likewise if
you deliberately select populations with haplotype-block allelles such
that one particular site is exactly 50% each single-site allelle,
you're going to have a hard time arranging that some other site on the
same block is 50% each way independent of the first site.
> There is, to my knowledge, no such block structure as you imply here.
> There are regions in which recombination is more or less frequent, but I
> wouldn't say that talking about "blocks" does more than confuse what
> structure there is.
I really got a different impression from browsing the summary report in
_Science_, that there's such a small number of hot spots, and such a
small amount of recombination anywhere else, that the intervals between
adjacent hot spots really do act like significantly-large indivisible
blocks over many generations lasting tens or hundreds of thousands of
years, such that many modern blocks are still intact from the major
miagrations out of Africa, whereby each such block provides a
non-recombinant hence pure-tree archive of mutations from then until
now. These many pure-trees are in addition to the two pure-trees we
already had from mitochondrial DNA and Y chromosomes, allowing us now
to trace many additinal trees-of-descent in addition to the
pure-female-female and pure-male-male trees.
Maybe one of the better experts in this group read the _Science_
article with more detail and can say better where the truth lies
somewhere between my personal impression and the other view (which I
myself held prior to seeing the HapMap summary article)?
.
an...@sci.sci
> > just now, to replace the misnomer "selection pressure" that has been in
> > use until now. Anybody like the replacement better than the original?)
> No.
Hey, what gives you the right to speak for *everyone*?
How about just saying "Not I" and speak only for yourself?
If all of the Little Red Hen's neighbors all say "Not I", then she
really knows nobody likes her idea, but none of them have the right to
speak for the others, right?
Anyway, speaking only for yourself, do you have any third term to
nominate to replace the original and confusing "selection pressure" in
lieu of my proposed replacement "fitness adaption pressure"?
Maybe something like "selection-drift bias" perhaps?
Or "fitness-hill-climbing gradient"?
> > So anyway, after we outline the evidence
> Evidence for what, precisely? This seems like evidence for common descent.
No no no, don't pre-judge what the evidence is "for". Just collect
evidence like any good police detective, and then after you have
all the evidence, see where it leads. For example, if you pre-judge
that the husband probably killed his wife, you'll be so busy collecting
evidence to "prove" he did it that you not bother to ask neighbors if
they saw any one-armed men around the neighborhood, and as a result
you'll achieve a false conviction Dr. Richard Kimble.
So the evidence is characteristics of various fossils, and their
location and date buried. Upon analysis, we determine that common
descent is the most likely (parsimonious and reasonable) explanation
for the evidence, and also anti-explanation of the many kinds of
evidence we never see (such as chimeras of parts from widely separated
times and places). But we come up with common descent only after
collecting and examining and analyzing the evidence. The evidence is
*for* determining where the species came from, not *for* proving a
pre-conceived idea of common descent. (In fact when such evidence was
originally collected and studied, the a priori belief was that
everything was specially created by a supernatural being during a
single week only a few thousand years ago, and the evidence was so
overwealmingly opposed to that belief that paleontologists were forced
to then think of some other theory to replace their old belief.)
In our case, if we pre-judge the case and collect evidence only in
support of common descent, the Creationists can legitimately argue that
we stacked the deck, collected only evidence in support and ignored
evidence against, so our conclusion is bogus. We don't want that. We
want to present honest evidence, and *then* show that common descent is
the only reasonable theory to explain it. We want to invite the
Creationists or anybody else to propose any other theory which equally
well might explain the evidence and anti-explain the non-evidence.
(Goddidit isn't a good theory, because it explains both the evidence
and anything else whatsoever that godmightdo. It fails to anti-explain
the non-evidence, to explain why cross-time cross-space chimeras don't
occur. On the other hand, Omphalos, whereupon a supernatural being
deliberately planted evidence to make it *appear* as if common descent
occurred, is a valid philosophical thoery, scientifically
indistinguishable from ordinary common descent.)
> what do you mean by "fossil chains"?
Fossil species A B and C appear successively within approximately the
same geographical region, and each appears to be very similar to its
chronological neighbors, but clearly the three species are not
identical to each other. How to explain that? Perhaps they form a chain
of ancestry? By comparison we do *not* see species A B and C in such
chronological sequence but each in a totally separate geographic area,
nor three such species so very similar yet separated by gaps of
hundreds of millions of years, at least hardly ever. Often we see ten
or more such species in sequence, spaced over time, with small
progressive change from each to the next, but rather large end-to-end
total accumulated change. The only *other* species we ever see so very
similar to any of these, are at what appear to be species-split
branch-points. Those side-branches sometimes go extinct promptly, but
sometimes start their own "fossil chains" which then run in parallel
with the one we were already looking at.
> In reality, what we have here is a single nested hierarchy,
> formed by data from various different sources. That's all.
When we fold together all the fossil chains, with the branch points
that replace one ancestral fossil chain with two or more new fossil
chains, we have a forest of trees. These trees are consistent with a
single tree where lots of links are missing from our evidence so-far.
We can state strongly that there are fossil chains, and branch points,
yielding lots of small common-ancestry trees. We can guess with less
assurance that they all belong to one universal common-ancestry tree.
We can state with intermediate assurance that we have a single
common-ancestry tree for each phylum, and sometimes for several phyla
together, from the Cambrian explosion to the present. The case for
joining all those trees into one universal tree before the Cambrian
explosion is not firmly established by fossil evidence alone.
This is where we need the other two lines of evidence (comparative
embryology, and DNA cladograms) to piece together the separate trees
into a single universal tree. And last I heard the best we could do is
establish three trees for the three major domains, with unknown linkage
earlier than that.
> > when we then state the underlying theory to explain such amazing
> > agreement of evidence, do we agree on the specific parts of the theory,
> > namely:
> > (t1) replication,
> > (t2) mutation (new variation),
> > (t3) stochastic mechanism for selecting who begats and who dies without
> > begatting, which can be mathematically factored into:
> > (t3a) "fitness adaption pressure",
> > (t3b) "random drift"?
> Those are, more or less, some of the parts of a particular theory of
> evolution. I think the first two are necessary for any evolutionary
> theory.
Replication is accepted by everyone, even YECs.
The source of new mutation is a point of debate between Darwinists and
some IDIots. Other IDiots reject the whole idea of mutation, preferring
special creation of each new species, where any "mutation" is actually
deliberate re-thinking of the design within the mind of the Designer.
Other IDiots compromise between those two extremes, allowing
micro-mutation to actually happen in live begatting-chains, but saying
that macro-mutation never happens, any species change is attributed to
installing a new Design, such as when MicroSoft starts shipping a new
major version of Windows rather than distributing patches to an old
version that convert it to a new minor-upgrade version.
Nit: A theory that denies physical evolution, which admits only
Design-evolution and shipping a new major version of the product, *is*
an evolutionary theory, since it involves evolution in the mind of the
Creator/Designer, but does *not* involve mutation in the physical
sense. So not all evolutionary theories require mutation as you claim.
But all *scientific* evolutionary theories do.
> The third is specific to some subset of theories.
Again, IDiots (actually more properlty ITiots) may claim some
supernatural Tinkerer (not Designer) is responsible for deciding which
of the random mutations to keep and which to discard, replacing natural
selection with deliberate breeding just as humans breed dogs.
But it's impossible for any physical process to be 100% correct, to
perform some task correctly 100% of the time, so no matter how accurate
a phyiscal process is, there is *some* stochastic element in the
process. So whereas a supernatural Intelligent Tinkerer might be
totally infallable in its breeding efforts, any natural process,
including humans-breeding-dogs, sexual selection, arms races, etc., all
have a stochastic element. Any *scientific* theory must restrict itself
to such natural processes, hence any *scientific* theory of selection
must be some form of stochastic selection. I don't think it's
mathematically possible for replication and mutation, with no selection
of any kind whatsoever, to yield the Origin of the Species, do you?
Accordingly *any* theory of evolution (to fit the evidence on Earth)
*must* involve some kind of selection, hence any *scientific* theory of
evolution must include stochastic selection.
Do you accept my argument to that point?
Then by simple math, we can separate the forcing toward a particular
result (the "selection pressure" or whatever we call it) from the
unexplained/random variance from that path (the "random drift").
> None of these seem to have all that much to do with e1-e3.
If you run simulations based on t1-3, assuming current theory about
polymeric DNA language and mapping to phenotypical characters, you get
evolutionary trees like e1, nested hierarchy in latest-generation
populations like e2, and DNA cladograms like e3, each agreeing with the
other two. Do you know any other local theory (some mechanism that
performs all action within local time and space, without any global or
time-travel coordination) that when simulated gives results likewise
matching all three lines of evidence?
It's not that t1-3 "have anything to do with" e1-3 in form, but that
if you simulate t1-3 you get e1-3 as consequence, so t1-3 explains
e1-3. It's like Kepler's orbits and Galileo's mechanics deal with conic
sections, while Newton's laws of gravity and motion deal with
differential equations, with no obvious similarity between the two. But
when you simulate Newton's laws with very small test masses (kilogram
size) under constant-gravity conditions you end up with Galileo's
mechanics, and when you simulate Newton's laws under zero external
gravity but very large test masses (planetary/stellar size) you end up
with Kepler's orbits, so Newton's theory explains both Kepler's and
Galileo's observables.
I want to clearly separate the evidence e1-3 from the explanatory
theory t1-3, partly because everyone except YECs agrees with the
evidence, so we can just stipulate that at the start of debates with
everyone else, but there's a lot of debate about the underlying theory
among IDiots etc. Most IDiots really do believe in evolution in the
narrow sense of e2 only (common ancestry), and even the YECs accept the
nested hierarchy e1, and the DNA cladograms can be shown live in the
laboratory so only a fool would dismiss them. But the IDiots don't
accept the Darwinian explanation as to how it works, *why* e1-3 turned
out that way. So our challenge should be to have them present any
alternate to our t1-3 to explain our agreed-upon e1-3 (and anti-explain
anti-e1-3, for example why don't we see each species using a totally
different hereditary mechanism, a different biochemistry, yet somehow
still containing the same basic units of organic chemicals so that food
webs work as they do, after all God could have designed life that way
if He chose, so why didn't he? Goddiddit fails to explain why such
differences are *not* present in life. IDiots gotta come up with
something better to anti-explain anti-e1-3.).
.
John Wilkins
>>>(I invented that term "fitness adaption pressure"
>>>just now, to replace the misnomer "selection pressure" that has been in
>>>use until now. Anybody like the replacement better than the original?)
>>
>>No.
>
>
> Hey, what gives you the right to speak for *everyone*?
> How about just saying "Not I" and speak only for yourself?
> If all of the Little Red Hen's neighbors all say "Not I", then she
> really knows nobody likes her idea, but none of them have the right to
> speak for the others, right?
>
> Anyway, speaking only for yourself, do you have any third term to
> nominate to replace the original and confusing "selection pressure" in
> lieu of my proposed replacement "fitness adaption pressure"?
> Maybe something like "selection-drift bias" perhaps?
> Or "fitness-hill-climbing gradient"?
Not sure who you're responding to, but "fitness-adaptation" is tautologically
pleonastically repeating the same thing twice over again. And selection and
drift are not polar opposites. The last offering works in some contexts.
...
--
John S. Wilkins, Postdoctoral Research Fellow, Biohumanities Project
University of Queensland - Blog: evolvethought.blogspot.com
Nihil tam absurdum quod non quidam Philosophi dixerit - adapted from Cicero
an...@sci.sci
> I am speaking of the probability that an individual possessing allele
> An (n=1, 2, ... N, the number of alternatives) will contribute a copy
> of that allele to the next generation.
If the cell is prokaryotic, or if eukaryotic and haploid and
reproducing via mitosis, then the probability is 1, regardless of
whether we're talking about symmetric cell division or big/large
budding. This probability of 1 applies to each and every daughter
thereby produced. The compound probability of at least one copy to
*any* of several daughters is of course also 1.
If we're dealing with sexual reproduction, involving two parents
mating, one male and one female, or equivalent, with both parents
diploid, each undergoing meiosis before yielding haploid sperm and egg
which fuse to yield diploid child, and the allelle is located on a
non-sexual chromosome, and we're considering this allelle independently
of all other sites on that chromosome or any other, the probability is
one half for each child thereby produced, (1 - (1/2)**N) compound
probability of at least one among N non-identical children.
If the allelle is on a sex chromosome, then the probability is 1 or 0
depending on the sex of the child, and compound probability is 1 or 0
depending on whether any child is of the correct sex.
If you're dealing with more than one gene at the same time, linkage
would affect the probability of various combinations of the genes.
What does this have to do with evolution, except for the fact that
replication occurs??
> If the probability is uniformly distributed, that is, does not depend
> on An, then all change in gene frequencies must be due to drift.
I don't see any relation between that sentence and the ones before.
What are you talking about here?? What probability is uniformly distributed??
What do you mean by "uniformly distributed" in this context??
> When <snip gibberish: the distribution is not uniform>
> -- some alleles contribute more to the next generation than others,
> they have greater fitness -- then part of the change in gene
> frequencies is attributed to the non-uniformity of the probabilities
> and part do [sic] to the stochastic nature of the process.
In general, at the moment when reproduction occurs, there is no
selection, any allelle that made it to that point is as good as any
other in being present at the very start of the next generation. A
moment later, *within* the lifetime of the next generation, *then* the
cell might promptly die if it has a bad genotype making a fatal
phenotype, or the cell might live for a while then die, or it might
live for a long time but never reproduce. But that's failure of the bad
allelle to keep the next generation alive, not failure of the bad
allelle to be passed at the moment of reproduction. (Occasionally
mitosis or meiosis or fission suffers a minor failure, where not all
the DNA from the parent are transmitted to the daughter cells. But such
instances are unrelated to the fitness of whatever DNA was lost.)
Now if you mark a particular point in the lifetime of a cell or an
organism, the correspondingly same point within each generation, and
check whether an allelle got all the way from that point in one
generation to that same point in the next generation, then you are
conflating the actual passing of DNA from parent to child and the
survival of parent before passing or child after being passed.
Is that what you tried to talk about above?? I prefer to speak of the
actual reproductive act, with nearly perfect passing of DNA from parent
to child, and the survival at all other phases in the life cycle,
whereby the cell might easily die or fail to reproduce because of the
phenotype, as two different processes to be kept separate in analysis.
.
r norman
wrote:
>an...@sci.sci wrote:
>>>>(I invented that term "fitness adaption pressure"
>>>>just now, to replace the misnomer "selection pressure" that has been in
>>>>use until now. Anybody like the replacement better than the original?)
>>>
>>>No.
>>
>>
>> Hey, what gives you the right to speak for *everyone*?
>> How about just saying "Not I" and speak only for yourself?
>> If all of the Little Red Hen's neighbors all say "Not I", then she
>> really knows nobody likes her idea, but none of them have the right to
>> speak for the others, right?
>>
>> Anyway, speaking only for yourself, do you have any third term to
>> nominate to replace the original and confusing "selection pressure" in
>> lieu of my proposed replacement "fitness adaption pressure"?
>> Maybe something like "selection-drift bias" perhaps?
>> Or "fitness-hill-climbing gradient"?
>
>Not sure who you're responding to, but "fitness-adaptation" is tautologically
>pleonastically repeating the same thing twice over again. And selection and
>drift are not polar opposites. The last offering works in some contexts.
>...
For anon1:
"No" is the correct response.
"Selection pressure" is a perfectly good term for the factors that
drive directional selection in a particular direction. No, there is
no physical pressure, but it is quite comparable to the notion of a
"diffusion force" that pushes material to flow from high concentration
to low. It is a convenient shorthand to describe factors that cause
processes to move in a particular direction.
Your policy of quoting material without the history is getting very
annoying. It becomes exceedingly difficult to find the original post
to see what you snipped out in the original. It is impossible to
resolve indefinite pronouns like "you" and "yourself".
r norman
>> You completely misunderstand the probability distribution I refer to.
>> I am speaking of the probability that an individual possessing allele
>> An (n=1, 2, ... N, the number of alternatives) will contribute a copy
>> of that allele to the next generation.
>
>If the cell is prokaryotic, or if eukaryotic and haploid and
>reproducig via mitosis, then the probability is 1, regardless of
>whether we're talking about symmetric cell division or big/large
>budding. This probability of 1 applies to each and every daughter
>thereby produced. The compound probability of at least one copy to
>*any* of several daughters is of course also 1.
Not every cell reproduces. That is the rub. The ability to pass an
allele to the next generation depends on survival (which is generally
stochastic) and whether or not the individual reproduces (which is
generally stochastic) and how many offspring are produced (which is
generally stochastic) and whether each offspring itself is capable of
surviving to reproductive age (which is generally stochastic). If the
cell is eukaryotic and diploid and reproducing sexually via meiosis
and fertilization, then you have even more probabilistic factors to
worry about. Life is uncertain. The only way to describe it
accurately is with probabilities.
I do believe you and I are talking about very different universes. I
claim mine more closely resembles the world of biology. Your opinion
may differ.
John Harshman
>>I don't know what you mean by "haplotype block".
>
>
> Any single contiguous segment of DNA that appears to have evolved as a
> single unit for hundreds of thousands of years according to the
> preliminary haplotype-map summary report that I saw in _Science_ a few
> weeks ago. Meiosis seems to have mixed-and-matched among the different
> blocks, yet seems to have kept each single block intact for all that
> time. At least that's the impression I got from reading the report.
> I had always thought that DNA can break just about anywhere, that
> some places it breaks slightly more often than other places, but in
> general there's only a small variation between hot spots and cool spots
> in regard to breaking and crossing during meiosis. But the report seems
> to say that old idea was wrong, that there's such a drastic difference
> between hot and cool spots that to a good approximation the chromosomes
> are strictly divided into fixed blocks that act as units during meiosis.
Very interesting if so, though it contradicts all sorts of other data.
But I can work with that.
Thanks for the information. I'm a bit dubious about the constancy of
these haplotype blocks. How big are they?
This definition is wrong. A chromosome would have to be very small to
form a single linkage group under most recombination rates. Or have a
special mechanism to prevent recombination.
> No! That's most definitely not what I mean here. (Note that "syntenic"
> means the same thing, but is unambiguous, applies to *any* two loci on
> the same chromosome, regardless of whether the linkage is strong enough
> to detect by standard tests, presumably Mendel-type breeding
> experiments.)
Agreed. That was a terrible definition of "linkage group".
No, that's not the centroid. The centroid would refer to non-independent
assortment.
> which is *not* what I mean, I think I'll avoid using
> that standard term when I mean the more specific kind of longterm
> linkage discovered by the haplotype-map project, where there are
> several different such blocks on each single chromosome.
Can you define "several" here? How long are these blocks? If they're so
big as to have only "several" per chromosome, then chromosomal mapping
by linkage would be impossible, and as this has been done with humans as
well as other species, I'm going to suppose that there are lots and lots
of blocks per chromosome, so they don't show up at all at the scale of
linkage maps.
John Harshman
>>I really don't know about recombination frequency through the human
>>genome.
>
>
> Did you see the article in _Science_ summarizing the results from the
> human haplotype-map project (based on Oct. 27 full report in _Nature_)?
> I'm not sure when exactly the _Science_ summary appeared, but probably
> close to the _Nature_ report.
No, I didn't.
> <http://groups.google.com/group/net-gold/msg/65477f6396258f93>
> It has been recently realized that in humans, most such swapping
> occurs primarily at a limited number of "hotspots" in the genome.
> By analyzing the HapMap data, the researchers have produced a genome-wide
> inventory of where recombination takes place.
>
>
>>But the presence of one site under selection does not alter the
>>question of whether even an adjacent site is evolving neutrally.
>
> I disagree. For any new mutation, it occurs on one haplotype-block
> allelle or another, among the several variants (allelles) of that
> particular haplotype-block. From that point onward, it's linked to all
> the other SNPs within that block-allelle, separate from any SNPs within
> other allelles of the same block. If natural selection is still working
> on that particular block, whereby different allelles will be
> retained/killed differently, because of any single selectable gene
> elsewhere on the block, then even if the new mutation is itself
> neutral, it'll be dragged with its mates on the same block, toward
> extinction or fixation, rather than drifting neutrally. Only if an
> entire haplotype block contains not a single non-neutral site, do all
> allelles of that block compete on a neutral basis, drifting neutrally
> as a unit, not dragged in one direction or another.
This is another confusion of nomenclature. If one site is under
selection, nearby sites will be dragged along with it to fixation. This
is called hitchhinking. Hitchhiking is not selection. The evolution of
those sites is still neutral.
> Furthermore if there are two or more allelles of a block, with *every*
> site on *every* allelle that block totally neutral, so that the block
> is drifting neutrally, but then a new mutation in one allelle causes a
> selectable variation between one allelle and the others, all the
> neutral sites on that one block-allelle will be dragged one way or the
> other depending on whether the mutation is beneficial or harmful (and
> the other allelles will go the other way, dragging all their neutral
> SNPs that other way). Only when that one mutated allelle has gone
> extinct, or it's become fixed by driving all the other allelles to
> extinction, will that block revert to neutral drift again.
You could if you want to talk about selection on blocks, but this is
sloppy language. It's sites that are being selected, and the rest of the
block is hitchhiking.
> And even while any such block is drifting neutrally (because no allelle
> of it has anything selectable), all the neutral SNPs on each allelle
> are drifting as a single unit, not drifting individually.
>
> So in any case there's a fundamental difference between a neutral site
> drifting all by itself, and a neutral site linked to other sites in a
> block so they drift or select as a unit.
Yes, it does make a difference to the outcome. But that doesn't change
the nature of the process. Some sites are evolving neutrally, and some
aren't. At neutral sites, the variant that is eventually fixed doesn't
depend on the variant itself at all: it's neutral.
>>It just means that if there is a selective sweep, variation at those
>>neutral sites will be reduced.
>
> If you look only at each site by itself, this is what you observe. But
> that's not the whole picture, of linked neutral drift whenever there
> isn't linked selection. Looking simultaneously at two sites on the same
> block, you see a very different picture from looking at two sites on
> different blocks. An analogy from quantum mechanics may be of value.
No, I don't think it will be.
> When two spins are entangled, each spin by itself seems truly random,
> but if you observe the two spins together you see the entanglement,
> which is decidely non-random. You see ud or du, each with 50%
> probability. but you never see uu or dd. If they weren't entangled,
> then all four combinations would be equally (25%) likely. Likewise if
> you deliberately select populations with haplotype-block allelles such
> that one particular site is exactly 50% each single-site allelle,
> you're going to have a hard time arranging that some other site on the
> same block is 50% each way independent of the first site.
I was right.
>>There is, to my knowledge, no such block structure as you imply here.
>>There are regions in which recombination is more or less frequent, but I
>>wouldn't say that talking about "blocks" does more than confuse what
>>structure there is.
>
> I really got a different impression from browsing the summary report in
> _Science_, that there's such a small number of hot spots, and such a
> small amount of recombination anywhere else, that the intervals between
> adjacent hot spots really do act like significantly-large indivisible
> blocks over many generations lasting tens or hundreds of thousands of
> years, such that many modern blocks are still intact from the major
> miagrations out of Africa, whereby each such block provides a
> non-recombinant hence pure-tree archive of mutations from then until
> now. These many pure-trees are in addition to the two pure-trees we
> already had from mitochondrial DNA and Y chromosomes, allowing us now
> to trace many additinal trees-of-descent in addition to the
> pure-female-female and pure-male-male trees.
>
> Maybe one of the better experts in this group read the _Science_
> article with more detail and can say better where the truth lies
> somewhere between my personal impression and the other view (which I
> myself held prior to seeing the HapMap summary article)?
Perhaps. I still haven't read it, and all this is news to me. There is
certainly abundant evidence for recombination working differently in
many species other than humans. A recent study of a long stretch of
avian DNA shows that sites as little as 500 bases apart are effectively
unlinked over very short timespans. Species differ in all sorts of ways,
so maybe humans (and some unknown number of other mammals) are special.
We'll see.
an...@sci.sci
> drive directional selection in a particular direction. No, there is
> no physical pressure, but it is quite comparable to the notion of a
> "diffusion force" that pushes material to flow from high concentration
> to low. It is a convenient shorthand to describe factors that cause
> processes to move in a particular direction.
Where "direction" means toward a particular goal, namely fixation of
all relatively advantageous allelles by means of extinction of all
relatively bad allelles at the same sites, ignoring all neutral sites,
right? (Caveat: Not a *teleological* goal!! Just an "attractor" or
"sink". If you turn the "fitness landscape" upside down, so you're
flowing toward a sink rather than climbing toward a peak, it makes more
sense. We already use this metaphor in thermodynamics, where we talk of
something falling down a potential well, or being jostled over a wall
by thermal energy, or tunneling through a wall if quantum mechanics is
applicable, although perhaps tunneling is really nothing more than
thermodynamics of vacuum virtual particles as with Hawking radiation.)
With diffusion, we talk about "pressure", but we might equally talk
about "gradient". There's a density gradient across a permeable
membrane, which causes particles to tend to travel from the more-dense
region toward the less-dense region, i.e. travel "down-gradient". It's
just that the membrane slows down such diffusion, compared to
free-fluid diffusion which is unimpeded. (OT: The funny thing about
permeable membranes is that they often allow only certain sizes or
types of particles to pass, so if you put one type of particle on one
side and another type on the other side, it's only the *partial*
pressure of the passible particles that matters, not the *total*
pressure in the fluid, so osmosis can actually build up a *total*
pressure difference across a membrane by equalizing the partial
pressure of passible particles while *not* equalizing partial pressure
of the non-passible particles.)
However it occurs to me that the whole metaphor of flow or pressure or
hill-climbing etc. here is wrong, because we're talking about change in
statistics of populations, not change in characters of individuals. The
individuals don't climb the hill (or flow toward the sink), only the
averages do. Lines of descent don't (in absense of new mutations) climb
the hill (etc.). Lines of descent either survive or don't survive,
either fan out to lots of descendents or stay narrow with only very few
descendents or none at all. The so-called "fitness landscape" which
life "climbs" over billions of years refers to individuals, i.e.
individuals nowadays are mostly higher up the hill than their distant
ancestors were, because surviving individuals nowadays are because of
lines of descent that happened to be so fortunate as to get new
beneficial mutations from time to time. When we talk about fixation of
a good allelle, or extincation of a bad allelle, or selection pressure,
we're merely talking about massive replication of individuals already
high on the hill and massive deaths of individuals already low on the
hill, not anybody actually moving from a low point to a high point.
Selection pressure moves the centroid of the population up the hill, by
killing of lows and replicating highs, but doesn't move any individual
or any line of descent. It's only a new mutation, or recombination of
good allelles that didn't previously appear within the same individual,
that establishes a single individual even higher up the hill than ever
before, thereby bootstrapping selection-pressure to the ability of
eventually shifting the centroid toward that new "base camp higher up
the slope of Everest".
Maybe an amoeba climbing a gradient (such as toward some food that it
can "smell" off in some direction, stronger smell at one end of the
amoeba than at the other end, hence food must be in the direction where
the smell is strongest) is a better metaphor: An amoeba moves by two
processes that somewhat alternate: First, the amoeba feebly extends its
cell membrane further along the path to its target, with only a tiny
bit of protoplasm moved into the new/extended pseudopod. Then the
amoeba flows the bulk of protoplasm from its rear toward its front
end, shifting the center of mass from where it was before to where it
is now closer to the goal. Eventually some extremely rear pseudopods
are completely drained of protoplasm and are withdrawn. Extending a new
pseudopod represents new mutations and recombinations that never
occurred before, while flowing protoplasm represents shifting the
statistics away from less-fit phenotypes toward most-fit-so-far
phenotypes, and withdrawl of rear-empty pseudopods represents
extinction of the least-fit allelles/phenotypes. Because fitness
involves simultaneous expression of many genes, with fixation or
extincation of each gene independently, but selection based on the
"whole ball of wax" rather than individual genes per se, and it's only
single allelles not phenotypes that can become extinct, i.e. old
phenotypes can re-appear due to recombination, the amoeba metaphor
isn't exact. But for asexual life, where there's no recombination so
it's impossible for old phenotypes to re-appear after they're gone
(except by very unlikely chance reverse mutation), and where again
there's no recombination so new phenotypes appear *only* by new
mutation, it may be very close to an exactly correct metaphor. Perhaps
when explaining evolution to beginners, we should explain only the
asexual case at first, using the amoeba metaphor, and then only after
it's well understood we move to the much more complicated sexual case?
OK, now I have a good metaphor for the sexual-reproduction case:
Ultra-conservative family values and housing market: When a young
couple gets married, they usually buy their own house, near one or both
of their parents' houses, but they don't live in either of their
parents' houses. Once they buy a house, they *never* move, and *never*
divorce. There is never any motion of an individual up the fitness
landscape, and children are always similar to their parents but not
identical to either. A community moves up the fitness landscape not by
moving houses or people moving to a new house, but only by bearing
young which establish new households in new locations. Those young
couples who happen to establish homes where there are no jobs, starve
to death (no welfare to bale them out, and parents already drained of
resources), whereas those young couples who happen to establish homes
where there are fine jobs have plenty of money to support large
families. As a result of differential jobs and differential family
size, the center-of-mass of communities tends to miagrate toward where
the best jobs are located. Getting married and buying a new house near
where their parents lived represents meiosis. Making babies within a
single household, and raising them until they're ready to go find a
mate, represents mitosis. Meiosis happens only once per generation.
Difficulty finding and getting together with a mate who lives far away,
and logistical difficulties of visiting relatives who live too far from
each other and from your new home, represent reproductive barriers that
tend to divide species into "ring species" and eventually into totally
separate species.
Am I now the master of metaphors?
(Splitting reply here, as I shift to a meta-topic raised by the turkey
known as "r norman".)
.
an...@sci.sci
> annoying. It becomes exceedingly difficult to find the original post
> to see what you snipped out in the original. It is impossible to
> resolve indefinite pronouns like "you" and "yourself".
It is even more annoying for people to quote the entire thread down to
their article just to reply to some small parts of it. If every Web
page quoted everything it referred to, instead of simply linking via an
anchor href, every Web page would be many gigabytes in size, making the
Web totally infeasible. Any decent Web client (browser) shows just the
name of the link instead of the entire content of the link, until and
unless the user clicks on the link, and then the user is shown the Web
page which is the target of the link.
Likewise every properly-linked newsgroup article has a References:
field in the header which is supposed to cite the path through the
thread leading to the particular article, or at least the last several
articles in that path, and any decent newsgroup client (newsreader)
should allow the user to look at any of those referenced articles.
Google Groups does this via "View Thread", although a few months ago
they changed their system to half break it, no longer showing an actual
tree ASCII-art diagram but instead just showing an outline (nested
indented) view, and if you click on "View Thread" you are taken to the
very top of the tree instead of to the location of the particular
article within the thread, and if you somehow successfully scroll down
the outline view to actually find where that article you started from
and then click on one of the ancestor articles, it takes you not to
that single article but to the first of a group of ten articles
containing the ancestor you wanted, so then you have to search for the
index number to find the actual ancestor article you wanted, but still
the basic functionalty of browsing the ancestors of an article in a
thread is vaguely available. (If anybody knows of a service better than
Google Groups for browsing threads, like Google Groups used to provide
a year ago before they half-broke their service, please let me know!)
Anyway, if you have your own computer, why don't you get a newsreader
that allows browsing the thread, including looking at ancestors of the
article you're in the midst of reading and/or responding to? If you
can't do that, it's trivial with either Google Groups or any NNTP
server, to manually copy the References: field to Note Pad or somesuch
text editor, then manually past the individual message IDs into Google
Groups search form or NNTP server to pull up that single article. Just
a bit extra work, and saves having to repeat-quoting the whole thread
in every followup as you seem to be proposing.
Note that I generally respond to individual points, in the context of
the original article. I quote just the succinct point I'm going to
respond to, then write my response. If the point isn't
self-explanatory, I cite the general context just before the text I
quote. For example, in this article of mine:
<http://groups.google.com/group/talk.origins/msg/abd1932c23808462>
I quoted the following text:
>> But what if cell membranes had a single common
>> ancestor (which they probably don't)? Wouldn't that be a good
>> candidate for a single species bottleneck? I mean, any HGT mechanisms
>> that worked prior to the first membrane probably wouldn't work for the
>> new membraned life forms. Instead they would need a modified version
>> of their own which could get through membranes. This would in effect
>> create an exclusive gene pool.
but since the topic of discussion wasn't clear from that text, I
prefaced that quote with the following topic specification:
>(Regarding HGT = Horizontal/lateral Gene Transfer)
Actually in that case I merely defined an abbreviation used in the
quoted text that might not be immediately obvious to all readers.
Here's an example where I prefaced a whole line of responses with
a general topic specification:
<http://groups.google.com/group/talk.origins/msg/acb7c317013add78>
Search for the word "Regarding" to find that point about 80 lines down
from the top where I preface like this:
>Regarding my example of development of feathers from the near-basal/LCA
>group toward modern birds:
I think if you will use the hands your parents provided you by their
genes, to manipulate your mouse and keyboard appropriately to look at
the ancestral articles in a thread any time the current article isn't
totally self-explanatory out-of-context, using features of your
newsreader, or copy&paste features of your computer and
message-ID-lookup features in both Google Groups and NNTP servers (and
most newsreaders too), you'll find my practice to be quite reasonable.
Or you can hire somebody to devise a new kind of news-reading service
that provides all those features via hypertext (in the original sense)
to make it so easy you can navigate it without hardly any effort on
your part. Or maybe there's an open-source project that has already
prodced such software you can just download and use, if you can find
it. I think there's a newsgroup dealing with neasreaders where somebody
might be able to help you find such software if it exists.
Or you can hire your secretarial staff to do all the work for you.
No, I'm not available for your hire to manually wet-nurse your needs.
But there's still a 5% unemployment rate, which includes a lot of
people who know how to use computers, so surely you can find somebody
willing to work at minimum wage to wet-nurse your online needs.
.
Nic
John Harshman wrote:
> an...@sci.sci wrote:
>
> >>I don't know what you mean by "haplotype block".
> >
> >
> > Any single contiguous segment of DNA that appears to have evolved as a
> > single unit for hundreds of thousands of years according to the
> > preliminary haplotype-map summary report that I saw in _Science_ a few
> > weeks ago. Meiosis seems to have mixed-and-matched among the different
> > blocks, yet seems to have kept each single block intact for all that
> > time. At least that's the impression I got from reading the report.
> > I had always thought that DNA can break just about anywhere, that
> > some places it breaks slightly more often than other places, but in
> > general there's only a small variation between hot spots and cool spots
> > in regard to breaking and crossing during meiosis. But the report seems
> > to say that old idea was wrong, that there's such a drastic difference
> > between hot and cool spots that to a good approximation the chromosomes
> > are strictly divided into fixed blocks that act as units during meiosis.
>
> Very interesting if so, though it contradicts all sorts of other data.
> But I can work with that.
>
<snip>
This is new to me. Even as recently as last month I was reading that
haplotype blocks were conserved in the same way linkage groups are;
i.e. low probability of being hit by a crossover. I had thus been
regarding the terms as synonymous.
The source I read was explaining that the evident conservation of
sizeable haplotype blocks underlies the mapping project which is going
on. It speculated that because the human species had been through a
severe population bottleneck in the distant past, there have not been
enough generations since that time for crossovers to have divided the
genome up very finely. I was satisfied with that explanation, but
wondered if it was the same for most species (i.e. is there always a
strong founder effect involved in speciation?).
It would be good to get to the bottom of this. What I read was a very
brief page, albeit in a very significant place:
Paragraph 3 of section _The power of haplotypes_ in the following link:
http://www.wellcome.ac.uk/en/genome/thegenome/hg04f001.html
Nic
John Harshman
>>>(I invented that term "fitness adaption pressure"
>>>just now, to replace the misnomer "selection pressure" that has been in
>>>use until now. Anybody like the replacement better than the original?)
>>
>>No.
>
> Hey, what gives you the right to speak for *everyone*?
> How about just saying "Not I" and speak only for yourself?
> If all of the Little Red Hen's neighbors all say "Not I", then she
> really knows nobody likes her idea, but none of them have the right to
> speak for the others, right?
Consider the interpretation changed in just that way. It's what I meant,
after all.
> Anyway, speaking only for yourself, do you have any third term to
> nominate to replace the original and confusing "selection pressure" in
> lieu of my proposed replacement "fitness adaption pressure"?
> Maybe something like "selection-drift bias" perhaps?
> Or "fitness-hill-climbing gradient"?
I don't find "selection pressure" objectionable.
>>>So anyway, after we outline the evidence
>>
>>Evidence for what, precisely? This seems like evidence for common descent.
>
> No no no, don't pre-judge what the evidence is "for". Just collect
> evidence like any good police detective, and then after you have
> all the evidence, see where it leads. For example, if you pre-judge
> that the husband probably killed his wife, you'll be so busy collecting
> evidence to "prove" he did it that you not bother to ask neighbors if
> they saw any one-armed men around the neighborhood, and as a result
> you'll achieve a false conviction Dr. Richard Kimble.
Actually, no scientist does anything like this. You don't collect
evidence at random, and you don't collect evidence to prove any
particular hypothesis. What you do is collect evidence that bears on a
family of hypotheses and (we hope) distinguishes among them.
Anyway, I'm not pre-judging. I'm post-judging. You will agree that
post-judging is necessary, won't you?
> So the evidence is characteristics of various fossils, and their
> location and date buried. Upon analysis, we determine that common
> descent is the most likely (parsimonious and reasonable) explanation
> for the evidence, and also anti-explanation of the many kinds of
> evidence we never see (such as chimeras of parts from widely separated
> times and places). But we come up with common descent only after
> collecting and examining and analyzing the evidence. The evidence is
> *for* determining where the species came from, not *for* proving a
> pre-conceived idea of common descent. (In fact when such evidence was
> originally collected and studied, the a priori belief was that
> everything was specially created by a supernatural being during a
> single week only a few thousand years ago, and the evidence was so
> overwealmingly opposed to that belief that paleontologists were forced
> to then think of some other theory to replace their old belief.)
I'm not sure what you're getting at here.
> In our case, if we pre-judge the case and collect evidence only in
> support of common descent, the Creationists can legitimately argue that
> we stacked the deck, collected only evidence in support and ignored
> evidence against, so our conclusion is bogus. We don't want that. We
> want to present honest evidence, and *then* show that common descent is
> the only reasonable theory to explain it. We want to invite the
> Creationists or anybody else to propose any other theory which equally
> well might explain the evidence and anti-explain the non-evidence.
> (Goddidit isn't a good theory, because it explains both the evidence
> and anything else whatsoever that godmightdo. It fails to anti-explain
> the non-evidence, to explain why cross-time cross-space chimeras don't
> occur. On the other hand, Omphalos, whereupon a supernatural being
> deliberately planted evidence to make it *appear* as if common descent
> occurred, is a valid philosophical thoery, scientifically
> indistinguishable from ordinary common descent.)
Or here.
>>what do you mean by "fossil chains"?
>
>
> Fossil species A B and C appear successively within approximately the
> same geographical region, and each appears to be very similar to its
> chronological neighbors, but clearly the three species are not
> identical to each other. How to explain that? Perhaps they form a chain
> of ancestry? By comparison we do *not* see species A B and C in such
> chronological sequence but each in a totally separate geographic area,
> nor three such species so very similar yet separated by gaps of
> hundreds of millions of years, at least hardly ever. Often we see ten
> or more such species in sequence, spaced over time, with small
> progressive change from each to the next, but rather large end-to-end
> total accumulated change. The only *other* species we ever see so very
> similar to any of these, are at what appear to be species-split
> branch-points. Those side-branches sometimes go extinct promptly, but
> sometimes start their own "fossil chains" which then run in parallel
> with the one we were already looking at.
I don't think we can actually identify such fossil chains. What we have
are trees. There are various statistical tests you can do that will
determine the degree of "strain" between the tree and stratigraphic
order. When applied to the fossil record, these tests commonly show a
much better than random fit. That's about as close as actual science
gets to these chains of yours.
>>In reality, what we have here is a single nested hierarchy,
>>formed by data from various different sources. That's all.
>
> When we fold together all the fossil chains, with the branch points
> that replace one ancestral fossil chain with two or more new fossil
> chains, we have a forest of trees. These trees are consistent with a
> single tree where lots of links are missing from our evidence so-far.
That's not how it actually works. The trees we build all have real taxa
only at the tips of the branches. There are no linear "chains", because
we have no way to tell if one species is ancestral to another.
> We can state strongly that there are fossil chains, and branch points,
> yielding lots of small common-ancestry trees. We can guess with less
> assurance that they all belong to one universal common-ancestry tree.
If what you're trying to say is that not all nodes are resolved, then
yes. But we don't in fact end up with lots of separate trees. We end up
with one tree some of whose nodes are unresolved.
> We can state with intermediate assurance that we have a single
> common-ancestry tree for each phylum, and sometimes for several phyla
> together, from the Cambrian explosion to the present. The case for
> joining all those trees into one universal tree before the Cambrian
> explosion is not firmly established by fossil evidence alone.
True. Were we limiting ourself to fossil evidence? I was unaware.
> This is where we need the other two lines of evidence (comparative
> embryology, and DNA cladograms) to piece together the separate trees
> into a single universal tree. And last I heard the best we could do is
> establish three trees for the three major domains, with unknown linkage
> earlier than that.
The problem is in defining the root. But don't confuse the lack of
resolution with a lack of a tree. The three domains are clearly united
by a host of characters. The questions are which two are more closely
related, and if a branching tree is in fact the best model.
I wouldn't call that an evolutionary theory. It has some similarities to
an evolutionary theory, but lacks others. Semantics.
>>The third is specific to some subset of theories.
>
> Again, IDiots (actually more properlty ITiots) may claim some
> supernatural Tinkerer (not Designer) is responsible for deciding which
> of the random mutations to keep and which to discard, replacing natural
> selection with deliberate breeding just as humans breed dogs.
>
> But it's impossible for any physical process to be 100% correct, to
> perform some task correctly 100% of the time, so no matter how accurate
> a phyiscal process is, there is *some* stochastic element in the
> process. So whereas a supernatural Intelligent Tinkerer might be
> totally infallable in its breeding efforts, any natural process,
> including humans-breeding-dogs, sexual selection, arms races, etc., all
> have a stochastic element. Any *scientific* theory must restrict itself
> to such natural processes, hence any *scientific* theory of selection
> must be some form of stochastic selection. I don't think it's
> mathematically possible for replication and mutation, with no selection
> of any kind whatsoever, to yield the Origin of the Species, do you?
> Accordingly *any* theory of evolution (to fit the evidence on Earth)
> *must* involve some kind of selection, hence any *scientific* theory of
> evolution must include stochastic selection.
I don't think so. For one thing, "stochastic selection" is a terrible
term, conflating drift and selection. For another, a scientific theory
can include supernatural elements. All that's required is that these
events be systematic. We can imagine a system in which there is no
stochastic element involved in fixation of new alleles, i.e. in which
the designer stacks the deck, perhaps causing enough simultaneous,
identical mutations in a population to make fixation assured. This is in
principle testable.
> Do you accept my argument to that point?
No. But I'm not sure what your argument really is.
> Then by simple math, we can separate the forcing toward a particular
> result (the "selection pressure" or whatever we call it) from the
> unexplained/random variance from that path (the "random drift").
>
>
>>None of these seem to have all that much to do with e1-e3.
>
> If you run simulations based on t1-3, assuming current theory about
> polymeric DNA language and mapping to phenotypical characters, you get
> evolutionary trees like e1, nested hierarchy in latest-generation
> populations like e2, and DNA cladograms like e3, each agreeing with the
> other two.
None of this corresponds with the way real science works. No
understanding of DNA language (whatever that is -- the genetic code?) or
mapping to phenotype or simulation is necessary to determine the nested
hierarchy. Contrariwise, all you need to get some form of nested
hierarchy is 1) branching, 2) inheritance, and 3) occasional heritable
changes.
> Do you know any other local theory (some mechanism that
> performs all action within local time and space, without any global or
> time-travel coordination) that when simulated gives results likewise
> matching all three lines of evidence?
I think you are using needlessly complex language to cover up some basic
confusion about what is and is not implied by particular sorts of evidence.
> It's not that t1-3 "have anything to do with" e1-3 in form, but that
> if you simulate t1-3 you get e1-3 as consequence, so t1-3 explains
> e1-3.
No, I don't think so. You are going backwards here. But you have
eliminated context and confused so much that it's hard to tell what
you're talking about. "A results in B" is not a good way to show that B
resulted from A, because there may be lots of other processes that also
would result in B.
> It's like Kepler's orbits and Galileo's mechanics deal with conic
> sections, while Newton's laws of gravity and motion deal with
> differential equations, with no obvious similarity between the two. But
> when you simulate Newton's laws with very small test masses (kilogram
> size) under constant-gravity conditions you end up with Galileo's
> mechanics, and when you simulate Newton's laws under zero external
> gravity but very large test masses (planetary/stellar size) you end up
> with Kepler's orbits, so Newton's theory explains both Kepler's and
> Galileo's observables.
>
> I want to clearly separate the evidence e1-3 from the explanatory
> theory t1-3, partly because everyone except YECs agrees with the
> evidence, so we can just stipulate that at the start of debates with
> everyone else, but there's a lot of debate about the underlying theory
> among IDiots etc. Most IDiots really do believe in evolution in the
> narrow sense of e2 only (common ancestry),
I don't think you are correct about that. A few idiots accept common
ancestry, but most appear not to.
> and even the YECs accept the
> nested hierarchy e1,
This also is not true. Some YECs accept it, but others claim it doesn't
exist. Sometimes the same YECs accept it one moment and claim it doesn't
exist in the next moment. I don't think any of them really believes it.
> and the DNA cladograms can be shown live in the
> laboratory so only a fool would dismiss them.
I don't know. But it's all just the nested hierarchy under another name,
and I expect the same responses: it exists but it's explained by
goddidit, not common descent, or it doesn't really exist.
> But the IDiots don't
> accept the Darwinian explanation as to how it works,
Not sure what "the Darwinian explanation" is here. But natural selection
is unnecessary to produce a nested hierarchy.
> *why* e1-3 turned
> out that way. So our challenge should be to have them present any
> alternate to our t1-3 to explain our agreed-upon e1-3 (and anti-explain
> anti-e1-3, for example why don't we see each species using a totally
> different hereditary mechanism, a different biochemistry, yet somehow
> still containing the same basic units of organic chemicals so that food
> webs work as they do, after all God could have designed life that way
> if He chose, so why didn't he? Goddiddit fails to explain why such
> differences are *not* present in life. IDiots gotta come up with
> something better to anti-explain anti-e1-3.).
This is just too confused. Too many sources of evidence and too many
separate conclusions from evidence have been smushed into a goopy mass
here. You need some rigor. What does any particular source of evidence
imply, all by itself?
r norman
I never cared for the "fitness landscape" metaphor because it assumes
that fitness is a continuous function of whatever is on the axes of
variation. These are not continuous parameters and subtle changes in
regulatory genes can produce rather abrupt and discontinuous changes
in phenotype, hence in fitness.
Still, the fact is that evolution is always about populations and
never about individuals. Selection pressure changes the allele
frequency in a population. It has nothing whatsoever to do with the
survival or reproduction if any individual.
An amoeba responding with chemotaxis to a food source may seem
superficially similar to hill climbing but has absolutely nothing to
do with evolution. When explaining evolution to beginners, something
I have done for some thirty years or more, now, you do NOT explain
only the asexual case. With the classical formulation of the theory
of evolution you start with how sexual reproduction in large
populations without selection (or mutation or migration) obeys the
Hardy-Weingerg equilibrium and no evolution results. You then
describe how any deviation from the Hardy-Weinberg assumptions --
mutation, gene flow, small population = drift, non-random mating, and
selection = non-uniform fitness allows evolution.
No, you are not at all the master of metaphor. It is rather a rather
poor one. In your metaphor, the entire burden of fitness rests on the
environment the offspring happen to inhabit and have absolutely
nothing at all to do with their genetic composition. The biological
fact is that, if the parents tend to survive beyond a single
reproductive effort, not moving away is a distinct disadvantage. It
is difficult for a newborn baby to compete with its own well
established and successful parent.
I don't understand how you say that "making babies within a single
household and raising them until they are ready to go find a mate"
represents mitosis. The babies do not have the same genetic
composition as their parents.
You seem to introduce ideas rather alien to the way biology actually
works. It is far simpler simply to talk about how sexual reproduction
produces recombination of alleles so that when you look at more than
one factor (gene locus), you get novel combinations. Fitness
differences result from an interaction between the genotype and the
environment.
And I prefer to think of myself as a guinea hen, not a turkey, though
(as Ben Franklin pointed out) that is a noble American bird in its
wild incarnation.
r norman
>> Your policy of quoting material without the history is getting very
>> annoying. It becomes exceedingly difficult to find the original post
>> to see what you snipped out in the original. It is impossible to
>> resolve indefinite pronouns like "you" and "yourself".
>
>It is even more annoying for people to quote the entire thread down to
>their article just to reply to some small parts of it. If every Web
>page quoted everything it referred to, instead of simply linking via an
>anchor href, every Web page would be many gigabytes in size, making the
>Web totally infeasible. Any decent Web client (browser) shows just the
>name of the link instead of the entire content of the link, until and
>unless the user clicks on the link, and then the user is shown the Web
>page which is the target of the link.
>
>Likewise every properly-linked newsgroup article has a References:
>field in the header which is supposed to cite the path through the
>thread leading to the particular article, or at least the last several
>articles in that path, and any decent newsgroup client (newsreader)
>should allow the user to look at any of those referenced articles.
>
Many news readers have problems when the nesting level of references
exceeds then or so. These threads just go on and on and the nesting
level increases without limit.
I am not the only one to complain about your style.
an...@sci.sci
Agreed. But in an asexually-reproducing strain, virtually *every* cell
that accumulates sufficient resources to reproduce, does in fact
reproduce. It's extremely rare that a strain of asexual critters that
has reproduced for billions of years would suddenly fail to reproduce
(due to a mutation that affects DNA replicase or some other mechanism
needed only for reproduction).
And in a multi-cellular animal or plant, the fact that some somatic
cells are programmed *not* to reproduce is irrelevant to what happens
to in the gonads.
And the fact (or urban legend) that in women's ovaries, only a fixed
number of eggs are made, and then the cells in the ovaries completely
stop reproducing, is pretty much irrelevant to the question as to
whether the woman as a whole makes babies, because there are in general
**plenty** enough eggs to last through a normal life before disease or
starvation due to dental loss ends the life. (Human women getting a
career and using birth control to postpone making babies to avoid
interfering with career advancement, then living past menopause without
having yet started a family, is somewhat of an exception compared to
most of nature. Generally a female animal either makes lots of babies
before menopause, or doesn't live that long.)
And the fact that sometimes a man or a woman can't find a mate, so
makes lots of eggs or sperm but never gets achieves fertilization, is
irrelevant to the question of whether they make the eggs or sperm in
the first place. At a cellular level, the ovaries etc. do indeed make
all the haploid half-babies as usual. When ocean-dwelling critters shed
large cloudes of sperm in the ocean, which drift around looking for
eggs, it's hard to say whether they should have counted as "already"
done their job or not.
To keep it simple, I'll stick to asexual strains such as bacteria,
where it's clear that virtually all cells that get fully fed to where
they have resources to reproduce, *do* reproduce, and those that don't
get resources simply live in stasis indefinitely until they either get
killed off somehow or they finally achieve resources to allow
reproduction. So the thing that varies is whether they live long enough
and get enough resources to reach the point where they are ready to
reproduce, not whether upon reaching that point they actually do
reproduce. I'd attribute failure to reproduce *not* to reproductive
failure per se but merely to death at an earlier point in the life
cycle. Summary: Every cell that reaches that point in life cycle, does
indeed then reproduce. The stochastic element appears elsewhere in the
life cycle, not at the instant of reproduction.
> The ability to pass an allele to the next generation depends on
> survival (which is generally stochastic)
Yes, but that was earlier in the life cycle.
> and whether or not the individual reproduces (which is generally
> stochastic)
No. If the cell survives to that point, it *does* split into two.
> and how many offspring are produced (which is generally stochastic)
There are always exactly two daughter cells after prokaryotic fission.
> and whether each offspring itself is capable of surviving to
> reproductive age (which is generally stochastic).
That's *after* reproduction has already occurred successfully. To
rewrite history by saying that it never reproduced in the first place
is a lie.
> If the cell is eukaryotic and diploid and reproducing sexually via
> meiosis and fertilization, then you have even more probabilistic
> factors to worry about.
If a cell reaches the point where it's ready to undergo meiosis, it
generally *does* do that, successfully virtually 100% of the time. At
that point is *has* reproduced in the sense of making sperms or egg.
Generally there's a surplus of sperm, so only the eggs count when
deciding success or failure of fertilization, and generally at least
one sperm does indeed find the egg, so fertilization runs to completion.
(In two clades of eukaryotes, Animalia, and Plantae, it's a little more
complicated sometimes. Let's skip those special cases, OK? In all the
other clades, it's like I said.)
> Life is uncertain.
I agree. Life, the 99% of the life cycle when reproduction isn't
happening, is uncertain. But the reproduction itself, either
prokaryotic fission, or simple non-animal non-plant sexual
reproduction, is pretty much a sure thing if life gets to that point in
the first place.
Now if you conflate the 99% of life cycle, and the 1% of reproduction
itself, into a single calculation of reproduction of the daughter cells
at exactly the same life-cycle point as we started the calculation but
exactly one generation down the line of descent, *that* conflated
calculation is indeed stochastic. But the stochastic part of it is the
life, not the reproduction. But if that's what you meant to be talking
about, you should have made it clear up front, instead of using the
word "reproduction" all by itself and claiming *that* was stochastic.
> The only way to describe it accurately is with probabilities.
If by "it" you mean the entire life cycle, all the way around the cycle
exactly once, back to the same phase you started but now one generation
later, then I agree.
> I do believe you and I are talking about very different universes. I
> claim mine more closely resembles the world of biology.
I believe you are screwing up the language to make it look like you are
saying something other than what you intend, and what you *say* is
wrong, but what you secretly believe is the same as what I say, but
unfortunately you failed to communicate that. I believe in the life
cycle of a bacterium, most of the deaths which occur are *not* at the
moment of reproduction (making daughter cells), but are somewhere else
in the life cycle, and likewise most of the extended delays are not due
to something wrong with reproduction itself, but due to something
earlier in the life cycle where not enough nutrients were built up to
be ready yet to reproduce, so it never gets to the point where
reproduction would occur, not that reproduction itself fails.
.
an...@sci.sci
> > to say that old idea was wrong, that there's such a drastic difference
> > between hot and cool spots that to a good approximation the chromosomes
> > are strictly divided into fixed blocks that act as units during meiosis.
> Very interesting if so, though it contradicts all sorts of other data.
Note that haplotype blocks that are very nearby may take several
generations before they have even a 50% chance of a cross-over between
them, thereby separating them. But over long terms (tens of thousands
of years), they eventually get separated, so what I'm saying above
refers to such long terms.
Note that there are just over 3000 million base pairs, which averages
to more than 100 million per chromosome. So even if there are thousands
of DNA bases per haplotype block (typically, not universally), that's
still less than a thousandth of a chromosome length, which is
virtually an infinitesimal on any regular chromosome map. At the
centimorgan scale, haplotype blocks of such size don't show up.
(That's my impression anyway. CMIIM)
> Thanks for the information. I'm a bit dubious about the constancy of
> these haplotype blocks. How big are they?
I don't remember from browsing the article in _Science_.
I'm hoping somebody here has that information and can answer.
(I think I already asked what the statistics of haplotype-block lengths
are, and nobody answered. So maybe I'll have to re-post my question on
the other newsgroup, the more professional one, which spends most of
its time arguing over Hamilton's equation, you know which one, right?)
> > <http://www.biochem.northwestern.edu/holmgren/Glossary/Definitions/Def-L/linkage_group.html>
> > A group of gene loci known to be linked; a chromosome. There are as
> > many linkage groups as there are homologous pairs of chromosomes.
> This definition is wrong. A chromosome would have to be very small to
> form a single linkage group under most recombination rates. Or have a
> special mechanism to prevent recombination.
You're mixing up two definitions, the linkage itself, which means
Mendel's rules for those two specific loci are violated enough to
directly measure the linkage, and the "linkage group", which is the
transitive closure of all chains of linkages. It's like the difference
between a micro-species (part of a "ring species" whereby each
individual can mate with any other of opposite sex), and the entire
"ring species" as a whole, where not all pairs of opposite-sex
individuals can pairwise mate, but there's a chain of such links from
any individual to any other individual. East-Mongolia and West-Mongolia
birds can't mate with each other, but East-Mongolia can mate with
East-China which in turn can mate with Tibet which in turn can mate
with West-China which in turn can mate with West-Mongolia, or something
like that. So among loci A B C D E, A may be linked with B and C, B
with C, C with D and E, D with E, so all five are a single "linkage
group". The entire chromosome is a single linkage group, by such a
daisy chain of A-link-B-link-C-..., assuming markers are known
sufficiently close each to the next.
> > Due to the ambiguity of definition, and the fact that the centroid of
> > the various definitions simply defines linkage group as the contents of
> > a chromosome,
> No, that's not the centroid. The centroid would refer to non-independent
> assortment.
Huh?? By "centroid" of those conflicting definitions, I merely mean a
semi-concensus, pick and choose the best of each definition to try to
guess what the "correct" definition would be, a middle-of-road
political position, not an extremist.
> > where there are
> > several different such blocks on each single chromosome.
> Can you define "several" here? How long are these blocks?
I don't remember from the article. Maybe somebody else here knows.
> If they're so big as to have only "several" per chromosome,
I didn't say *only* several. I said merely *several*, which is most
likely an understatement. (I must be turning into a Brit!)
> then chromosomal mapping by linkage would be impossible, and as this
> has been done with humans as well as other species, I'm going to
> suppose that there are lots and lots of blocks per chromosome, so they
> don't show up at all at the scale of linkage maps.
Correct, they're very large by kindergarden arithmetic (longer than a
hundred bases, maybe longer than that utterly impossible "thousand"),
but very very small by whole-chromosome-length standards. But I really
don't remember how long they are. Maybe they really do affect
linkage-maps, for example where two different genes seem to be at
exactly the same "locus", which is impossible on first thought, but a
very large haplotype block would nicely explain that occasional anomaly.
(I do actually seem to recall seeing in _Science_ some genome of
something portrayed as just a linkage map, with two genes indicated at
exactly the same spot on the map, not once but several different such
cases. But I might be mis-remembering.)
By the way, in the midst of just starting to compose this reply, I did
a Google search about the human genome, to find out the total number of
DNA bases (which I quoted above, just over 3 billion, actually just
over 3.1 billion), and saw something somewhere that said there are 30
thousand genes, of which only half have known function. So I got to
thinking: Are there really 30 thousand genes, or only 30 thousand open
reading frames, which aer *presumed* to all be genes, but for all we
know half of them are suppressed, inactive, don't actually generate
either RNA or polypeptides. But I wasn't sure I knew the correct
definition of ORF so I did another Google search, and found rather
conflicting definitions. Which of these (grouped by synonyms, each
group contradicts the others) is the correct ORF definition?
<http://www.biochem.northwestern.edu/holmgren/Glossary/Definitions/Def-O/open_reading_frame.html>
A section of a sequenced piece of DNA that begins with an initiation
(methionine ATG) codon and ends with a nonsense codon. ORFs all have
the potential to encode a protein or polypeptide, however many may not
actually do so.
<http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/O/O.html>
DNA sequence beginning with ATG and read in triplets until it
ends with a STOP codon. An ORF is potentially able to encode a
polypeptide.
<http://66.102.7.104/search?q=cache:IqPKpi5TztwJ:manatee.sourceforge.net/pdf/overview.pdf+open+reading+frame&hl=en&ie=UTF-8>
o Telling the difference between random ORFs and
genes is the goal in the gene finding process.
(That's the definition I thought was correct, but see the others:)
<http://www.medterms.com/script/main/art.asp?articlekey=4644>
Open reading frame: A long sequence of DNA that has no stop codon (no
signal to stop reading) and therefore may encode part or all of a
protein.
<http://opbs.okstate.edu/~melcher/MG/MGW2/MG241.html>
Regions of nucleotide sequences devoid of termination
codons in one frame of reading are designated "open reading
frames".
<http://66.102.7.104/search?q=cache:l1xHsaVtaNIJ:hal.weihenstephan.de/genglos/asp/genreq.asp%3Fnr%3D292+open+reading+frame&hl=en&ie=UTF-8>
A sequence of DNA following an initiation codon that does not contain
a stop codon. Detection of an open reading frame in DNA implies the
presence of a gene that codes for a protein.
<http://www.iscid.org/encyclopedia/Open_Reading_Frame>
A reading frame, in biology, consists of three-nucleotide codon sets
in either DNA or RNA that are contiguous and non-overlapping. An open
reading frame (ORF) is a similar sequence that can be translated into
a protein or a polypeptide.
In any open reading frame, the start-code sequence or initiation codon
that begins the protein is methionine ATG, and then stop-code sequence
or termination codon ends it. The stop sequence is coded by what is
termed a nonsense codon, or a codon that does not have an RNA match.
There are only three nonsense codons: amber(UAG) ochre(UAA) and opal
(UGA). As you can see, each one contains a "U" nucleotide, not normal
to DNA.
(That last one is bonkers!! Who wrote it, Archimedes Plutonium??)
I also found currently-online Web pages which disagree about number of
human genes. What I read so-far in _Science_ over the past five years
was that the person who guessed 30K won the bet, because the number was
actually about 25-30K, but then the number went down lower to maybe
only 10-20K, but then the number came back up to 30K again. What's the
latest best-estimate of ORFs (any sequence starting with ATG and being
of sufficient length before the next stop codon, including both actual
coding genes and pseudo-genes), and number of only the actual coding
genes, as two separate questions?
<http://www.thedoctorwillseeyounow.com/articles/other/genome_4/index.shtml>
It is believed that
the human genome has between 50,000 and 100,000 genes.
(Somebody needs to update their Web site!)
It is estimated that the sequence of the human genome should be
completely mapped by approximately the year 2005.
(Yeah!)
<http://www.infoaging.org/b-human-12-glossary.html>
The human genome is estimated to contain
about 30-40,000 genes,
<http://www.schoolscience.co.uk/content/4/biology/abpi/genome/genome3.html>
With between 30 000 and 40 000 genes,
<http://www.ornl.gov/sci/techresources/Human_Genome/project/info.shtml>
The human genome is estimated to contain 20,000-25,000 genes.
<http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/H/HGP.html>
1. The number of genes turned out to be much smaller than once predicted.
The two groups came up with slightly different estimates of the number
of protein-encoding genes, but both in the range of 30 to 38 thousand:
* and a third of the 100,000 genes that many had predicted would be
found.
* (By the end of 2004, the number had been reduced to some
20,000-25,000.)
.
BruceW
an...@sci.sci wrote:
>>Your policy of quoting material without the history is getting very
>>annoying. It becomes exceedingly difficult to find the original post
>>to see what you snipped out in the original. It is impossible to
>>resolve indefinite pronouns like "you" and "yourself".
>
> It is even more annoying for people to quote the entire thread down to
> their article just to reply to some small parts of it.
> ...
There *is* middle ground here, you know.
Also, you might want to consider that when you post childish insults,
those insults say much more about you then they say about the person to
whom the insults are directed.
-BruceW
Nic
an...@sci.sci wrote:
<snip>
> And the fact that sometimes a man or a woman can't find a mate, so
> makes lots of eggs or sperm but never gets achieves fertilization, is
> irrelevant to the question of whether they make the eggs or sperm in
> the first place. At a cellular level, the ovaries etc. do indeed make
> all the haploid half-babies as usual. When ocean-dwelling critters shed
> large cloudes of sperm in the ocean, which drift around looking for
> eggs, it's hard to say whether they should have counted as "already"
> done their job or not.
The ocean dwelling critters have indeed done their job. The humans
haven't even got off the starting blocks. A component of the
subsequent generation's success is creditable to the parental genotype.
That is how evolution 'sees' it, anyway. That component is *in
addition to* the component due to the parental alleles being
represented in the offspring. It is due to the actual combination
expressed in the parents. A successful oak tree producing offspring
only of relatively poor fitness, will still be well represented in the
grandchild generation because it can throw enough acorns at the
problem. That is not a shortcoming of the system. Far from it. It is
not just alleles that behave as if they were selfish - it also any
combination thereof, and up to any size of combination. Maybe
something like the original successful oak will emerge in the
grandchild generation. In the case of the humans, if they put their
baby on a raft let it float down the river into an area known to be
uninhabited, then that is a black mark against their genotypes, not
those of the offspring.
<snip>
> > The ability to pass an allele to the next generation depends on
> > survival (which is generally stochastic)
>
> Yes, but that was earlier in the life cycle.
>
> > and whether or not the individual reproduces (which is generally
> > stochastic)
>
> No. If the cell survives to that point, it *does* split into two.
>
> > and how many offspring are produced (which is generally stochastic)
>
> There are always exactly two daughter cells after prokaryotic fission.
>
> > and whether each offspring itself is capable of surviving to
> > reproductive age (which is generally stochastic).
>
> That's *after* reproduction has already occurred successfully. To
> rewrite history by saying that it never reproduced in the first place
> is a lie.
A white lie. See baby on raft example above.
You seem to be hanging back from the idea of meiosis introducing any
randonimity into the fate of alleles. Is that because you know it
produces gametes in pairs, so that no parental allele actually gets
left out? If so, I see your point. The randonimity/selection comes in
after this, when the haploid phase of the life cycle takes its 'test'.
However, two points. Gamete survival, although it 'reads in' some
environmental information, it reads in practically none that it is
responsible for species diversity. I think this is purposeful - it is
inefficient for one genome to have to master two niches (hence the
pressure for neoteny, and for marine mollusks to keep ditching their
planktonic phase). That is not really a point against what you are
saying, just an observation.
My second point is that meiosis *does* introduce randonimity into the
fate of alleles because every time it produces an unviable combination
it blocks the reproduction of alleles which aren't part of that
combination, as well as ones that are. Maybe I'm falling into the trap
of double counting here, and that effect is already included in haploid
phase selection. Surely though, gametes (especially female ones) are
not produced in sufficient quantity to average this effect out.
Suppose the parents are contributing rival alternative coadapted gene
complexes, one of which is wholly dominant. In this case meiosis will
*almost* always break both complexes, and fatally. Maybe this is what
happens in mules.
<snip non-argument>
Nic
r norman
I still believe we are talking about two very different universes,
only one of which resembles real biology.
Take a steady-state ecosystem including, say, one million asexually
reproducing organisms, whether bacteria or amoeba or whatever. Wait a
year after which time these organisms will have gone through many cell
divisions. Because the ecosystem is in a steady-state, you will end
up with one million organisms. Trace the ancestry of these one
million back to the original population and you will get a tiny
fraction of the original successfully reproducing.
The fact is that most systems are usually in something close to a
steady state. As Darwin (and Malthus) noticed a few years ago, the
propensity to procreate far exceeds the reality of procreation. So
your very lengthy analysis of how mitosis works and how many
unfertilized eggs and sperm exist and everything you say above is
completely meaningless for evolution. The only thing that matters is
having ones particular alleles passed down to future generations.
At the end of your screed you do say ">If by "it" you mean the entire
life cycle, all the way around the cycle exactly once, back to the
same phase you started but now one generation later, then I agree".
The only thing that makes sense biologically in reproduction is the
entire life cycle, from adult organisms in one generation all the way
to adult organisms in the next generation. There are events that
occur constantly throughout that complete cycle that are either random
or depend on so many uncontrolled and external factors that they are
best described as random. On the other hand, the probability of
successfully completing the cycle also does depend on the phenotype
and genotype of the original parent. Hence, drift plus selection in
the fact that different genotypes have different probabilities in
their ability to successfully reproduce (going the full cycle). The
probability distribution, defined over the space of genotypes, is
non-uniform.
John Harshman
>>>the report seems
>>>to say that old idea was wrong, that there's such a drastic difference
>>>between hot and cool spots that to a good approximation the chromosomes
>>>are strictly divided into fixed blocks that act as units during meiosis.
>>
>>Very interesting if so, though it contradicts all sorts of other data.
>
>
> Note that haplotype blocks that are very nearby may take several
> generations before they have even a 50% chance of a cross-over between
> them, thereby separating them. But over long terms (tens of thousands
> of years), they eventually get separated, so what I'm saying above
> refers to such long terms.
I wasn't talking about independent assortment between haplotype blocks.
I was talking about the very existence of haplotype blocks as stable
entities.
> Note that there are just over 3000 million base pairs, which averages
> to more than 100 million per chromosome. So even if there are thousands
> of DNA bases per haplotype block (typically, not universally), that's
> still less than a thousandth of a chromosome length, which is
> virtually an infinitesimal on any regular chromosome map. At the
> centimorgan scale, haplotype blocks of such size don't show up.
> (That's my impression anyway. CMIIM)
That's what I was wondering about.
That's not the way it's generally used in my experience.
>
>>>Due to the ambiguity of definition, and the fact that the centroid of
>>>the various definitions simply defines linkage group as the contents of
>>>a chromosome,
>>
>>No, that's not the centroid. The centroid would refer to non-independent
>>assortment.
>
> Huh?? By "centroid" of those conflicting definitions, I merely mean a
> semi-concensus, pick and choose the best of each definition to try to
> guess what the "correct" definition would be, a middle-of-road
> political position, not an extremist.
Right. That's what I mean too. I don't think your semi-consensus is the
correct semi-consensus based on the definitions you have given.
It's convoluted and contains much unnecessary information, but it's
really the same as all the others, which match each other. Some forget
to mention the initiation codon, and that's the only difference I can
see. But the way the number of genes is estimated is more complicated
than this. Gene-estimation programs are used to identify genes, introns,
exons, and such, and they rely on a large set of clues. I've never
looked at this in detail, but references can be found on the UCSC genome
browser at least for the gene-finding programs they have used.
> I also found currently-online Web pages which disagree about number of
> human genes. What I read so-far in _Science_ over the past five years
> was that the person who guessed 30K won the bet, because the number was
> actually about 25-30K, but then the number went down lower to maybe
> only 10-20K, but then the number came back up to 30K again. What's the
> latest best-estimate of ORFs (any sequence starting with ATG and being
> of sufficient length before the next stop codon, including both actual
> coding genes and pseudo-genes), and number of only the actual coding
> genes, as two separate questions?
I don't know.
> <http://www.thedoctorwillseeyounow.com/articles/other/genome_4/index.shtml>
> It is believed that
> the human genome has between 50,000 and 100,000 genes.
> (Somebody needs to update their Web site!)
> It is estimated that the sequence of the human genome should be
> completely mapped by approximately the year 2005.
> (Yeah!)
Ah, internet paleontology, or vestiges of the natural history of web
creation.
an...@sci.sci
> No, I didn't.
Oh well, just as well. It's pretty big and hard for a layman to find
the meaningful (to the layman) information of interest. I wish I could
find a nicely organized Web site that outlined the major findings
so-far, something like a HapMap1-results-FAQ, so that we could
efficiently find the key info I want, such as statistics of
haplotype-block length (to answer questions about how it affects
locus-mapping efforts), statistics of estimated age of most recent
cross-over at each boundary between adjacent blocks (to then estimate
time of last serious bottleneck i.e. when founders of our clade lived),
etc.
> If one site is under selection, nearby sites will be dragged along
> with it to fixation. This is called hitchhinking. Hitchhiking is not
> selection. The evolution of those sites is still neutral.
Well the evolution *caused*by* those sites is neutral, but those sites
are dragged along with their neighbors, so the contingent statistics of
those sites aren't neutral at all. I.e. suppose we have 50% mix of two
allelles each of two very closely linked sites, exactly uncorrelated at
the start (each combination AB Ab aB ab has exactly 25% occurrance; I'm
pretending they're haploid to make the math easier here). One site is
neutral, and the other is strongly selected. Now a few generations
later the selected site is biassed 90% 10%, while the neutral site is
still sitting at 50-50. (The combinations are AB=45% Ab=5% aB=45%
ab=5%.) But that's rarely the initial condition. More likely the
initial condition, due to longterm strong linkage (haplotype blockage)
is that the two loci are highly correlated. For sake of argument, let's
say they are 100% correlated (AB=50% Ab=0% aB=0% ab=50%). Now do the
same few generations of selection, and now the selected site is 90-10
just like before, but the neutral site has been dragged to 90-10 too
(AB=90% Ab=0% aB=0% ab=10%). Looking only at the contingent
probabilities of the neutral site, they started at 50-50 and are now
90-10, so they don't satisfy the hypothesis of neutral-drift contingent
statistics. This is because of being dragged, not because of their own
fitness, but nevertheless they still don't look statistically like they
are neutral. Now look at them 10 thousand years later. Because of very
strong linkage (they are in the same haplotype block, no hot spots
between them), both selected and neutral loci are fixed at 100%-0%.
Now let's imagine the neutral site is observable (color of eyes), while
the selected site is invisible (more efficient conversion of citrate to
isocitrate when stressed by presence of a toxin that only recently
started polluting the local water from a local industry, hence lack of
selection to fix the better allelle prior to the start of this time
period we're studying). So we do all the measurements we can do with
our level of technology, and the actually-neutral site appears to be
actively selected, no way to tell it isn't being selected, so we have
to treat it as if selected in all our predictions of results of
experiments. Yeah, it's not *really* selected, but at present that's a
philosophical not scientific point of view.
So what I'm saying overall is that due to hitchhiking, together with
the very strong difference between hottest and coldest recombination
sites (large blocks of no recombination whatsoever over tens of
thosands of years, spaced by hot spots that recombine with observable
measurable occurrance in even a single meiotic generation, with not as
many spots of intermediate hotness as would be naively expected, at
least that's my impression from the HapMap1 summary in _Science_,
CMIIM), to a first approximation it's as if continguous blocks evolve
as single units over tens of thousands of years, *not* as separate loci
which are "linked" to an intermediate degree. AIUI, intermediate
linkage (halflife of remained "joined at the hip" on the order of a few
generations to a few hundred years) occurs between nearby haplotype
blocks, not within a single such block. So if you have two loci in
nearby blocks, they follow linked statistics, but if you have two loci
in a single block they act as if they were a single locus over time
spans longer than you're willing to study in the lab or in nature.
(In another article in this same thread, posted sometime within the
past 24 hours, somebody seemed to say that the only reason there are
*any* conserved haplotype blocks now is because it's still only a short
time, less than a million years, since a serious population bottleneck.
So the full truth may be that unlinked "haplotype blocks" are an
accurately valid hypothesis only over medium-length spans of time, more
than 20 generations but less than a million years, while linked
"haplotype blocks" are the best hypothesis over shorter time spans, and
the traditional view of each locus separate from each other locus is
valid over spans of many millions of years when *all* blocks have been
internally crossed at least once. However it may be that some pairs of
adjacent bases are *never* crossed during meiosis, that the mechanism
of crossing over somehow "misses" them due to their very strong binding
and lack of any mechanism to ever break them, or because of very rare
breaking and 100% successful repair to original state. Due to recent
population bottleneck in human species, I don't think we'll ever be
able to answer this question for humans, at least not within the next
several million years it'll take to achieve linkage equilibrium
everywhere possible in our genome.)
> You could if you want to talk about selection on blocks, but this is
> sloppy language. It's sites that are being selected, and the rest of the
> block is hitchhiking.
Yes. We're in agreement on the science. It's just the way to word it
where we disagree. I claim that whereever there's a haplotype block
that persists over tens of thousands of years, due to the lack of even
one hot or even warm crossing site within it, for all practical
purposes (for evolution studies over such time spans and any shorter
spans) the various sites on that block are co-evolving as a single
unit, with selection pressure on them all merged into a single
effective selection pressure on the group, so it's group selection that
we need be concerned about here, and the actual selection on individual
loci within that block can't be seen by any experiment. We might be able
to measure the selection on individual loci by contrived experiments
wherein we deliberately apply really strong selection on a single site
compared to normal, for example if we have good reason to believe we
can *double* selection pressure for a given site. Then if we measure
the total selection effect (non-neutral drift in population) with and
without our doubling, and subtract, we thereby compute the actual
selection pressure of that one site, although not with good
signal/noise ratio. Alternately we might directly observe deaths of
individual cells/organisms, perform autopsies on each to determine
cause of death, and thereby attribute some allelle of some locus at
fault for the death, and by compiling statistics of such
causes-of-death we might then calculate selection pressure for each
non-neutral locus. But that's not yet feasible, so I think for now we
have to treat all loci on a block as if a single locus, and learn to
deal with more than four possible allelles that code for more than one
polypeptide or RNA and thereby affect more than one aspect of life.
Anyway, it's good that we at least agree on the science here. It's much
pleasenter to debate you on models to use than to debate IDiots or YECs
where they are totally wrong on the science yet won't budge.
> Yes, it does make a difference to the outcome. But that doesn't change
> the nature of the process. Some sites are evolving neutrally, and some
> aren't. At neutral sites, the variant that is eventually fixed doesn't
> depend on the variant itself at all: it's neutral.
Aha, yes in *that* sense even with hitchhiking the particular locus is
still neutral. But still thinking of it all as a "single site" may be
the best way. At the start you have a fixed number of allelles for the
entire block (due to recent mutations prior to the start of our time
span of study, and due to ancient mutations that were neutral and all
linked to the one best allelle which got fixed prior to those recent
mutations). We can factor those block-allelles into the single allelle
for the fixed selected loci, and the various allelles for the neutral
loci, and the various allelles for the loci that have recently mutated.
But because of their longterm entanglement (100% linkage) I would treat
them as a unit, a single giant locus with several allelles, some
identical in fitness because they vary only at neutral sites, and some
different in fitness because they vary from the rest at recently
mutated sites. So let's say three mutations are bad, and one is good
but only slightly better than the original. So first the three
block-allelles containing the bad mutations are eliminated, but the
corresponding three not-mutated block-allelles remain. Meanwhile the
frequency of the one block-allelle containing the good mutation
increases slightly in frequency. Over longer times, gradually that one
block-allelle increases to fixation, driving all the other block
allelles to extinction, and since there's been zero recombination
within this block in all this time, all the neutral sites within this
block that had variation at the start are likewise driven to fixation
according to which of their allelles was mated with the good mutation.
(Is this what is meant by a selective sweep?) Meanwhile a few new
neutral mutations occurred within the blocks not having the good
mutation, which were swept to extinction too, while a few new neutral
mutations occurred in copies of that one block having the good
mutation, and these new mutations are retained to the end of the time
span. So now, once again, we have a situation of a single selected
allelle that is fixed, within a block containing a few recent mutations
at neutral sites, in effect a large block-site having a mix of allelles
all of equal fitness.
Note that while those sites are neutral *now*, they might not be
neutral for all time. Some change in environment could later make one
or another of them selectable.
The same logic works at a single DNA base due to (1) more than two
bases possible, (2) some base-differences important and some not, (3)
possible frame shift which changes the importance of some base
differences. Consider this particular set of bases at one site, the
third of a triple:
AAT and AAC both code for N (Asparagine)
AAA and AAG both code for K (Lysine)
If that third base is T or C, it makes no difference which, because the
result is the same, Asparagine. Likewise if it's A or G, no difference
between those two. But it *does* make a difference whether it's for
example, T or G, because that change codes for a different amino acid.
We can pretend like the two aspects of this single base are two
different factors, as if two different loci, which happen to be 100%
linked within a haplotype block, with one aspect neutral and the other
aspect selected. Suppose for example we have a population with lots of
AAT and AAC, a little bit of AAA, but no AAG at all. So with respect to
the neutral factor, it's just about balanced, but with respect to the
selectable factor, it's mostly code for Asparagine. Now suppose there's
selection for Lysine in preference to Asparagine. Suppose in fact
Lysine gets fixed. That eliminates all the AAT and AAC, fixes the AAA,
and there never was any AAG so the neutral factor which distinguishes T
from C and A from G has been driven to fixation too by hitchhiking.
Now suppose later a copy of the block is made, and a frame shift occurs
in one copy, whereby now the difference in that formerly neutral factor
is very important. Suddenly the fact that the second factor hitchhiked
its way to fixation is important. So it really is like a second locus,
in that it affects fitness under different circumstances from when the
other factor affects fitness.
Anyway, done with that thinking way outside the box...
> A recent study of a long stretch of avian DNA shows that sites as
> little as 500 bases apart are effectively unlinked over very short
> timespans.
Does that observation apply to *all* pairs of sites located at that
pairwise distance, or only to carefully selected sites that straddle a
recombination hot spot?
> Species differ in all sorts of ways, so maybe humans (and
> some unknown number of other mammals) are special.
The most fundamental question to ask is what causes there to be hot
spots in the first place. Are certain sequences of DNA bases more
stable as molecules in the first place, as confronted with mere
thermodynamic jostling? Or are there enzymes that tend to actively
break DNA at certain places based on matching either a pattern there or
a pattern upstream or downstream some fixed distance, or are there
preferential places where other chemicals can attach to the DNA strand
to protect those regions and anywhere they don't attach is more
vulnerable to breakage, or what? Is anybody working on that question?
.
John Harshman
Where did you get the idea that neutral evolution requires equal
frequencies of all alleles, which you seem to have here? You understand
that neutral evolution will always result in either fixation or
extinction of the allele, right?
Yes, we attempt to detect selection by looking at nucleotide diversity
in a small region. But we don't therefore suppose that an entire region
is under selection, merely that some site within that region is under
selection.
> Now let's imagine the neutral site is observable (color of eyes), while
> the selected site is invisible (more efficient conversion of citrate to
> isocitrate when stressed by presence of a toxin that only recently
> started polluting the local water from a local industry, hence lack of
> selection to fix the better allelle prior to the start of this time
> period we're studying). So we do all the measurements we can do with
> our level of technology, and the actually-neutral site appears to be
> actively selected, no way to tell it isn't being selected, so we have
> to treat it as if selected in all our predictions of results of
> experiments. Yeah, it's not *really* selected, but at present that's a
> philosophical not scientific point of view.
Or, to put it more reasonably, with present technology it's impossible
to tell whether the allele is being selected or is hitchhiking.
Careful. You just used the term "group selection" when all you meant was
the sum of selection. Now in actuality, we could do experiments on
individual loci or individual sites by genetically engineering changes,
even if in nature the blocks never recombined internally. We could also
examine the specific effects of individual differences on physiology,
anatomy, or whatever by looking at the actual biological mechanisms more
closely, and that would let us predict the effects on fitness of
individual bits, even if they were never separated into different
individuals.
> We might be able
> to measure the selection on individual loci by contrived experiments
> wherein we deliberately apply really strong selection on a single site
> compared to normal, for example if we have good reason to believe we
> can *double* selection pressure for a given site. Then if we measure
> the total selection effect (non-neutral drift in population) with and
> without our doubling, and subtract, we thereby compute the actual
> selection pressure of that one site, although not with good
> signal/noise ratio. Alternately we might directly observe deaths of
> individual cells/organisms, perform autopsies on each to determine
> cause of death, and thereby attribute some allelle of some locus at
> fault for the death, and by compiling statistics of such
> causes-of-death we might then calculate selection pressure for each
> non-neutral locus. But that's not yet feasible, so I think for now we
> have to treat all loci on a block as if a single locus, and learn to
> deal with more than four possible allelles that code for more than one
> polypeptide or RNA and thereby affect more than one aspect of life.
Actually, similar sorts of studies are quite feasible with current
technology.
I'm also not sure what sort of point you are trying to make here. I've
lost sight of all the context, since you continually snip all of it away.
> Anyway, it's good that we at least agree on the science here. It's much
> pleasenter to debate you on models to use than to debate IDiots or YECs
> where they are totally wrong on the science yet won't budge.
>
>
>>Yes, it does make a difference to the outcome. But that doesn't change
>>the nature of the process. Some sites are evolving neutrally, and some
>>aren't. At neutral sites, the variant that is eventually fixed doesn't
>>depend on the variant itself at all: it's neutral.
>
> Aha, yes in *that* sense even with hitchhiking the particular locus is
> still neutral. But still thinking of it all as a "single site" may be
> the best way.
I don't think it is for most purposes. The purpose for which it's the
best way is when you are computing a fitness, and you want to know which
block is going to be fixed.
> At the start you have a fixed number of allelles for the
> entire block (due to recent mutations prior to the start of our time
> span of study, and due to ancient mutations that were neutral and all
> linked to the one best allelle which got fixed prior to those recent
> mutations). We can factor those block-allelles into the single allelle
> for the fixed selected loci, and the various allelles for the neutral
> loci, and the various allelles for the loci that have recently mutated.
> But because of their longterm entanglement (100% linkage) I would treat
> them as a unit, a single giant locus with several allelles, some
> identical in fitness because they vary only at neutral sites, and some
> different in fitness because they vary from the rest at recently
> mutated sites. So let's say three mutations are bad, and one is good
> but only slightly better than the original. So first the three
> block-allelles containing the bad mutations are eliminated, but the
> corresponding three not-mutated block-allelles remain. Meanwhile the
> frequency of the one block-allelle containing the good mutation
> increases slightly in frequency. Over longer times, gradually that one
> block-allelle increases to fixation, driving all the other block
> allelles to extinction, and since there's been zero recombination
> within this block in all this time, all the neutral sites within this
> block that had variation at the start are likewise driven to fixation
> according to which of their allelles was mated with the good mutation.
> (Is this what is meant by a selective sweep?)
Yes.
That was a great deal of verbiage to wade through, and I think it was a
terrible analogy filled with sloppy and confusing language. All we have
is selection among 4 possible alleles at the site, only 3 of which are
realized in the population. A beats C or T, and is therefore selected.
That's all.
> Now suppose later a copy of the block is made, and a frame shift occurs
> in one copy, whereby now the difference in that formerly neutral factor
> is very important. Suddenly the fact that the second factor hitchhiked
> its way to fixation is important. So it really is like a second locus,
> in that it affects fitness under different circumstances from when the
> other factor affects fitness.
Or, more reasonably, you could say that frameshifts can change the
selective value of different nucleotides at a given site.
> Anyway, done with that thinking way outside the box...
>
>
>>A recent study of a long stretch of avian DNA shows that sites as
>>little as 500 bases apart are effectively unlinked over very short
>>timespans.
>
> Does that observation apply to *all* pairs of sites located at that
> pairwise distance, or only to carefully selected sites that straddle a
> recombination hot spot?
As I recall, they could find no significant linkage between any two
sites separated by 500 bases or more.
>>Species differ in all sorts of ways, so maybe humans (and
>>some unknown number of other mammals) are special.
>
> The most fundamental question to ask is what causes there to be hot
> spots in the first place. Are certain sequences of DNA bases more
> stable as molecules in the first place, as confronted with mere
> thermodynamic jostling? Or are there enzymes that tend to actively
> break DNA at certain places based on matching either a pattern there or
> a pattern upstream or downstream some fixed distance, or are there
> preferential places where other chemicals can attach to the DNA strand
> to protect those regions and anywhere they don't attach is more
> vulnerable to breakage, or what? Is anybody working on that question?
There is literature on this subject, but I'm not really up on it. If you
have access to a scientific literature database like BIOSIS or Medline,
you could look this up. PubMed is freely available and probably has what
you want. Pick a few search terms and go to it. (I commonly use PubMed
through the NCBI web site, though I'm sure there are other ways.)
an...@sci.sci
> were conserved in the same way linkage groups are; i.e. low
> probability of being hit by a crossover. I had thus been regarding
> the terms as synonymous.
They are not synonymous. Of the various definitions of "linkage group"
that I saw online, the most reasonable definition I saw was essentially
any maximal connected subset per linkage pairs as the connection links.
<http://opbs.okstate.edu/~melcher/MG/MGW1/MG111122.html>
* Markers that have measurable recombination frequencies are said to
be linked.
* Markers related through a chain of linkage constitute a linkage
group.
For example, in mathematical graphs, if there are points A,B,C,D,E,
where A is connected to B, B is connected to C, C is connected to E,
and there are no other connections between any of those points, then
A,B,C,E are one maximal connected set and D all by itself is another
maximal connected set. By necessity, membership in such maximal
connected subsets is an equivalence relation.
Now a larger example in genetics: Suppose along a chromosome there are
sites/markers/loci A,B,C,D,E,F,G such that each marker is linked to its
immediate neighbors and also to its next-to-immediate neighbors, but
not to its next-to-next-to-immediate neighbors. Now consider the most
distant loci, A and G. The following are just a few of the chains of
linkage from A to G: A-B-C-D-E-F-G, A-C-E-G, A-B-D-F-G,
A-C-B-D-C-E-D-G-E-G, etc. It doesn't matter how many chains there are
from A to G, it just takes one such chain to establish that A and G are
in the same linkage group. Obviously in this case all seven loci are in
a single linkage group.
In most cases in genomes, a linkage group is in fact an entire
chromosome, because for any two loci A,Z anywhere within a single
gehome there's soem chain of other loci B,C,...,X,Y such that there's
linkage between A and B, between B and C, ...etc..., between X and Y,
between Y and Z, thereby establishing a chain A,B,C,...,X,Y,Z between A
and Z. The only exception would be if there was a rcombination "hot
spot" so very very hot that there was no linkage between loci
immediately on opposite sides of it, i.e. during every single instance
of meiosis there's 50% chance of the two sides of that hotspot swapping
and 50% chance of them not swapping (or swapping and then swapping back
before meiosis is finished). Note it's *because" the whole chromosome
is a single linkage group that it's possible to map entire chromosomes
using fractions of a "Morgan", usually centiMorgans.
<http://www.sizes.com/units/centimorgan.htm>
The genetic distance between two loci is 1 cM if their statistically
corrected recombination frequency is 1%; the genetic distance in
centimorgans is numerically equal to the recombination frequency
expressed as a percentage. Typically a genetic distance of 1 cM
corresponds to a physical distance of roughly one million base pairs.
Note that if the distance between two loci is 1 centiMorgan, then
there's 99% linkage between them, clearly showing they are "linked". So
if you have a map of loci showing them at various centiMorgan distances
from the start, i.e. each locus is about 1 centiMorgan from any at the
next tick mark on the scale, clearly you have a linkage every little
step, hence a chain all the way from end to end. (Caveat: When the
total accumulated distance approaches 50 centiMorgans, that does not
mean that two loci spaced that far apart have 50% chance of a crossing
between them, because compound probabilities aren't exactly additive,
and aren't even close to additive over such large distances. Does
anybody need a mathematical explanation of why it's not exactly
additive even though it's close to additive over short distances?
Hint: P*P vs. 2*P*(1-P)) vs. (1-P)*(1-P), where P=0.99, and that middle
term is what you want to look at.)
> It speculated that because the human species had been through a
> severe population bottleneck in the distant past, there have not been
> enough generations since that time for crossovers to have divided the
> genome up very finely.
You wrote the premise backwards. You meant to say because the human
species had been through a severe population bottleneck in the
*not*very*distant* past ...
Yes, with that correction, I agree that is one good reason why today we
observe a small number of different haplotype blocks compared to what
would happen with an unlimited-size population over an unlimited amount
of time. I think the other factor is the great disparity of "hotness"
between various recombination sites. See later below...
> I ... wondered if it was the same for most species (i.e. is there
> always a strong founder effect involved in speciation?).
That is one of the major hypothesis as to what triggers a species split
(which is then followed by one of three scenerios: new tiny-population
species goes extinct before it leaves any fossils, new tiny-population
species drives parent species to extinction, both species survive long
enough to leave parallel set of fossils, sometimes both surviving long
enough to split further). It seems to me that haplotype studies could
answer this question, by checking several other species to see whether
they likewise show a population bottleneck as evidenced by a
restriction of haplotype variety just about at the point where the
species diverged visibly (fossil characters) from the parent species.
> http://www.wellcome.ac.uk/en/genome/thegenome/hg04f001.html
20/3/03. By RT
A new map of human variation will greatly aid research on the genetic
origins of disease.
(Note this was written *before* the project was done, indicating what
they *expected* to learn, not what was actually learned after
completion and post-analysis.)
Genetically speaking, humans are incredibly similar to one another.
Any two unrelated genome sequences differ at only one position in a
thousand, on average. The 0.1 per cent difference, which amounts to
about three million base pairs of DNA in total, ...
Much genetic diversity (around 90 per cent) consists of single
nucleotide polymorphisms (SNPs), ...
(What do the other kinds of diversity consist of? Tandem repeats?
Rearrangements? Duplications? Block deletions? Point inserts and
deletes? How do you compare large block changes with point changes on a
fair basis? For example, does a single block of 50 bases that gets
duplicated count the same as 50 separate SNPs, or just 1 or 2 SNPs?)
SNPs are scattered liberally through the genome. While most of them
are found outside genes and probably do not have any effect, those
located in and around genes may contribute to the genetic basis of our
biological individuality, ...
SNPs outside any phenotype-affecting regions (exons, regulatory, etc.)
are neutral, hence drift randomly, and it takes a long time for them to
be fixed one way or the other, so SNPs tend to remain for a long time.
But SNPs inside phenotype-affecting regions are sometimes neutral
(3base->1aa coding synonyms) and sometimes not neutral, and the latter
under selection pressure, so they tend to be rapidly moved to one
extreme or the other i.e. eliminating one allelle and fixing the other.
So statistically, how much more frequent are neutral SNPs than selected
SNPs compared to what you'd expect based on how many places they'd be
neutral vs. selected? I.e. how much is SNP diversity reduced in
selected places compared to neutral places?
For many years, researchers have been aware of a phenomenon called
linkage disequilibrium - the tendency for alleles at separate sites in
the genome (in this case SNP alleles) to be found together more
frequently than would be expected by chance.
Note, as was carefully explained in a Web site I found the other day
<http://linkage.rockefeller.edu/wli/lld.html>
linkage disequilibrium is a historical record, i.e. a state caused by
accumulated history, as opposed to linkage which is merely conditional
probabilities that apply regardless of the initial state. The cause of
linkage disequilibrium is that in the distant past there were only one
allelle each at the two sites, but then a mutation happened at one of
the sites, so now there's one line of descent (at the single-chromosome
level, half of a diploid genome) with that mutation and one line of
descent without it. Then later another mutation happened at the other
site. If it happened in the original line of descent, then there are
now three lines of descent, with each single mutation and also the
parent allelle without either mutation. If the second mutation happened
in the line of descent having the first mutation, then again there are
now three lines of descent, but one with neither mutationn, and one
with both mutations, and one with just the first. Note in both cases
only three of the four possible combinations occur at all, which
already constitutes linkage disequilibrium. Over subsequent
generations, if there are meiotic crossings frequently between those
two sites, then the two independent variations are mixed in all
combinations and the linkage disequilibrium disappears. But if the two
sites are sufficiently closely linked that crossing-over hardly ever
happens between them, then even over moderately long spans of time
(tens of thousands of years) the linkage disequilibrium persists.
If we observe linkage disequilibrium between two SNPs nowadays, that
could mean either that two mutations occurred so recently that there
hasn't been time for crossing-overs between them to homogenize them
back to equilibrium, or that there are no hot spots between them so
that even over long spans of time the ancestral linkage disequilibrium
from the two long-ago mutations has persisted.
If we observe the same two SNPs appearing in many widely separate
native populations which couldn't possibly have all exchanged DNA with
each other recently (assuming we see linkage disequilibrium between
these two SNPs in the first place), that eliminates the recent-mutation
hypothesis, leaving only the no-hot-spot hypothesis. Apparently there
are in fact huge blocks of DNA bases which almost never cross over
during meiosis anywhere within them, not a single crossing in tens of
thousands of years apparently, the "haplotype blocks" being studied.
Most such blocks contain many more than just two SNPs, whereby the
linkage disequilibrium is much more obvious and "proveable".
Recent studies suggest that the genome may be divided into a
remarkably small number of blocks - just 200 000 or so. Recombination
seems to be focused between the haplotype blocks, so large groups of
alleles end up travelling together.
Math: 3 billion total DNA bases, divided by 200 thousand blocks, yields
an average of 15 thousand bases per block. 3 million base pairs
different, 90% of them SNPs, gives 2.7 million SNPs, and divided by 200
thousand blocks, yields an average of 13 or 14 SNPs per block.
That means if each SNP is binary (two different allelles), there would
be a potential for 2**13 to 2**14 different combinations of the
allelles within a single block, hence a potential of appx. ten thousand
different allelles at the block level. As I understand it, only a tiny
fraction of those ten thousand possible combinations actually appear.
If there has been no recombination within a single block in all the
time since those 13 or 14 mutations occurred, then there should be 14
or 15 block-allelles present (the original, plus one extra for each new
mutation that occurred, regardless of whether the new mutation occurred
in a tree of descent where earlier mutations had already occurred or
not). Does anybody know the average number of alelles of such a
haplotype block that has 13 or 14 SNPs within it? Is it like 20
allelles, which means that recombination within that block has happened
only about five times in all of human history, or more like 100
allelles, which means that recombinaton within that block has happened
about 90 times, or even more?
Note that if at the time of the population bottleneck (or shortly
after) only a single allelle of a nowaday haplotype block survived, we
should be able to identify it. Survey all allelles of this block in
populations around the world. If a block-allelle occurs in *all* of
them, it's probably the ancestral block-allelle, whereas if it appears
only in a few populations in just a few geographical regions and/or
along a route of miagration, then the block-allelle originated due to a
new mutation after the bottleneck and didn't have time to be
distributed to all regions of Africa before the first miagration out of
Africa, or it originated *after* the first such miagration. If more
than one globally-distributed block-allelle is found, it indicates that
all such block-alleles were ancestral and survived through the
bottleneck. It's remotely possible that more than one block-allelle
survived the bottleneck and for a while after, but then eventually
drifted to fixation in all the world's populations, but that's highly
unlikely. So if we identify only a single allelle of a particular
block, we can assume it probably was the only survivor of the
bottleneck. If we find a lot of such single-bottleneck-allelle blocks
and hardly any multiple-bottleneck-allelle blocks, it would indicate a
very small population through the bottleneck, a really severe
bottleneck for sure! On the other hand if we discover that lots of
multi-allelle blocks survived the bottleneck, then maybe the bottleneck
was only mild, that a decent-sized population existed at all times
during the bottleneck.
Therefore, rather than the
millions of allele combinations potentially available, the human
population seems to be made up of a more limited set of haplotype
patterns.
That math is not correct. With 2.7 million SNPs, if they were
rearranged in all possible combinations, assuming each has only two
allelles, the total number of combination allelles would be
**HOLD YOUR BREATH** 2 ** (2.7 million) =appx= 10 ** (800 thousand)
not merely "millions" as implied above, not a billion, not a
quadrilion, not even as small as a GOOGOL, more like a GOOGOL to the
eight power!!!!!!!!
But there are only 7 billion total people on Earth, so it'd be
impossible to have more than an infinitesimal number of different
combinations represented in our population. The 7 billion population
would be the bottleneck, reducing the total number of combinations to
no more than 14 billion (two copies of each combination in each
individual, if you somehow decide which of the two copies of each
chromosome is the first copy so you can think of all first-copies of
all the chromosomes as one allelle-combination). It's probably better
to treat both copies of chromosomes together, so then there is only one
combinination of allelles per individual, 7 billion total, out of
GOOGLE to the sixteenth possible combinations.
Now if we break the genome into arbitrary blocks of size 15 thousand
bases each, and calculate number of possible combinations of only the
SNPs located within a single block, we get 2 ** (13 or 14) which is the
ten thousand possible allelles of each block that I calculated earlier.
So even if they said it wrong, "millions of combinations" isn't correct.
This haplotype arrangement appears to be similar in all the different
populations around the world, suggesting that many of them represent
ancestral haplotypes that existed in the earliest humans.
Suppose the bottleneck occurred a million years ago, and the first
major miagration out of Africa occured only 30 thousand years ago. Then
we would expect only about 3% of the mutations to have occurred since
the first miagration, the remaining 97% occured while everyone was
still in Africa. But only a tiny portion of East Africa was occupied by
humans for most of that time, so that small region might have been
homogenized by intermarriage between different local groups many times
throughout most of the pre-miagration period, before the population
started to spread through a larger portion of Africa (just prior to the
first out-of-Africa miagration) and got too spread around for the
different locales to exchange genes with each other. So perhaps 90% of
the mutations occurred before the spreading apart occurred, likewise
90% of the breakages of old blocks to make two new smaller blocks
occurred during that early homogeneous time. SO 90% of the block
structure would be shared across all human populations, with only 10%
of block structure appearing in just part of the human population. So
is that basically what they're saying?
The map will be based on DNA samples obtained from hundreds of people
in geographically distinct populations: Nigerian Yorubas, Han Chinese,
Japanese, and US residents of European origin.
(That seems to have been the original plan. Apparently they later
decided to survey only a small local population of Utah residents for
the US sample. And for Japan, they picked a group in Tokyo.)
A good way to think about haplotypes and haplotype blocks is to
imagine the SNP alleles as children sharing school minibuses.
Oh boy, I'm not the master of metaphor after all!
By the way, I thought I invented the term "haplotype block" myself,
after reading the article in _Scince_, becuase when I started using the
term here people said they had no idea what I meant. But I see the term
was already in use by the HapMap founders before the project even
started, so I claim innocence of the coining.
Additional note: Very few DNA-base-pair-neighbors (appx. 200 thousand,
out of 3.1 billion total) have crossed over even once since the
population bottleneck, most (99.99%) haven't crossed even once during
all that time, yet some very "hot" places have crossed many many times
during the same time. Is that right, or did I miss something? Does
anybody know where I might find online the statistics of hotness of all
the known DNA-base-pair-neighbors
More thoughts: During a time of statis, when the local population stays
the same, each couple on the average having two offspring that survive
to the same point in the generation-cycle, when a neutral mutation
occurs, generating a new allelle of that locus, the average number of
copies of that new allelle remains at 1, which means that it's very
likely to drop to 0 i.e. go extinct. As a result, very very few new
allelles at such times would survive to the present. On the other hand,
during a time of local population growth, a new allelle would have a
good chance of quickly attaining reasonble number of copies, after
which it'd have a decent chance of surviving over long term. Eventually
it'd have p chance of surviving, where p is the proportion of
individuals with that allelle, which is 1/2n where n is the population
size at the time of the mutation in a single individual. But there
hasn't been enough time for drift to play out like that, so there'd
still be a few copies remaining almost for sure. As a result of that
analysis, the SNPs we see today should be of two types, those which
appeared very very recently, like in the current individual or
immediate parents, so there hasn't been time for them to dissappear by
accident, and those which appeared during a time of population growth.
But *now* *is* a time of population growth, so the population-growth
case includes the very-recent case, so we need consider only that one
case.
Now by measuring the allelles of SNPs in lots of different people, we
should be able to observe how widely distributed they are in various
populations in different geographic areas, and from that information we
should be able to estimate when each SNP first occurred due to point
mutation. Likewise by mapping the haplotype blocks, in particular by
measuring how widely distributed crossing-overs at boundaries between
adjacent-block are compared to inheritance of the same ancestral pair
of adjacent blocks, we may be able to estimate how long ago each
crossing-over between two adjacent haplotype blocks occurred. Just like
SNPs, new combinations of haplotype blocks due to crossing between
blocks would tend to disappear except when such cross-overs occurred
during times of population growth. I predict that when we calculate
statistics of age of neutral SNPs and adjacent-haplotype-block-pair
cross-overs, we'll get the same age-distribution within any single
local population group, which will clearly show when various spurts of
population growth occurred, either globally or within local
populations.
Furthermore, if we collect haplotype data for each living human, we may
be able to use both SNP-origin ages and block-boundary cross-over ages
to compute a complete pedigree for each such human, that is a complete
family tree for all of Homo sapiens, going back all the way to the
population bottleneck!! Should I explain how I believe this would be
possible, what specific calculational methodology would accomplish that
task/goal, or can you-all figure it out for yourselves after I have now
spurred you with the idea?
(Final note: I was thinking maybe this one article was getting too
long, so I might split it into parts. But it's only 21k bytes, so what
the heck, I'll keep it intact and see if the auto-moderator accepts
it.)
.
r norman
This is a rather involved explanation of a simple idea. Another
definition is, simply: "A group of gene loci known to be linked; a
chromosome. There are as many linkage groups as there are homologous
pairs of chromosomes."
http://www.biochem.northwestern.edu/holmgren/Glossary/Definitions/Def-L/linkage_group.html
Simply put, two genes (gene loci) are linked if they are on the same
chromosome. If two genes are linked, they do not assort
independently; they show linkage. However, if genes are far enough
apart on the chromosome, the probability of crossover between them is
so high that essentially that there is a 50:50 chance that they assort
together in meiosis. That is, they behave as if they are not linked.
So the definition as the "maximally connected subset". A and B show
linkage, as does B and C as does C and D .... as does Y and Z. But A
and Z are so far apart that you can't measure linkage. Still, they
belong to the linkage group of A, B, C, D, ... Z.
an...@sci.sci
> > No no no, don't pre-judge what the evidence is "for". Just collect
> > evidence like any good police detective, and then after you have
> > all the evidence, see where it leads. For example, if you pre-judge
> > that the husband probably killed his wife, you'll be so busy collecting
> > evidence to "prove" he did it that you not bother to ask neighbors if
> > they saw any one-armed men around the neighborhood, and as a result
> > you'll achieve a false conviction Dr. Richard Kimble.
> Actually, no scientist does anything like this. You don't collect
> evidence at random, and you don't collect evidence to prove any
> particular hypothesis. What you do is collect evidence that bears on a
> family of hypotheses and (we hope) distinguishes among them.
You and I are in agreement, although before you have any hypothesis you
just collect evidence regarding some particular kind of phenomen. After
you have studied some evidence then maybe you are ready to guess
various possible hypotheses, and from that point you may do as you say,
collect evidence to distinguish the various possible hypotheses.
To have the most convincing case against Creationists and IDiots,
especially those leaning that way but not absolutely 100% certain that
evolution is right and YEC or ID is correct, I want to take a
historical approach. For example, first collect *all* evidence from the
fossil record, without any single hypothesis or set of hypotheses in
your mind. Then you look at it and see what it seems to show, namely
some sort of descent (linear chains of related fossils), and then some
sort of common descent (branchings whereby several modern species come
from a single ancestral species).
So part 1 is just to look at the three kinds of evidence:
characteristics of individuals that seem to be clumped into species
that can mate together and clustering of such species at increasingly
larger scales to form a nested hierarchy, fossils that seem to be lined
up in time-consecutive lines with branchings from time to time, and
cladograms from DNA sequences that actually can be computed and are to
an unexpected degree stable unlike what you'd expect if the genomes
were randomly generated. Then part 2 is to ask how to explain that
evidence. At this point we have a prima facie case in favor of
evolution and common descent, regardless of whatever mechanism might
have caused it. I'm trying to show that if you just look at the
evidence, with no pre-conceived idea what to expect, you come up with
evolution and common descent. I leave for later the exhaustive testing,
trying to find violations of the theories, failing to find any, and
thereby getting more and more confident of the correctness of the
theory. I'm just trying to collect a nice organization of the evidence,
to ask who accepts it and who disputes it, and then for whose who agree
to stipulate the evidence is correct, a nice organization of the
theories of evolution and common descent, and ask them whether they
accept those conclusions. For those who accept all the evidence, and
evoution and common descent, then I ask them what theory, other than
mutation + replication + selection/drift, they have to explain the
cause of evolution and common descent.
I really want a check-list of evidence, accept or deny, and then a
check-list of theory, accept or deny, so we can pin down where exactly
the others disagree with modern evolutionary fact or theory.
> Anyway, I'm not pre-judging. I'm post-judging.
Post-judging what? The topic of this thread (you dispute either the
evidence or the theory to explain the evidence)? Or my style of
debating the anti-evolutionists (you don't like my idea of asking them
whether they dispute the evidence, and then if they accept the evidence
askign whether they dispute the theory)? Or what??
> I don't think we can actually identify such fossil chains. What we have
> are trees. There are various statistical tests you can do that will
> determine the degree of "strain" between the tree and stratigraphic
> order. When applied to the fossil record, these tests commonly show a
> much better than random fit. That's about as close as actual science
> gets to these chains of yours.
Do you know of any Web page showing a complete worked out example for
one "kind" of animal ("kind" in the naive sense, such as horses or dogs
or cats etc.), whereby there's a description of each specific character
being compared, with a criterion for assigning numeric values to each
character (such as length of rear femur measured in centimeters, or
curvature of spine between shoulders and rump measured somehow, or
interpolation between claws and hoofs), and then a complete matrix
showing the value (mean and standard deviation) for each character for
each species? Then that matrix is fed to such-and-such computer program
which generates the best-fit tree, if any, and shows a measure of
fitness such as a P (confidence) value for tree-ness (and/or that
specific tree model) preferred over the null hypothesis of random
number input not belonging to any tree? It would be nice to
have two such examples, one very small, such as apes, only five or ten
species, such that the whole thing could be processed by hand or at
least the final result verified by hand, and one much larger to show
how really interesting trees can be demonstrated.
> > When we fold together all the fossil chains, with the branch points
> > that replace one ancestral fossil chain with two or more new fossil
> > chains, we have a forest of trees. These trees are consistent with a
> > single tree where lots of links are missing from our evidence so-far.
> That's not how it actually works. The trees we build all have real taxa
> only at the tips of the branches. There are no linear "chains", because
> we have no way to tell if one species is ancestral to another.
If within a single geographic area there's an isolated species that
lasts a few million years, then immediately afterward there's another
species only slightly different from the previous but clearly showing a
significant change from the previous, and it's in the same area as the
first, we can reasonably guess that the second descends from the first.
Are there no such examples whatsoever in the fossil record to show
blatantly apparent descent? Alternately, consider a marine species that
is distributed worldwide, followed by a similar worldwide marine
species, no examples whatsoever of that ever in the fossil record? I'm
saying if there are even a few such examples, they can establish the
prima facie suspicion that lines of descent might be apparent, and then
we can test that hypothesis to see whether there are other
not-so-obvious lines of descent. But we need some really blatantly
obvious lines of descent so that it makes it reasonble to even conceive
such a hypothesis.
Now Darwin's finches wouldn't be like that. He observed a whole bunch
of species, which all seemed be be variants upon a common theme, and he
guessed that they all descended from a common ancestor, but his "tree"
had a single level wherein all modern species suddenly appeared in
parallel, no successive splittings over time that he could discern
without any fossils to look at. But surely among the actual fossils
back on the mainland around the world, there might have been examples
of apparent chains of descent?? Surely Lamarck knew of such when he
formulated his linear evolution theory??
> > We can state strongly that there are fossil chains, and branch points,
> > yielding lots of small common-ancestry trees. We can guess with less
> > assurance that they all belong to one universal common-ancestry tree.
> If what you're trying to say is that not all nodes are resolved, then
> yes. But we don't in fact end up with lots of separate trees. We end up
> with one tree some of whose nodes are unresolved.
I believe such trees would show low levels of confidence for some
branchings, expecially pre-Cambrian, and if the program is forced to
exclude all low-confidence parts of the tree, it would yield a bunch of
separate trees instead of just a single tree. (Some programs would
simply refuse to give any result at all. I'm suggesting a program that
would break apart low-confidence trees to achieve a forest of
high-confidence trees instead.)
By the way, by "unresolved", you mean three or more subs in a rooted
tree, or four or more branches in an unrooted tree, can't be resolved
into a binary-rooted-tree or ternary-unrooted tree respectively, right?
Note I'm assuming that if you have only a bunch of current-day species,
with no information about past species, such as when making DNA
cladograms, or when including embryiological data which you have only
for present-day organisms, you generate unrooted trees, but if you have
time-dated species, such as when dealing with the fossil record, you
can then generate rooted trees, right?
Here's an interesting data-analysis experiment: Use all the fossil
evidence, but discard all time and place data, include *only* the
characters of the various species, and use that data to generate an
unrooted cladogram. Then check if the that unrooted tree matches the
rooted tree that you got when you included time-and-place data from the
outset. If they indeed do match, that's refutation of ID, because a
supernatural Designer wasn't restricted to placing designed species in
a natural-looking chronological sequence, He could have designed a
Jurassic species, kept the design in his Mind, then designed five
different variations on it for the Cambrian, then manufactured those
five when the Cambrian time came, then waited for the Jurassic and
manufactured that ancestor species at that time, no need for ancestral
designs to be manufactured before descendent designs. Even computer
software designers sometimes do that: Design a primitive program, but
not release it to the public, then design variations on it, and release
those variations, then later finally release the prequel for those who
have been asking for it. Some TV programs or movies are like that too,
releasing the main movie before the prequel (although in those cases
the prequel was designed after the main movie, but they were
retro-evolved to make them look like ancestral movies).
> > We can state with intermediate assurance that we have a single
> > common-ancestry tree for each phylum, and sometimes for several phyla
> > together, from the Cambrian explosion to the present. The case for
> > joining all those trees into one universal tree before the Cambrian
> > explosion is not firmly established by fossil evidence alone.
> True. Were we limiting ourself to fossil evidence? I was unaware.
What, you're recanting your claim that we *always* get a single tree?
Yes, I'm treating the three lines of evidence (modern taxonomy,
historical fossil, and DNA cladograms) as three independent lines of
evidence at first. For fossils, we get a rooted tree or forest of trees
of which the largest start just before the Cambrian. For each of the
other two we get an unrooted tree with a bunch of unresolved nodes or
even a forest of disconnected unrooted trees. The fact that we can get
most of the data fitted into high-confidence trees is powerful evidence
already. *then* we compare those three trees (or forests of trees) to
see whether they are approximately mutually consistent), and the fact
that they are indeed is really powerful evidence. The two sets of
unrooted trees merely show there's hierarchial clustering of
characteristics, a Platonic sort of thing, but the one set of rooted
fossil trees shows some actual chronological pattern of descent with
modification (either actual descent of living species, whereby species
begats species, or descent of Intelligently-Designed productions where
design begats design and each design is released into production
separately, like the evolution of MicroSoft Windows software designs).
So indeed for the one line of evidence from fossils, indeed at first
I'm limiting the evidence *only* to fossils, nothing else. That's why
the trees are disconnected before the Cambrian.
When we later try to merge the disconnected rooted trees (from fossil
evidence) with DNA evidence, we then hope to fit those various rooted
fossil-trees onto branches of the DNA unrooted trees, and then use the
central parts of the DNA trees to link the fossil trees together into a
single unrooted tree.
> > This is where we need the other two lines of evidence (comparative
> > embryology, and DNA cladograms) to piece together the separate trees
> > into a single universal tree. And last I heard the best we could do is
> > establish three trees for the three major domains, with unknown linkage
> > earlier than that.
(Let me clarify that slightly: When drawing whole-cell-species trees,
at each endosymbiosis event I expect to start a new tree, since it
doesn't make sense to fit a combination of separate symbionts into a
single ancestor. If we ignore plastids however, we may be able to get
by with just two trees, one for prokaryotes and one for eukaryotes. Or
if we ignore the endo-horizontal gene flow from mitochondria to
nucleus, pretend like the DNA just magically appeared out of nowhere
when it first got into the nucleus, and only the nuclear DNA is
calculated in our cladograms, then we might get a single unrooted tree
for both prokaryotes and eukaryotes. I've seen published reports where
they seem to have done that last thing, but failed to make that clear,
so their report seemed poor-quality to me.)
> The problem is in defining the root.
For unrooted trees, such as cladograms of modern DNA or cladograms of
modern characters including embryology, there's never a root, so who
cares? When we fit rooted trees from fossils in with unrooted trees
from cladograms (both kinds), we lose the roots. We have directed links
where pairs of dated fossils show the direction, but non-directed links
where there are no dated fossils, so the central part of the unrooted
tree consists of non-directed links while the rest is all directed
links.
> But don't confuse the lack of resolution with a lack of a tree. The
> three domains are clearly united by a host of characters. The questions
> are which two are more closely related, and if a branching tree is in
> fact the best model.
Yes, we agree about this. In particular I actually am leaning against a
whole-cell (or whole-genome) tree-system of heredity for the very early
eras when the three domains originally came to be. I think horizontal
gene flow and/or frequent merging of genomes to yield new composite
genomes (formation of endosymbiosis by merging of formerly separate
genomes) may have been very common way back then. Two strains of
RNA-world pre-prokaryote might have separately evolved different
aspects of biochemistry, such as one of them evolving a way to make DNA
bases and the other evolving an efficient form of sulfur respiration,
or one of them producing formate and the other consuming it, and
initially they survive as exo-symbionts via external exchange of
chemical products, and later the two merge to form a single strain that
performs both types of chemosynthesis within a single cell. It's
possible that all such genomic merging occurred before the first
DNA-world true-prokaryote happened, but it may also be possible that
exo-symbiotic biochemistry webs persisted right into the age of DNA
prokaryotes and only later a bunch of additional merging of genomes
happened, yielding five or ten different domains, of which three were
so much more fit/successful than the others that they drove the others
to extinction.
> ...
> > Accordingly *any* theory of evolution (to fit the evidence on Earth)
> > *must* involve some kind of selection, hence any *scientific* theory of
> > evolution must include stochastic selection.
> I don't think so. For one thing, "stochastic selection" is a terrible
> term, conflating drift and selection.
But as somebody else pointed out, drift and selection *are* already
conflated in fact. They are caused by the same mechanism: Differential
survival rates among different allelles, which *may* be partly caused
by some difference in fitness or may be caused entirely by random
events unrelated to any difference of fitness. (Or as you point out
below, with supernatural intervention, all random elements whatsoever
might be eliminated, and the differential survival rate might be
*entirely* caused by differential fitness in the Mind of the
Intervener.) In most cases, both difference in fitness (not necesarily
in the allele under consideration, but possibly in nearby genes which
are linked to the allelle under consideration) and randomness are
involved in the differential survival, not just statistically, but for
each and every life/death decision which occurs. An individual might
die tonight because its cells lacked the ability to synthesize citric
acid, and by chance it wasn't able to consume any food containing
citric acid for a while, and its stockpile ran out, and Kreb's cycle
slowed down so much that the cell exhausted its ATP supply and could no
longer perform basic functions of life. How can you say that death was
*entirely* 100% due to the faulty citric-acid pathway, or it was
*entirely* 100% due to the chance lack of citric-acid-containing food?
I say that both the differential fitness and the bad luck in the food
department caused the death. Virtually all deaths are like that,
combination of bad genome and bad luck. Even Bambi being stepped-on by
Gozilla could have been avoided if Bambi had a genotype giving it the
ability to sense the approaching Godzilla and take cover and avoid
emitting any smell that would tell Godzilla where it was hiding.
Fact: Differential survival is conflation of genome and chance.
Math: It's possible to calculate separate drift and selection factors.
Truth: That math is just a calculation, not evidence that two separate
mechanisms exist in nature for drift and selection.
> For another, a scientific theory can include supernatural elements.
> All that's required is that these events be systematic. We can imagine
> a system in which there is no stochastic element involved in fixation
> of new alleles, i.e. in which the designer stacks the deck, perhaps
> causing enough simultaneous, identical mutations in a population to
> make fixation assured. This is in principle testable.
I agree. Do the math to separate the selection and the randomness, and
if the randomness equals zero, whooptie doo.
> > and the DNA cladograms can be shown live in the
> > laboratory so only a fool would dismiss them.
> I don't know. But it's all just the nested hierarchy under another name,
> and I expect the same responses: it exists but it's explained by
> goddidit, not common descent, or it doesn't really exist.
That's why I want to clearly organize the various kinds of evidence,
and the various elements of the theoretical explanation for the
evidence, and ask each individual person which items he/she accepts or
disputes. Once they are pinned down as to where they disagree with our
chain of evidence and logic, we can work on discussing thet particular
breakdown-point with them. Maybe they'll learn they were overlooking
something, and they convert to our logic. Or maybe we'll discover
there's a flaw in our logic and our case isn't yet proven.
.
an...@sci.sci
> that fitness is a continuous function of whatever is on the axes of
> variation. These are not continuous parameters and subtle changes in
> regulatory genes can produce rather abrupt and discontinuous changes
> in phenotype, hence in fitness.
I suppose the rationale is that most non-fatal mutations are really
tiny steps, and a continuous landscape is a reasonable mathematical
model for a descrete but very finely-divided staircase. Really huge
regulatory changes usually screw up the overall interlocked system so
badly as to be virtually always fatal. Or such changes make one
individual totally incompatible with all other members of what was his
species, so even if he can survive, he can't produce offspring. So most
of the time giant steps can be ignored as a course of evolution. But
all it takes is one such giant mutation to succeed every few million
years, something making an allelle very much more successful than the
ancestral allelle, yet not mating-incompatible, and suddenly there's a
big evolutionary jump. So maybe the correct metaphor is a fitness
landscape with the ability of individuals to shout "Geronimo" and jump
into a strong breeze with a hang-glider and take their chances where
they might land on a new island far from home.
> Still, the fact is that evolution is always about populations and
> never about individuals. Selection pressure changes the allele
> frequency in a population.
You're seeing only half the picture of evolution, the half whereby
already-exiting variation is tested and only the best is retained.
You're overlooking the other half of evolution, whereby one individual
suffers a mutation whch changes the fitness of that individual and of
any descendents that individual might begat, thereby adding new
variation to the population. That individual's mutation, and the
mutations within other individual, are essential for evolution to
produce new characters rather than merely narrow down the whole world's
life to a single genotype.
The anti-evolutionists, who claim evolution can only purify the race,
never produce a new race, make the same omission you made there. You
are playing into their hands, giving them quotes to mine.
.
an...@sci.sci
> exceeds then or so. These threads just go on and on and the nesting
> level increases without limit.
First, if your newsreader is broken, switch to a different one that works.
Or don't use a newsreader at all. Do all your replies manually as I do.
Second, your "bug report" is too vague. When you refer to "nesting
level", are you referring to the actual number of message-IDs in the
Received line of the single article you are looking at, or are you
referring to the entire path upward by chaining Received lines upward
all the way to the start of the thread? Thinking of the various
articles cited in the Received line as if they were linkage between
loci, are you referring to direct linkage or to "linkage groups" which
are maximal chains of single linkages?? And what "problems" does your
broken newsreader exhibit? Obviously it doesn't crash totally, or you
wouldn't have been able to see my article at all. Does it allow you to
see the full header, whereon you could copy&paste the message IDs
manually yourself? Does your computer system allow you to run your
newsreader and a Web browser simultaneously? If so, why don't you use
that capability to ask Google Groups for any particular article whose
message ID you see in my article?
Third, just about every newsreader has a way to "killfile", i.e. ignore
articles written by any particular author you don't like. So if you
really don't like my posting style, then you should do that. I use
Google Groups to browse newsgroups, so I don't have the option of
killfiling you so I don't have to see your inane complaints about bugs
in your newsreader which you are blaiming on me.
.
r norman
>> I never cared for the "fitness landscape" metaphor because it assumes
>> that fitness is a continuous function of whatever is on the axes of
>> variation. These are not continuous parameters and subtle changes in
>> regulatory genes can produce rather abrupt and discontinuous changes
>> in phenotype, hence in fitness.
>
>I suppose the rationale is that most non-fatal mutations are really
>tiny steps, and a continuous landscape is a reasonable mathematical
>model for a descrete but very finely-divided staircase. Really huge
>regulatory changes usually screw up the overall interlocked system so
>badly as to be virtually always fatal. Or such changes make one
>individual totally incompatible with all other members of what was his
>species, so even if he can survive, he can't produce offspring. So most
>of the time giant steps can be ignored as a course of evolution. But
>all it takes is one such giant mutation to succeed every few million
>years, something making an allelle very much more successful than the
>ancestral allelle, yet not mating-incompatible, and suddenly there's a
>big evolutionary jump. So maybe the correct metaphor is a fitness
>landscape with the ability of individuals to shout "Geronimo" and jump
>into a strong breeze with a hang-glider and take their chances where
>they might land on a new island far from home.
Those fatal mutations are also quite striking discontinuities in the
landscape.
>> Still, the fact is that evolution is always about populations and
>> never about individuals. Selection pressure changes the allele
>> frequency in a population.
>
>You're seeing only half the picture of evolution, the half whereby
>already-exiting variation is tested and only the best is retained.
>You're overlooking the other half of evolution, whereby one individual
>suffers a mutation whch changes the fitness of that individual and of
>any descendents that individual might begat, thereby adding new
>variation to the population. That individual's mutation, and the
>mutations within other individual, are essential for evolution to
>produce new characters rather than merely narrow down the whole world's
>life to a single genotype.
>
>The anti-evolutionists, who claim evolution can only purify the race,
>never produce a new race, make the same omission you made there. You
>are playing into their hands, giving them quotes to mine.
>.
Selection pressure, a term that you disparage, occurs only after
mutation has introduced the new allele into the mix. Nothing in my
statement even suggests that I am listing all the factors behind
evolution. I am only describing one of them. I was also making the
extremely important point that you seem to have missed in earlier
posts, and whose significance you snipped away, about the population
nature of evolution.
r norman
<plonk>
John Harshman
>>>>Evidence for what, precisely? This seems like evidence for common descent.
>>>
>>>No no no, don't pre-judge what the evidence is "for". Just collect
>>>evidence like any good police detective, and then after you have
>>>all the evidence, see where it leads. For example, if you pre-judge
>>>that the husband probably killed his wife, you'll be so busy collecting
>>>evidence to "prove" he did it that you not bother to ask neighbors if
>>>they saw any one-armed men around the neighborhood, and as a result
>>>you'll achieve a false conviction Dr. Richard Kimble.
>>
>>Actually, no scientist does anything like this. You don't collect
>>evidence at random, and you don't collect evidence to prove any
>>particular hypothesis. What you do is collect evidence that bears on a
>>family of hypotheses and (we hope) distinguishes among them.
>
>
> You and I are in agreement, although before you have any hypothesis you
> just collect evidence regarding some particular kind of phenomen.
This is very seldom done. I can't really think of an example right off.
You almost always start with some kind of theory. That theory may be
based on evidence collected for some other purpose, by someone else. But
you never really start in a vacuum.
> After
> you have studied some evidence then maybe you are ready to guess
> various possible hypotheses, and from that point you may do as you say,
> collect evidence to distinguish the various possible hypotheses.
>
> To have the most convincing case against Creationists and IDiots,
> especially those leaning that way but not absolutely 100% certain that
> evolution is right and YEC or ID is correct, I want to take a
> historical approach. For example, first collect *all* evidence from the
> fossil record, without any single hypothesis or set of hypotheses in
> your mind. Then you look at it and see what it seems to show, namely
> some sort of descent (linear chains of related fossils), and then some
> sort of common descent (branchings whereby several modern species come
> from a single ancestral species).
Feel free to do whatever you want. But that's not how science is done.
Also, both of your chosen forms of evidence seem to rely on identifying
actual fossils as ancesors, which is just impossible. Instead you should
be thinking about nested hierarchy.
> So part 1 is just to look at the three kinds of evidence:
> characteristics of individuals that seem to be clumped into species
> that can mate together and clustering of such species at increasingly
> larger scales to form a nested hierarchy, fossils that seem to be lined
> up in time-consecutive lines with branchings from time to time, and
> cladograms from DNA sequences that actually can be computed and are to
> an unexpected degree stable unlike what you'd expect if the genomes
> were randomly generated.
It's probably not a good idea to think of these as separate. The first
and third especially seem like just two ways of saying the same thing.
The middle one is a garbled form of something that paleontologists
sometimes do, which is to test the stratigraphic fit of their
cladograms. It usually comes out as highly non-random despite the patchy
nature of the record.
> Then part 2 is to ask how to explain that
> evidence. At this point we have a prima facie case in favor of
> evolution and common descent, regardless of whatever mechanism might
> have caused it. I'm trying to show that if you just look at the
> evidence, with no pre-conceived idea what to expect, you come up with
> evolution and common descent.
That's pretty easy to do. Several people here and in the TO FAQs have
done just that. But you have a strange way to go about it. As I recall
we began this argument because you were trying to use the evidence above
as some kind of argument for natural selection, not just common descent.
> I leave for later the exhaustive testing,
> trying to find violations of the theories, failing to find any, and
> thereby getting more and more confident of the correctness of the
> theory. I'm just trying to collect a nice organization of the evidence,
> to ask who accepts it and who disputes it, and then for whose who agree
> to stipulate the evidence is correct, a nice organization of the
> theories of evolution and common descent, and ask them whether they
> accept those conclusions. For those who accept all the evidence, and
> evoution and common descent, then I ask them what theory, other than
> mutation + replication + selection/drift, they have to explain the
> cause of evolution and common descent.
You have a strange way of going about it.
> I really want a check-list of evidence, accept or deny, and then a
> check-list of theory, accept or deny, so we can pin down where exactly
> the others disagree with modern evolutionary fact or theory.
Good luck on that. I've had trouble getting creationists to admit what
they believe, especially in any consistent fashion, or to accept the
consequences of what they believe.
>>Anyway, I'm not pre-judging. I'm post-judging.
>
> Post-judging what? The topic of this thread (you dispute either the
> evidence or the theory to explain the evidence)? Or my style of
> debating the anti-evolutionists (you don't like my idea of asking them
> whether they dispute the evidence, and then if they accept the evidence
> askign whether they dispute the theory)? Or what??
Lost context is probably the reason for your confusion here. If you
hadn't snipped everything, you would be able to determine from context
just what I'm post-judging. As I recall, I was judging what the sorts of
evidence you were talking about were capable of telling us.
>>I don't think we can actually identify such fossil chains. What we have
>>are trees. There are various statistical tests you can do that will
>>determine the degree of "strain" between the tree and stratigraphic
>>order. When applied to the fossil record, these tests commonly show a
>>much better than random fit. That's about as close as actual science
>>gets to these chains of yours.
>
> Do you know of any Web page showing a complete worked out example for
> one "kind" of animal ("kind" in the naive sense, such as horses or dogs
> or cats etc.), whereby there's a description of each specific character
> being compared, with a criterion for assigning numeric values to each
> character (such as length of rear femur measured in centimeters, or
> curvature of spine between shoulders and rump measured somehow, or
> interpolation between claws and hoofs), and then a complete matrix
> showing the value (mean and standard deviation) for each character for
> each species? Then that matrix is fed to such-and-such computer program
> which generates the best-fit tree, if any, and shows a measure of
> fitness
Better call that "fit"; fitness has a different meaning.
> such as a P (confidence) value for tree-ness (and/or that
> specific tree model) preferred over the null hypothesis of random
> number input not belonging to any tree? It would be nice to
> have two such examples, one very small, such as apes, only five or ten
> species, such that the whole thing could be processed by hand or at
> least the final result verified by hand, and one much larger to show
> how really interesting trees can be demonstrated.
Only the one I did myself with mtDNA sequences. Did you read that post?
I've put it up many times. You will find something of the sort in
hundreds of scientific papers, but they never test a tree against
separate creation. However, we can suppose that any strong, nested
hierarchy is good evidence against the absence of a tree, and those are
plentiful in the scientific literature.
>>>When we fold together all the fossil chains, with the branch points
>>>that replace one ancestral fossil chain with two or more new fossil
>>>chains, we have a forest of trees. These trees are consistent with a
>>>single tree where lots of links are missing from our evidence so-far.
>>
>>That's not how it actually works. The trees we build all have real taxa
>>only at the tips of the branches. There are no linear "chains", because
>>we have no way to tell if one species is ancestral to another.
>
> If within a single geographic area there's an isolated species that
> lasts a few million years, then immediately afterward there's another
> species only slightly different from the previous but clearly showing a
> significant change from the previous, and it's in the same area as the
> first, we can reasonably guess that the second descends from the first.
No, we can't. We could if we could be confident that there had been no
movement between unsampled geographic areas, and that we had sampled all
species in that particular area, and that we could actually recognize
which individuals belonged to the same or different species. But we
really can't do any of those things.
> Are there no such examples whatsoever in the fossil record to show
> blatantly apparent descent?
Not to my knowledge. But feel free to propose one.
> Alternately, consider a marine species that
> is distributed worldwide, followed by a similar worldwide marine
> species, no examples whatsoever of that ever in the fossil record?
Not to my knowledge. There are certainly many with quite wide distribution.
> I'm
> saying if there are even a few such examples, they can establish the
> prima facie suspicion that lines of descent might be apparent, and then
> we can test that hypothesis to see whether there are other
> not-so-obvious lines of descent. But we need some really blatantly
> obvious lines of descent so that it makes it reasonble to even conceive
> such a hypothesis.
No, we don't. That good old nested hierarchy works fine. Also, a general
pattern of change in world biota (life of the Ordovician different from
Cambrian, Silurian different from Ordovician, etc.) argues for evolution
too, even without a hierarchy.
> Now Darwin's finches wouldn't be like that. He observed a whole bunch
> of species, which all seemed be be variants upon a common theme, and he
> guessed that they all descended from a common ancestor, but his "tree"
> had a single level wherein all modern species suddenly appeared in
> parallel, no successive splittings over time that he could discern
> without any fossils to look at.
In fact Darwin didn't even propose a tree for Darwin's finches, just
common ancestry for the lot. However, we don't need fossils to determine
this successive splitting. There are several papers on the phylogeny of
Darwin's finches, the most recent using DNA sequences.
> But surely among the actual fossils
> back on the mainland around the world, there might have been examples
> of apparent chains of descent?? Surely Lamarck knew of such when he
> formulated his linear evolution theory??
I don't know whether Lamarck used fossil evidence of descent. But I
don't know of any "chains", just trees. You can turn trees into chains
by ignoring some of the branches, if you really want a chain. But the
assumptions are unwarranted and unnecessary.
>>>We can state strongly that there are fossil chains, and branch points,
>>>yielding lots of small common-ancestry trees. We can guess with less
>>>assurance that they all belong to one universal common-ancestry tree.
>>
>>If what you're trying to say is that not all nodes are resolved, then
>>yes. But we don't in fact end up with lots of separate trees. We end up
>>with one tree some of whose nodes are unresolved.
>
> I believe such trees would show low levels of confidence for some
> branchings, expecially pre-Cambrian, and if the program is forced to
> exclude all low-confidence parts of the tree, it would yield a bunch of
> separate trees instead of just a single tree. (Some programs would
> simply refuse to give any result at all.
Not true. I only know of one such program, and I doubt sincerely that
you have ever heard of it.
> I'm suggesting a program that
> would break apart low-confidence trees to achieve a forest of
> high-confidence trees instead.)
That's not what we would get. The node of all eukaryotes, for example,
is a very high confidence one. So whatever lack of resolution there may
be above that node, everything must come together there.
> By the way, by "unresolved", you mean three or more subs in a rooted
> tree, or four or more branches in an unrooted tree, can't be resolved
> into a binary-rooted-tree or ternary-unrooted tree respectively, right?
Yes. Or to put it more simply, a node at which 4 or more branches come
together. All fully resolved nodes have exactly 3 branches. (In the
rooted tree, that's one ancestral branch and two descendants.)
> Note I'm assuming that if you have only a bunch of current-day species,
> with no information about past species, such as when making DNA
> cladograms, or when including embryiological data which you have only
> for present-day organisms, you generate unrooted trees, but if you have
> time-dated species, such as when dealing with the fossil record, you
> can then generate rooted trees, right?
No. Trees can be rooted in many ways, and we seldom if ever, even with
fossils, use stratigraphic position to root a tree. The most common
method is by outgroup. This of course requires the assumption that some
particular species is outside the clade of interest.
> Here's an interesting data-analysis experiment: Use all the fossil
> evidence, but discard all time and place data, include *only* the
> characters of the various species, and use that data to generate an
> unrooted cladogram. Then check if the that unrooted tree matches the
> rooted tree that you got when you included time-and-place data from the
> outset. If they indeed do match, that's refutation of ID, because a
> supernatural Designer wasn't restricted to placing designed species in
> a natural-looking chronological sequence, He could have designed a
> Jurassic species, kept the design in his Mind, then designed five
> different variations on it for the Cambrian, then manufactured those
> five when the Cambrian time came, then waited for the Jurassic and
> manufactured that ancestor species at that time, no need for ancestral
> designs to be manufactured before descendent designs.
You have conflated ID, which at least claims to be a theory about
mechanism, with separate creation. This would be evidence against
separate creation, not against ID.
Something like what you have proposed here (though quite different in
detail) is actually done by paleontologists. They assess the strain
between trees constructed without stratigraphic data and the
stratigraphic data themselves. There are a number of different tests.
There are also a couple of methods that allow incorporation of
stratigraphic data into tree-building, but they are rarely used. If
you're interested, look up stratocladistics.
> Even computer
> software designers sometimes do that: Design a primitive program, but
> not release it to the public, then design variations on it, and release
> those variations, then later finally release the prequel for those who
> have been asking for it. Some TV programs or movies are like that too,
> releasing the main movie before the prequel (although in those cases
> the prequel was designed after the main movie, but they were
> retro-evolved to make them look like ancestral movies).
>
>
>>>We can state with intermediate assurance that we have a single
>>>common-ancestry tree for each phylum, and sometimes for several phyla
>>>together, from the Cambrian explosion to the present. The case for
>>>joining all those trees into one universal tree before the Cambrian
>>>explosion is not firmly established by fossil evidence alone.
>>
>>True. Were we limiting ourself to fossil evidence? I was unaware.
>
> What, you're recanting your claim that we *always* get a single tree?
If we limit ourselves to fossils, that single tree is not so clear or so
strongly supported. I'm not really sure what characters you would even
use that you could score across all metazoan fossils.
> Yes, I'm treating the three lines of evidence (modern taxonomy,
> historical fossil, and DNA cladograms) as three independent lines of
> evidence at first. For fossils, we get a rooted tree or forest of trees
> of which the largest start just before the Cambrian. For each of the
> other two we get an unrooted tree with a bunch of unresolved nodes or
> even a forest of disconnected unrooted trees.
No, not true. You are confused about rooting. There is also no reason to
separate any of these. Are you proposing that analyses of morphology
should include fossils only, or living species only? Are you proposing
that no combined analyses of morphology and molecules should be done?
And your (apparent) distinction between morphological and molecular data
is artificial. There is no "twin nested hierarchy". And independently
assembled data set is testable against any other: one gene vs. another
gene, one part of the body vs. another part, etc.
> The fact that we can get
> most of the data fitted into high-confidence trees is powerful evidence
> already. *then* we compare those three trees (or forests of trees) to
> see whether they are approximately mutually consistent), and the fact
> that they are indeed is really powerful evidence.
> The two sets of
> unrooted trees merely show there's hierarchial clustering of
> characteristics, a Platonic sort of thing,
Not so. In fact (and ignoring your idea of rooting), the nested
hierarchy itself is the most powerful evidence of common descent. Nobody
has ever been able to come up with an alternative explanation. Try it
yourself. Attempt to flesh out your "Platonic sort of thing" into an
explanation that holds up.
> but the one set of rooted
> fossil trees shows some actual chronological pattern of descent with
> modification (either actual descent of living species, whereby species
> begats species, or descent of Intelligently-Designed productions where
> design begats design and each design is released into production
> separately, like the evolution of MicroSoft Windows software designs).
No, it doesn't. Like I said, we can't actually see the begats. We can
only infer from the nested hierarchy that they have happened. However,
stratigraphic fit is a separate criterion that we can use to evaluate
the evidence, so you're not entirely off base. There's a germ of a real
idea there, though you state it confusingly.
> So indeed for the one line of evidence from fossils, indeed at first
> I'm limiting the evidence *only* to fossils, nothing else. That's why
> the trees are disconnected before the Cambrian.
> When we later try to merge the disconnected rooted trees (from fossil
> evidence) with DNA evidence, we then hope to fit those various rooted
> fossil-trees onto branches of the DNA unrooted trees, and then use the
> central parts of the DNA trees to link the fossil trees together into a
> single unrooted tree.
Again, this is not the way it's done, or the way it should be done, or
even a way it could be done.
>>>This is where we need the other two lines of evidence (comparative
>>>embryology, and DNA cladograms) to piece together the separate trees
>>>into a single universal tree. And last I heard the best we could do is
>>>establish three trees for the three major domains, with unknown linkage
>>>earlier than that.
>
> (Let me clarify that slightly: When drawing whole-cell-species trees,
> at each endosymbiosis event I expect to start a new tree, since it
> doesn't make sense to fit a combination of separate symbionts into a
> single ancestor. If we ignore plastids however, we may be able to get
> by with just two trees, one for prokaryotes and one for eukaryotes. Or
> if we ignore the endo-horizontal gene flow from mitochondria to
> nucleus, pretend like the DNA just magically appeared out of nowhere
> when it first got into the nucleus, and only the nuclear DNA is
> calculated in our cladograms, then we might get a single unrooted tree
> for both prokaryotes and eukaryotes. I've seen published reports where
> they seem to have done that last thing, but failed to make that clear,
> so their report seemed poor-quality to me.)
!
>>The problem is in defining the root.
>
> For unrooted trees, such as cladograms of modern DNA or cladograms of
> modern characters including embryology, there's never a root, so who
> cares?
Like I have said already, this is just plain wrong.
> When we fit rooted trees from fossils in with unrooted trees
> from cladograms (both kinds), we lose the roots. We have directed links
> where pairs of dated fossils show the direction, but non-directed links
> where there are no dated fossils, so the central part of the unrooted
> tree consists of non-directed links while the rest is all directed
> links.
And this isn't even wrong.
>>But don't confuse the lack of resolution with a lack of a tree. The
>>three domains are clearly united by a host of characters. The questions
>>are which two are more closely related, and if a branching tree is in
>>fact the best model.
>
> Yes, we agree about this. In particular I actually am leaning against a
> whole-cell (or whole-genome) tree-system of heredity for the very early
> eras when the three domains originally came to be. I think horizontal
> gene flow and/or frequent merging of genomes to yield new composite
> genomes (formation of endosymbiosis by merging of formerly separate
> genomes) may have been very common way back then. Two strains of
> RNA-world pre-prokaryote might have separately evolved different
> aspects of biochemistry, such as one of them evolving a way to make DNA
> bases and the other evolving an efficient form of sulfur respiration,
> or one of them producing formate and the other consuming it, and
> initially they survive as exo-symbionts via external exchange of
> chemical products, and later the two merge to form a single strain that
> performs both types of chemosynthesis within a single cell. It's
> possible that all such genomic merging occurred before the first
> DNA-world true-prokaryote happened, but it may also be possible that
> exo-symbiotic biochemistry webs persisted right into the age of DNA
> prokaryotes and only later a bunch of additional merging of genomes
> happened, yielding five or ten different domains, of which three were
> so much more fit/successful than the others that they drove the others
> to extinction.
I think you have confused all manner of early events here. RNA world
comes way before any of the prospective symbioses you speculate about
here. It's also very unlikely that any two prokaryotes actually merged.
Endosymbiosis would require that one cell incorporate the other, and
that wouldn't happen until the invention of endocytosis by
proto-eukaryotes. What you can have is exhange of genes by plasmid or
viral transmission.
>>>Accordingly *any* theory of evolution (to fit the evidence on Earth)
>>>*must* involve some kind of selection, hence any *scientific* theory of
>>>evolution must include stochastic selection.
>>
>>I don't think so. For one thing, "stochastic selection" is a terrible
>>term, conflating drift and selection.
>
> But as somebody else pointed out, drift and selection *are* already
> conflated in fact.
No, they aren't. They are entangled, if you like, which is quite a
different thing. Using terms that confuse two different mechanisms is
just, well, confusing.
> They are caused by the same mechanism: Differential
> survival rates among different allelles,
That's not a mechanism. It's a result. Selection and drift are
mechanisms. Or if you want to think of it another way, being eaten by a
frog is a mechanism. Any affect your phenotype has on the chances that
this does or does not happen is selection. Other effects (being in the
wrong place at the wrong time, say) are drift.
> which *may* be partly caused
> by some difference in fitness or may be caused entirely by random
> events unrelated to any difference of fitness. (Or as you point out
> below, with supernatural intervention, all random elements whatsoever
> might be eliminated, and the differential survival rate might be
> *entirely* caused by differential fitness in the Mind of the
> Intervener.) In most cases, both difference in fitness (not necesarily
> in the allele under consideration, but possibly in nearby genes which
> are linked to the allelle under consideration) and randomness are
> involved in the differential survival, not just statistically, but for
> each and every life/death decision which occurs. An individual might
> die tonight because its cells lacked the ability to synthesize citric
> acid, and by chance it wasn't able to consume any food containing
> citric acid for a while, and its stockpile ran out, and Kreb's cycle
> slowed down so much that the cell exhausted its ATP supply and could no
> longer perform basic functions of life. How can you say that death was
> *entirely* 100% due to the faulty citric-acid pathway, or it was
> *entirely* 100% due to the chance lack of citric-acid-containing food?
> I say that both the differential fitness and the bad luck in the food
> department caused the death.
That's a bad example. It's selection if some other individual can
synthesize citric acid (really you mean ascorbic acid, i.e. vitamin C,
by the way) and, encountering the same circumstances, doesn't die.
Anyway, this has been covered at length already. That's not the point at
all. The point is that "stochastic selection" is a bad description of
any of this. The features that are selection are not stochastic, and the
features that are stochastic are not selection, regardless of whether or
not they are happening at the same time. Use the right word, not its
third cousin.
[snip]
>>>and the DNA cladograms can be shown live in the
>>>laboratory so only a fool would dismiss them.
>>
>>I don't know. But it's all just the nested hierarchy under another name,
>>and I expect the same responses: it exists but it's explained by
>>goddidit, not common descent, or it doesn't really exist.
>
> That's why I want to clearly organize the various kinds of evidence,
> and the various elements of the theoretical explanation for the
> evidence, and ask each individual person which items he/she accepts or
> disputes. Once they are pinned down as to where they disagree with our
> chain of evidence and logic, we can work on discussing thet particular
> breakdown-point with them. Maybe they'll learn they were overlooking
> something, and they convert to our logic. Or maybe we'll discover
> there's a flaw in our logic and our case isn't yet proven.
> .
I'm not sure you know what you want to do, or have any idea how to go
about it.
Nic
> > Even as recently as last month I was reading that haplotype blocks
> > were conserved in the same way linkage groups are; i.e. low
> > probability of being hit by a crossover. I had thus been regarding
> > the terms as synonymous.
>
> They are not synonymous.
Agreed. Whether they are compatible is more the question, or do they
assume incompatible underlying mechanisms? The alleged 'hotspots' seem
to be a new fact about the crossover mutation mechanism which
invalidates assumptions from before the genomics era, possibly
including the very assumptions the Morgan measure is based upon.
I think there are two alternative pictures here.
One is that in meiosis a chromosome may or may not undergo a single
crossover - more than one at a push. In this case, the chance of two
short subsequences finding themselves on different sides is linear in
the distance between them. Here, the outcome boils down to a question
of whether a subsequence was crossed over at all or not at all.
The other picture is that crossover routinely occurs at *many* points.
Here, the outcome boils down to a question of whether a subsequence was
crossed over an even or an odd number of times.
I suspect the prevailing view has been changing from the former to the
latter picture over the last 25 years. It certainly makes a difference
to how long ago you might think the human race went through a
population bottleneck. Or you might just buy straight into the hotspot
idea, in which case it's irrelevant.
> Does anybody need a mathematical explanation of why it's not exactly
> additive even though it's close to additive over short distances?
> Hint: P*P vs. 2*P*(1-P)) vs. (1-P)*(1-P), where P=0.99, and that middle
> term is what you want to look at.)
>From the link you gave, it looked to me as if it is additive by
definition.
> > It speculated that because the human species had been through a
> > severe population bottleneck in the distant past, there have not been
> > enough generations since that time for crossovers to have divided the
> > genome up very finely.
>
> You wrote the premise backwards. You meant to say because the human
> species had been through a severe population bottleneck in the
> *not*very*distant* past ...
Yes, I did put it the wrong way round, perhaps because I had in mind
that it must have been a long time ago in terms of human history.
Hence my inclination to identify it with the last time of speciation.
Suppose instead the population bottleneck happened 75% of the way in to
the life of the species. We would of course, have to consider the
people before the bottleneck as fully modern humans. Then how come it
was only after that time that they managed to spread to all the
continents (except Antarctica)?
I *guess* the percentages are of loci with/without polymorphism, i.e.
it's real estate, with no weighting for higher degrees of polymorphism.
So 50 is 50 whether SNPs or not. I guess this, because surely they
would have to say if they meant anything else.
> SNPs are scattered liberally through the genome. While most of them
> are found outside genes and probably do not have any effect, those
> located in and around genes may contribute to the genetic basis of our
> biological individuality, ...
> SNPs outside any phenotype-affecting regions (exons, regulatory, etc.)
> are neutral, hence drift randomly, and it takes a long time for them to
> be fixed one way or the other, so SNPs tend to remain for a long time.
> But SNPs inside phenotype-affecting regions are sometimes neutral
> (3base->1aa coding synonyms) and sometimes not neutral, and the latter
> under selection pressure, so they tend to be rapidly moved to one
> extreme or the other i.e. eliminating one allelle and fixing the other.
> So statistically, how much more frequent are neutral SNPs than selected
> SNPs compared to what you'd expect based on how many places they'd be
> neutral vs. selected? I.e. how much is SNP diversity reduced in
> selected places compared to neutral places?
I haven't the faintest idea. Because you speak of SNP diversity, I
assume you are not talking about prevalence of variants (i.e. allele
populations), but about prevalence of variation (i.e. sites having
variants).
If we are talking about non-neutral sites responsible for vital cell
chemistry or physiology, I would of course expect them to be very clean
of variation. I guess these make up the majority of the non-junk
genome, I guess they are the oldest, and I guess they are also
comparitively clean of interspecies variation. Fitness of these sites
is under constant maintenace selection, and this selection is very low
cost as it only requires the death of gametes or early stage embryos.
The above sorts of sites aside, I can see no a priori reason why the
prevalence of polymorphism in selected for sites should be lower than
that of neutral sites. Remember you are looking at a snapshot of
ephemeral polymorphisms. The prevalences are equilibrium levels. Yes,
selectable sites are constantly being depolymorphised (either by
extinction or by becoming fixed). Likewise neutral ones are constantly
being taken out by drift to fixation. I believe because of sex
(heterozygosity shelters damaged sites from selection), there is a high
outstanding population of deleterious variants on the way to becoming
extinct, or nearly so. It is therefore not clear to me that selectable
sites should be particularly clean, if we are talking about the ones
actively evolving, or about the ones responsible for traits which make
a difference between species.
That's selling it to me. Especially if, as it seems, the population
genetics is backing up the genomics.
> Recent studies suggest that the genome may be divided into a
> remarkably small number of blocks - just 200 000 or so. Recombination
> seems to be focused between the haplotype blocks, so large groups of
> alleles end up travelling together.
>
> Math: 3 billion total DNA bases, divided by 200 thousand blocks, yields
> an average of 15 thousand bases per block. 3 million base pairs
> different, 90% of them SNPs, gives 2.7 million SNPs, and divided by 200
> thousand blocks, yields an average of 13 or 14 SNPs per block.
> That means if each SNP is binary (two different allelles), there would
> be a potential for 2**13 to 2**14 different combinations of the
> allelles within a single block, hence a potential of appx. ten thousand
> different allelles at the block level. As I understand it, only a tiny
> fraction of those ten thousand possible combinations actually appear.
> If there has been no recombination within a single block in all the
> time since those 13 or 14 mutations occurred, then there should be 14
> or 15 block-allelles present (the original, plus one extra for each new
> mutation that occurred, regardless of whether the new mutation occurred
> in a tree of descent where earlier mutations had already occurred or
> not). Does anybody know the average number of alelles of such a
> haplotype block that has 13 or 14 SNPs within it? Is it like 20
> allelles, which means that recombination within that block has happened
> only about five times in all of human history, or more like 100
> allelles, which means that recombinaton within that block has happened
> about 90 times, or even more?
The complication here is that unless you sample every little village in
the world, you won't find the most recent few mutations that went in.
Also you won't see if the most recent few is actually thousands of
different very local ones because of the population explosion.
> Note that if at the time of the population bottleneck (or shortly
> after) only a single allelle of a nowaday haplotype block survived, we
> should be able to identify it. Survey all allelles of this block in
> populations around the world. If a block-allelle occurs in *all* of
> them, it's probably the ancestral block-allelle, whereas if it appears
> only in a few populations in just a few geographical regions and/or
> along a route of miagration, then the block-allelle originated due to a
> new mutation after the bottleneck and didn't have time to be
> distributed to all regions of Africa before the first miagration out of
> Africa, or it originated *after* the first such miagration. If more
> than one globally-distributed block-allelle is found, it indicates that
> all such block-alleles were ancestral and survived through the
> bottleneck. It's remotely possible that more than one block-allelle
> survived the bottleneck and for a while after, but then eventually
> drifted to fixation in all the world's populations, but that's highly
> unlikely. So if we identify only a single allelle of a particular
> block, we can assume it probably was the only survivor of the
> bottleneck.
As I tried to say above, a single allele doesn't make a case of
polymorphism. Also a large part of the non-junk genome is highly
conserved, but capable of carrying SNPs. Certainly universal
distribution of an allele would strongly suggest it is ancient.
Chances are however the blocks have all undergone changes to neutral
sites, and each such change is not globally distributed geographically.
I assume it is possible to construct a genealogical tree explaining
the currently extant alleles of a particular block based on shortest
mutational distance from a single common ancestor (which probably will
not still be around, and so whose nature has to be hypothesised). This
technique can however give multiple solutions. The bottleneck
hypothesis is supported if there aren't multiple solutions, and the
root isn't too far back in the past given some assumption about
mutation rate.
> If we find a lot of such single-bottleneck-allelle blocks
> and hardly any multiple-bottleneck-allelle blocks, it would indicate a
> very small population through the bottleneck, a really severe
> bottleneck for sure! On the other hand if we discover that lots of
> multi-allelle blocks survived the bottleneck, then maybe the bottleneck
> was only mild, that a decent-sized population existed at all times
> during the bottleneck.
That would be very strange, because the original reason for supposing
such a bottleneck was that human mitochondria have a recent common
ancestor. Your decent sized population would therefore have to be one
woman and many men - either that or they kept very complicated records
to enable them to breed all but one line of mitochodria to extinction.
I think we can take it that there was a bottleneck. The question is
whether it helps or hinders deducing whether haplotype blocks are a
reality.
> Therefore, rather than the
> millions of allele combinations potentially available, the human
> population seems to be made up of a more limited set of haplotype
> patterns.
>
> That math is not correct. With 2.7 million SNPs, if they were
> rearranged in all possible combinations, assuming each has only two
> allelles, the total number of combination allelles would be
> **HOLD YOUR BREATH** 2 ** (2.7 million) =appx= 10 ** (800 thousand)
> not merely "millions" as implied above, not a billion, not a
> quadrilion, not even as small as a GOOGOL, more like a GOOGOL to the
> eight power!!!!!!!!
It makes you realise that even if there were never any mutations, there
is a near infinite fount of variability in sexually reproducing
organisms. Despite that, there are bounds. You couldn't for example
get from fur to feathers.
I'm starting to think they were weaseling because they didn't know.
> The map will be based on DNA samples obtained from hundreds of people
> in geographically distinct populations: Nigerian Yorubas, Han Chinese,
> Japanese, and US residents of European origin.
>
> (That seems to have been the original plan. Apparently they later
> decided to survey only a small local population of Utah residents for
> the US sample. And for Japan, they picked a group in Tokyo.)
>
> A good way to think about haplotypes and haplotype blocks is to
> imagine the SNP alleles as children sharing school minibuses.
>
> Oh boy, I'm not the master of metaphor after all!
>
> By the way, I thought I invented the term "haplotype block" myself,
> after reading the article in _Scince_, becuase when I started using the
> term here people said they had no idea what I meant. But I see the term
> was already in use by the HapMap founders before the project even
> started, so I claim innocence of the coining.
>
> Additional note: Very few DNA-base-pair-neighbors (appx. 200 thousand,
> out of 3.1 billion total) have crossed over even once since the
> population bottleneck, most (99.99%) haven't crossed even once during
> all that time, yet some very "hot" places have crossed many many times
> during the same time. Is that right, or did I miss something?
Dunno. And I'm getting innocenter by the minute.
Tending to disappear? It seems that way, but the population does a
random walk, which has no 'tendency'.
> I predict that when we calculate
> statistics of age of neutral SNPs and adjacent-haplotype-block-pair
> cross-overs, we'll get the same age-distribution within any single
> local population group, which will clearly show when various spurts of
> population growth occurred, either globally or within local
> populations.
>
> Furthermore, if we collect haplotype data for each living human, we may
> be able to use both SNP-origin ages and block-boundary cross-over ages
> to compute a complete pedigree for each such human, that is a complete
> family tree for all of Homo sapiens, going back all the way to the
> population bottleneck!! Should I explain how I believe this would be
> possible, what specific calculational methodology would accomplish that
> task/goal, or can you-all figure it out for yourselves after I have now
> spurred you with the idea?
This might show everything but that which is most of interest - the big
evolutionary innovations. The extant gene pool contains historical
information about changes that did *not* encounter strong selection.
Strong selection, and the past is obliterated.
> (Final note: I was thinking maybe this one article was getting too
> long, so I might split it into parts. But it's only 21k bytes, so what
> the heck, I'll keep it intact and see if the auto-moderator accepts
> it.)
> .
It wasn't truncated.
Nic
John Harshman
> an...@sci.sci wrote:
>
>>>Even as recently as last month I was reading that haplotype blocks
>>>were conserved in the same way linkage groups are; i.e. low
>>>probability of being hit by a crossover. I had thus been regarding
>>>the terms as synonymous.
>>
>>They are not synonymous.
> Agreed. Whether they are compatible is more the question, or do they
> assume incompatible underlying mechanisms? The alleged 'hotspots' seem
> to be a new fact about the crossover mutation mechanism which
> invalidates assumptions from before the genomics era, possibly
> including the very assumptions the Morgan measure is based upon.
Depends on their size. If these blocks are only a few thousand bases, or
even a few hundred thousand, then nothing really changes. If they're
millions of bases, then we get problems.
The average human gene is about 20kbp (the bulk of that being introns)
and there's about twice that space on average between genes. So even
with these blocks, unless they're very big, we should get recombination
between adjacent genes.
I pulled the numbers from here: http://cnx.rice.edu/content/m11317/latest/
I would have to say that just browsing the genome my impression would
have been that the average size of a gene was even larger than that.
>> Genetically speaking, humans are incredibly similar to one another.
>> Any two unrelated genome sequences differ at only one position in a
>> thousand, on average. The 0.1 per cent difference, which amounts to
>> about three million base pairs of DNA in total, ...
>> Much genetic diversity (around 90 per cent) consists of single
>> nucleotide polymorphisms (SNPs), ...
>>(What do the other kinds of diversity consist of? Tandem repeats?
>>Rearrangements? Duplications? Block deletions? Point inserts and
>>deletes?
All of the above.
> How do you compare large block changes with point changes on a
>>fair basis? For example, does a single block of 50 bases that gets
>>duplicated count the same as 50 separate SNPs, or just 1 or 2 SNPs?)
Depends on what you are trying to count. If you want (for some reason)
to count simple sequence differences, it might count as 50. If you want
to count evolutionary events (usually a more useful thing to count),
then it's one. If you want to count the probability of it having
happened identically more than once (i.e., homoplasy), that gets to be a
complicated question.
All those other kinds of diversity count for much fewer evolutionary
events, or single mutations, than do the SNPs.
> I *guess* the percentages are of loci with/without polymorphism, i.e.
> it's real estate, with no weighting for higher degrees of polymorphism.
> So 50 is 50 whether SNPs or not. I guess this, because surely they
> would have to say if they meant anything else.
I can't interpret this. I would guess that there are very few loci
without polymorphisms. If the probability of a mutation is in the
neighborhood of 10^-9 per site, per individual, per generation (and
that's a reasonable ballpark figure), then most sites will experience
multiple mutations per generation, somewhere within the population. That
makes total polymorphism pretty much a meaningless number. And that's
why mean polymorphism is the figure we count, and why frequency is
important.
[snip]
>>Now by measuring the allelles of SNPs in lots of different people, we
>>should be able to observe how widely distributed they are in various
>>populations in different geographic areas, and from that information we
>>should be able to estimate when each SNP first occurred due to point
>>mutation. Likewise by mapping the haplotype blocks, in particular by
>>measuring how widely distributed crossing-overs at boundaries between
>>adjacent-block are compared to inheritance of the same ancestral pair
>>of adjacent blocks, we may be able to estimate how long ago each
>>crossing-over between two adjacent haplotype blocks occurred. Just like
>>SNPs, new combinations of haplotype blocks due to crossing between
>>blocks would tend to disappear except when such cross-overs occurred
>>during times of population growth.
>
>
> Tending to disappear? It seems that way, but the population does a
> random walk, which has no 'tendency'.
The great majority of all neutral mutations disappear. Is that a
tendency? the frequency of a new mutation is 1/2N, and the probability
of a neutral mutation eventually becoming fixed is its frequency.
Therefore a new mutation has a probability of 1/2N of being eventually
fixed, and a probability of (1 - 1/2N) of becoming extinct. Is that a
tendency?
an...@sci.sci
> > eggs, it's hard to say whether they should have counted as "already"
> > done their job or not.
> The ocean dwelling critters have indeed done their job. The humans
> haven't even got off the starting blocks.
Ah, auto-racing metaphor. Continuing: The humans have no more than
started their engines at that point, still need to drive off starting
block, line up in 3 columns, take one warm-up lap, before the race
starts. (No, I won't try to match the different start-up phases of a
race as listed above with the various stages in nurture of children to
adulthood, like saying the warm-up lap is young adulthood while still
living with parents and attending college and posting singles ads.)
(BTW I'm glad you switched from the usual baseball metaphor.)
> A component of the subsequent generation's success is creditable to
> the parental genotype. That is how evolution 'sees' it, anyway. That
> component is *in addition to* the component due to the parental alleles
> being represented in the offspring. It is due to the actual
> combination expressed in the parents.
Indeed, child's genotype-based phenotype, and parents' phenotype and
memotype as expressed in nurturing, are conflated in effect on
children's survival rate. And if you include grandparents and
community, we have the "it takes a village to raise a child"
conflation. Only by mathematical analysis of variance (teasing apart
the various factors) can we then get the data for correctly calculating
the fitness of a genotype-based phenotype. (For example, how much of
the responsibility for the crime is the individual, how much the
parents, and how much the inner-city ghetto where they all lived.)
Anyway, I like your clear (easy to understand) nicely worded statement
of that principle/understanding/enlightenment.
> > Now if you conflate the 99% of life cycle, and the 1% of reproduction
> > itself, into a single calculation of reproduction of the daughter cells
> > at exactly the same life-cycle point as we started the calculation but
> > exactly one generation down the line of descent, *that* conflated
> > calculation is indeed stochastic. But the stochastic part of it is the
> > life, not the reproduction. But if that's what you meant to be talking
> > about, you should have made it clear up front, instead of using the
> > word "reproduction" all by itself and claiming *that* was stochastic.
> You seem to be hanging back from the idea of meiosis introducing any
> randonimity into the fate of alleles.
I seem to recall I was thinking more of bacterial fission, which
doesn't introduce hardly any random element whatsoever in any healthy
strain of bacteria. Only if there was radiation or other mutagen damage
immediately before or during reproduction would a problem happen during
the act of fission to create daughter cells with different genome from
parent.
In the case of mitosis, same analysis.
In the case of meiosis, more complicated:
> Is that because you know it produces gametes in pairs, so that no
> parental allele actually gets left out? If so, I see your point. The
> randonimity/selection comes in after this, when the haploid phase of
> the life cycle takes its 'test'.
I wasn't thinking of that at the time, but indeed in the case of
symmetric production of male gametes (sperm), this is true. In the case
of assymetric production of female gametes (single egg, and
non-continuing polar bodies), this argument doesn't apply. There is
indeed a random element as to which half the genome gets into the one
egg. So in that one case, we have a stochastic element, but no
selection element, pure random-drift.
In species that have more than two sexes, or a caste system with both
diploid and haploid individuals or clones of a single genome
cooperating in a hive, etc., it's even more complicated but I don't
know the details so I won't discuss such cases here.
> Gamete survival, although it 'reads in' some environmental
> information, it reads in practically none that it is responsible for
> species diversity. I think this is purposeful - it is inefficient for
> one genome to have to master two niches
Superficially that seems true, but in animal bodies with
differentiation of cell types and many tissues organized into many
organs into several systems, a single genome satisfies a wide variety
of different local environments within the body, via turning various
genes on or off in different tissues and even under different
circumstances within the same tissue, and already genes are turned
on/off at different stages of development, for purpose of growing new
tissue for a while then *stopping* that new growth when adulthood has
been achieved. So it seems only a tiny additional use of the same
mechanism for animals to also turn genes on/off during different stages
of life cycle in order to adapt to different environments. Given that
they are already well adapted to adult circumstances per your scenerio,
it seems there would be selection pressure to then adapt the larval
stage to better adapt to its own environment in a way that doesn't
interfere with already-fine adult adaption, and all it takes is one
species that has chanced upon that adaption to overrun all the other
species via a major radiation. All that's really needed is a complete
copying of the entire genome, and then changes in the regulatory
sequences to use one copy during larval stage and another copy during
adult stage, and voila, the two copies can now adapt independently from
now on.
And of course I'm *sure* you don't mean "purposeful" in the ontological
(Intelligent Design) sense, right?
> (hence the pressure for neoteny,
Unfortunately that word is ambiguous and I don't know which you mean:
<http://dictionary.reference.com/search?q=neoteny&db=*>
1. Retention of juvenile characteristics in the adults of a species,
as among certain amphibians.
2. The attainment of sexual maturity by an organism still in its
larval stage.
n : an evolutionary trend to be born earlier so that development is
cut off at an earlier stage and juvenile characteristics are retained
in adults of the species
> and for marine mollusks to keep ditching their planktonic phase).
By "ditching" you mean evolve to a life-cycle where there is no longer
any planktonic phase, the gametes fuse right on a substrate (rock or
coral) where they immediately start developig the adult sessile form,
as opposed to their ancestors who developed planktonic forms then
metamorphised into sessile forms after landing on a suitable substrate?
Or for those which aren't sessile as adults, such as octopus,
developing the adult crawling/tentacled form immediately?
Perhaps the eggs were already laid directly on the substrate, so any
time sperm find them they're already at their new home?
I guess there are two possible mega-adaptions, double much of the
genome to allow separate adaption at two life phases (as insects *must*
have done to have caterpillars/grubs and adults each well-adapted), or
find a way to no longer need any juvenile form (as you say some
mollusks have done). So once any species discovers one of those two
adaptions, its descendents are forever locked into that mega-decision,
so one clade would do it one way and another clade the other way, and
lots of clades haven't yet chanced upon either way yet but still manage
to survive in different niches.
Notice how animals and plants have chanced upon two very different
adaptions to life: Plants have strict alternating generations whereby
both haploid and diploid cells can undergo cellular reproduction,
perhaps two different versions of mitosis, whereas animals never
chanced upon how to do haploid mitosis, so sperm and eggs can't ever
grow and produce daughter cells, so they must immediately find mates to
achieve fertilization, or they die. But animals chanced upon a way to
regulate genes so flexibly as to be able to create organs and systems,
which plants never chanced upon, so plants are very limited in the body
plans they're capable of, in fact they seem to have only about two,
namely one for haploid generation predominating (absorb water directly
into tissues) and one for diploid generation predominating (absorb
water in roots and forcibly pump it to other tissues), the various
phyla of vascular plants all using the same body plan, only the way
they encase eggs in maternal tissue and either throw them to the wind
or attract insects to distribute sperm (pollen) to eggs and/or
distribute the resultant seeds, differ in the various vascular-plant
phyla. Whereas animals have more than thirty different body plans (one
per phylum), about half of which are worm-like but with so very many
different ways to "design a worm" compared to what any human might
happen to think of.
Hmm, a few days ago I submitted an article to sci.bio.evolution
proposing that Fungi be demoted from status as "kingdom", because they
don't have any major adaption across the entire "kingdom" as plants and
animals do across theirs respectively. Their definition consists only
of a loss of a major adaption, i.e. lack of undulipodium. But my
article hasn't yet appeared. I'll have to ask about that right after
sending off this article.
> My second point is that meiosis *does* introduce randonimity into the
> fate of alleles because every time it produces an unviable combination
> it blocks the reproduction of alleles which aren't part of that
> combination, as well as ones that are.
No. Suppose the following combinations are viable: AB Ab aB, while ab
is not viable. (Warning: I'm not using capitalization to indicate
dominance here. Rather just two different allells of same site. Think
of a case where *both* allelles of diploid are *always* expressed, but
one combination of allelles ab at related sites creates a biochemical
pathway that produces a toxin, whereas the other combinations produce
biochemical pathways that are non-toxic and even slightly useful. So
even if the other parent gives nice AB, still the diploid child has
AaBb which expresses enzymes to produce the toxin.) Suppose the (one)
parent's strands are Ab and aB, but recombination produces AB and ab.
(Hmmm, I must be contradicting myself: If this one parent had two
strands aB and Ab, then that parent expressed the toxin already. Think
instead a case of A/a being regulatory and B/b being exon, so ab must
be on same strand for the toxin to be produced, OK nice example now?)
The ab dies out, but the AB has no problem surviving.
> Maybe I'm falling into the trap of double counting here, and that
> effect is already included in haploid phase selection.
Yes. I think so. The actual meiosis proceeds to completion just fine,
it's only later when the egg dies, or the zygote dies, due to
expression of the ab juxtaposition on a single strand.
Anyway, if we do all accounting around a complete life cycle, back to
the same phase as we started, conflating all cell-division and
cell-survival factors in a single life-cycle-effective-fecundity-number,
we can hide all these disputes under the rug. We just need to make it
clear when we're bundling a complete life cycle like this, and when
we're trying to tease apart the various mechanisms during the different
phases of the life cycle, and not write as if treating separately when
we're actually bundling/conflating. Given the effect of paternal
genetype/phenotype on the survival of nurtured children, which you
pointed out earlier, it may be difficult to find a particular point in
the life cycle where all parental influence is gone and we can "start"
a life cycle with success based on genotype of only self not parents.
> Surely though, gametes (especially female ones) are not produced in
> sufficient quantity to average this effect out.
Agreed. As I said earlier in this reply, my original remarks were
mostly directed at prokaryotic fission, not eukaryotic meiosis, and in
the case of female meiosis there *is* a stochastic element in meiosis,
and as you pointed out just here, with linkage changing (and my example
of ab juxtaposition on same chromosome being fatal) even the male lines
occasionally may experience a stochastic element in meiosis.
Thanks for helping me clarify the limits of my argument.
> Suppose the parents are contributing rival alternative coadapted gene
> complexes, one of which is wholly dominant. In this case meiosis will
> *almost* always break both complexes, and fatally.
By use of "gene complexes" here you're assuming that these complexes
function normally *only* when present along a single chromosome, not
when split between the two aligned pairs in a diploid cell, as with my
ab example earlier except the opposite case where the juxtaposition is
essential for life rather than fatal to life?
It is my belief that this factor is actually a major incentive for
haplotype blocks to exist in the first place, and also a major
incentive for "gene complexes" too large to fit within a single
haplotype block to at least fit within a contiguous line of blocks so
that even though meiosis breaks them up it's moderatley rare for that
to happen, so that fatal cross-mixups between competing complexes
doesn't reduce survival rate enough to produce reproductive isolation.
> Maybe this is what happens in mules.
That's a reasonable hypothesis. More specifically, in the distant past,
in the common ancestor of horses and donkeys, there was no hot spot in
the midst of this gene complex, any mutation that created a hot spot
would cause the kind of fatal crossing-mixmash you propose, so it'd be
rapidly eliminated. But once horses and donkeys were separate species,
there was no further need to keep these complexes intact, because each
species had only one version of the complex anyway, so it'd be crossing
with itself. With such relaxation on stabilization pressure, eventually
a hot spot chanced into existance, and persisted. Now when the two
species try to mate, the hot spot mixes the two complexes into a
non-useful combination, causing sterility. Actually a single hot spot
would mix them fatally only part of the time. More likely there are a
whole bunch of hot spots, so that it's very unlikely that the resultant
combination would be exactly correct. But even then, there would
be occasional fertile mules. So if *none* are ever observed, the reason
must be a bit more complicated. Maybe also the two different gene
complexes are so grossly incompatible that even if they are retained
each intact on a pair of matching chromosomes, still each expresses
something that together are fatal.
.
an...@sci.sci
first new line, and I had to scroll down all that distance and eyeball
all of it just to make sure I didn't miss the start of your reply.
Then the first 2 lines of your reply contained this personal insult:
> I still believe we are talking about two very different universes,
> only one of which resembles real biology.
No, I'm talking about the same Universe as you are, but I'm using words
in a more normal way, see below:
> Take a steady-state ecosystem including, say, one million asexually
> reproducing organisms, whether bacteria or amoeba or whatever. Wait a
> year after which time these organisms will have gone through many cell
> divisions. Because the ecosystem is in a steady-state, you will end
> up with one million organisms. Trace the ancestry of these one
> million back to the original population and you will get a tiny
> fraction of the original successfully reproducing.
Yes, that's correct. You're talking about longterm survival of the
various clades of a population, not about individual acts of
reproduction. With bacterial reproductive time of 20 minutes, that
would be equivalent to more than 26 thousand generations, which in
human terms would be half a million years, longer than modern Homo
sapiens has even existed.
If you're going back that far in time, and rewriting history to say a
particular individual "never reproduced" just because none of its
descendents exist half a million years later, you're abusing the
English language, IMO. By that standard of language misuse, you could
say that none of the Neandertals ever reproduced.
> The fact is that most systems are usually in something close to a
> steady state. As Darwin (and Malthus) noticed a few years ago, the
> propensity to procreate far exceeds the reality of procreation.
Um, that should read "the reality of subsequent survival after
procreation has already occurred". Per normal English usage,
"procreation" means:
<http://dictionary.reference.com/search?q=procreation&db=*>
1. To beget and conceive (offspring).
n : the sexual activity of conceiving and bearing offspring
which says nothing about subsequent survival of the next generation,
much less longterm survival of distant descendents therefrom.
> The only thing that matters is having ones particular alleles passed
> down to future generations.
The Universe as a whole will cease to support life in a few trillion
years, so per the extreme view that *only* survival forever matters, we
are all already dead, nothing whatsoever ever matters. You have to cut
the time span to a finite threshold to give the definition any meaning,
such that some things matter and others don't matter. One reasonable
threshold I would propose at this time is to live long enough to
deposit a bunch of fossils and/or artifacts that persist long enough to
be discovered by somebody later, in which case the non-avian dinosaurs
mattered for a long time, even if not successful ultimately as we now
know. Another possible threshold would be that they survive as long as
there's any life whatsoever, in which case the non-avian dinosaurs
don't matter at all, and we don't yet know whether *anything* in
particular matters at all, not even birds or humans, because we don't
yet know whether our lines of descent will survive right to the Cosmic
end. I personally think that a threshold of 20 million years would be
just about right, since many species go extinct in just a few million
years, or even less, so achieving 20 million years of living
descendents is quite an accomplishment.
But really, defining what matters and what doesn't matter, as any sort
of absolute criterion, is silly. There are varying degrees of
mattering. We should just measure how long any given clade survives,
and compare one clade as mattering more than another, but not say that
one matters and the other doesn't.
> The only thing that makes sense biologically in reproduction is the
> entire life cycle, from adult organisms in one generation all the way
> to adult organisms in the next generation.
<flip> But none of that matters, because it's only a single generation,
not 26 thousand generations. </flip>
I agree that conflating all the processes all the way around one life
cycle is the correct way to calculate effective fecundity-number,
providing the life cycle works that way. In some kinds of life, this
isn't possible, for example:
- Cnidaria have both a sexual cycle and an asexual cycle, but both pass
through a stage called "colony", so perhaps if you conflate both cycles
and talk about how many new descent "colonies" are produced from a
starting colony it works, just barely.
- Apicocomlexa have an even more complicated intermix of life cycles,
although again they all pass through a common phase, namely trophozoite
within host cell, so again we might be able to declare that as the
phase for "start" of life cycle where we measure fecundity, just barely
again.
- Oomycota, specifially Saprolegnia, again have two interlocked life
cycles, but the single point of intersection is hypha, and I'm not sure
you can count how many there are in order to compute fecundity from
that point to the same point in the next generation.
- Angeiospermophyta, specifically vines such as blackberries or grapes,
where there is both sexual reproduction, and vegetative growth and
happenstance splitting of a single "plant" into separate plants. How
can you count how many "plants" there are, in order to compute
fecundity number???? How can you even decide where to draw the line
between "generations"??
The majority of life admit the possiblity of a clearcut life cycle
wherein fecundity number can be precisely defined and perhaps measured,
but clearly the idea can't be universally applied. Until recently
humans were among that majority where there's only a single life cycle.
But within the past week I became aware of a company that had developed
a way to culture skin cells to grow them into flat sheets, then wrap
them around a rod to form a cylinder, and further culture them to heal
the winding-splices to thereby form nicely uniform tubes, and then
graft them back into the original cell-donor to form blood vessels,
which then grow just like natural blood vessels as the human being
grows to adulthood. This general technique might someday be developed
to grow replacement tissues of all types, eventually replacement organs
of all types, which can then be spliced together to yield entirely
Frankenstein-style beings, perhaps more quickly than traditional
cloning to yield a single-cell zygote could yield an adult human. At
this point, the well-defined natural life cycle will no longer be fact,
and it'll be difficult to count generations and measure fecundity.
Hey, we've deviated considerably from my original purpose of this
thread, to create a concensus checklist of aspects of evolutionary
theory, to present to anti-evolutionists and open-minded freethinkers
to ask them which they accept and which they dispute, as a way to
target our debates toward specific points of dispute. Maybe we should
change the Subject of these sub-threads that have gone off on special
topics not necessary for the simple checklist? If you change the
Subject field, please cite my an...@sci.sci fake address in what you
post, so I'll be able to find it by Google search, OK? No, wait, Google
Groups now builds threads by references, not by Subject fields, so
it'll link everything correctly and I'll find your reply anyway, never
mind. Still at least one person must mention my fake address or else my
Google Groups search won't find this thread at all and I won't browse
it to find your followups. So please do refer to my fake address from
time to time regardless of whether you change the Subject field or not.
Never mind the "never mind" above. As you were.
.
Nic
John Harshman wrote:
> Nic wrote:
>
> > an...@sci.sci wrote:
> >
> >>>Even as recently as last month I was reading that haplotype blocks
> >>>were conserved in the same way linkage groups are; i.e. low
> >>>probability of being hit by a crossover. I had thus been regarding
> >>>the terms as synonymous.
> >>
> >>They are not synonymous.
>
> > Agreed. Whether they are compatible is more the question, or do they
> > assume incompatible underlying mechanisms? The alleged 'hotspots' seem
> > to be a new fact about the crossover mutation mechanism which
> > invalidates assumptions from before the genomics era, possibly
> > including the very assumptions the Morgan measure is based upon.
>
> Depends on their size. If these blocks are only a few thousand bases, or
> even a few hundred thousand, then nothing really changes. If they're
> millions of bases, then we get problems.
To say that, you have to have a view on how often crossover events
happen and how long it is since there was this population bottleneck.
I haven't looked up that information, let alone how big the blocks are
supposed to be.
> The average human gene is about 20kbp (the bulk of that being introns)
> and there's about twice that space on average between genes. So even
> with these blocks, unless they're very big, we should get recombination
> between adjacent genes.
I didn't realise the genes were spread out like that. In that case,
the haplotype blocks could indeed be an artifact. I had always
imagined the genes would be closer together because in the past there
must have been fewer chromosomes. But I guess that's silly because
chromosome number increases by accidental duplication, not by the
creation of new, empty ones. Still, it's hard to see how genes could
spread out to fill the space available. Maybe it's because insertion
mutations are guarranteed survivable if they land in the gaps between
genes.
> I pulled the numbers from here: http://cnx.rice.edu/content/m11317/latest/
>
> I would have to say that just browsing the genome my impression would
> have been that the average size of a gene was even larger than that.
Implying they are for proteins with structural roles rather than
catalytic?
I meant they would look at the total number of base pairs different,
whether comparing two human beings from the same locality, from
different continents, or if comparing a human with a chimp, or with a
geranium. Any more complicated comparison would demand an explanation.
I know. A random walker starting one step in from the edge of a cliff
hasn't got the best of prospects, despite there being no bias to its
meanderings. Still, it's got 50:50 of staying one more round, which is
exactly the same as for the unmutated variant in my sister.
Nic
Nic
I think possibly the linkages don't steretch very far, but you can
bridge two linkages with a third which overlaps parts of the original
two. So they could deduce the adjacenies and concatenate them to a
large extent in the days before mass transcription of genomes.
> Simply put, two genes (gene loci) are linked if they are on the same
> chromosome. If two genes are linked, they do not assort
> independently; they show linkage. However, if genes are far enough
> apart on the chromosome, the probability of crossover between them is
> so high that essentially that there is a 50:50 chance that they assort
> together in meiosis. That is, they behave as if they are not linked.
> So the definition as the "maximally connected subset". A and B show
> linkage, as does B and C as does C and D .... as does Y and Z. But A
> and Z are so far apart that you can't measure linkage. Still, they
> belong to the linkage group of A, B, C, D, ... Z.
Oh, perhaps you're getting at the same point I'm getting at above.
Nic
John Harshman
Not that I can see. Where did you get that idea? Remember that most of a
gene is non-coding sequence, largely introns. If I recall, the mean gene
has 5 exons, averaging 200 bases each. So that's 1000 bases of coding
sequence, and 19,000 bases of non-coding sequence.
Still don't understand. Who would do this, and why?
[snip]
an...@sci.sci
> I was talking about the very existence of haplotype blocks as stable
> entities.
Nothing is absolutely stable. Even protons will eventually decay.
The question is how stable are they, stable enough or not?
It's my understanding that among the 3.1 billion
adjacent-base-junctions in the DNA strands in our human genome, only
0.01% of those have crossed-over even once in all our history since the
major bottleneck that probably caused a founder effect to yield our
present species. The blocks between those very sparse crossing-points
have been stable all that time. No, they're not absolutely stable. A
few of them will surely experience a cross-over sometime in the next
century or so, while the vast majority will continue as unbroken blocks
during that next century or so. Once our population extends into space,
with vastly more habitat available, no limits to growth, and our
population grows to trillions, quadrillions, then with such a larger
number of individuals living in parallel, the rate of such new
breakages will increase even if the rate per individual remains very
very rare. Eventually the blocks may all be broken in one clade or
another, rendering the haplotype method no longer useful. But by that
time, we will have the genome of *every* individual at some epoch
registered in a database, and it'll be possible to use that database to
work backwards to the Dawn of Mankind however we want to study it, and
of course we'll have the pedigree of every new individual after that
time registered as well, so breakdown of this methodology won't be a
problem. We'll use the birth registry to track any family tree back to
the earliest all-genome-registry, and then use haplotype blocks to
track the family tree back from there to the Dawn of Mankind, with a
period of overlap where birth-registry and haplotype-tracking are
redundant.
> > ... The entire chromosome is a single linkage group, by such a
> > daisy chain of A-link-B-link-C-..., assuming markers are known
> > sufficiently close each to the next.
> That's not the way it's generally used in my experience.
What experience is that? In your particular lab, you have only a small
set of markers, with sufficient gaps between them that the markers near
one end of a chromosome aren't linked with markers near the other end,
giving you two separate linkage groups on one chromosome, like this:
a b c d e x y z
5cM 12cM 8cM 10cM (more than 50 cM gap hence "not linked") 9cM 8cM
per one of the definitions I read on the Web which defines linkage
groups only in terms of a given set of markers, not absolutely for a
chromosome, so a,b,c,d,e and x,y,z would be two linkage groups?
But including other markers would achieve
a b c d e k p t x y z
5cM 12cM 8cM 10cM 20cM 30cM 30cM 20cM 9cM 8cM
so now with those extra markers to fill the gap, you can measure
linkage across each medium-gap in the middle, so now there's a linkage
chain all the way from a to z, hence only one large linkage group
a,b,c,d,e,k,p,t,x,y,z for the whole chromosome?
Note I have no bio-lab experience. I go by what I read on the Web:
<http://opbs.okstate.edu/~melcher/MG/MGW1/MG111122.html>
* Markers that have measurable recombination frequencies are said to
be linked.
(i.e. two different markers, such that recombination frequency between
that pair is high enough to measure, are said to be linked.)
* Markers related through a chain of linkage constitute a linkage
group.
(i.e. linkage group is maximal set connected via pairwise linkage, and
is a function of which particular markers are being studied, not an
absolute aspect of a chromosome)
* As more and more markers are studied, linkage groups become larger
(and the number of groups smaller) until the number of linkage
groups equals the number of chromosomes.
(thereby defining "linkage group" in an absolute sense, as function of
chromosome per se, as I used the term)
Are we in agreement, except your particular lab is using only a few
markers so the first part of that applies relative to your set of
markers, but if using all known markers what I said would apply, per
the last quote above?
> > <http://www.iscid.org/encyclopedia/Open_Reading_Frame>
> > A reading frame, in biology, consists of three-nucleotide codon sets
> > in either DNA or RNA that are contiguous and non-overlapping. An open
> > reading frame (ORF) is a similar sequence that can be translated into
> > a protein or a polypeptide.
> > In any open reading frame, the start-code sequence or initiation codon
> > that begins the protein is methionine ATG, and then stop-code sequence
> > or termination codon ends it. The stop sequence is coded by what is
> > termed a nonsense codon, or a codon that does not have an RNA match.
> > There are only three nonsense codons: amber(UAG) ochre(UAA) and opal
> > (UGA). As you can see, each one contains a "U" nucleotide, not normal
> > to DNA.
> > (That last one is bonkers!! Who wrote it, Archimedes Plutonium??)
> It's convoluted and contains much unnecessary information, but it's
> really the same as all the others, which match each other.
Well when I said it was bonkers, I was referring mostly to the claim
that the U nucleotide appears in the DNA genome. Is that correct or
bonkers?? If the U is in DNA, how does it get put there during
anti-replication from the opposite strand of DNA?
> Gene-estimation programs are used to identify genes, introns,
> exons, and such, and they rely on a large set of clues.
There's a difference between the following claims:
- Our best estimate, based on all those clues, is 30k genes in human
genome, of which only half have been proven to be genes, by means of
showing that they either code for polypeptides or code for regulatory
RNA or have some other known function.
- There *are* 30k genes in human genome, of which only half have known
function.
> > <http://www.thedoctorwillseeyounow.com/articles/other/genome_4/index.shtml>
> > It is believed that
> > the human genome has between 50,000 and 100,000 genes.
> > (Somebody needs to update their Web site!)
> > It is estimated that the sequence of the human genome should be
> > completely mapped by approximately the year 2005.
> > (Yeah!)
> Ah, internet paleontology, or vestiges of the natural history of web
> creation.
Yeah. In that one case, I'm complaining about an article from years ago
which is still online as if were current information, accessible from
Google search (how I found it), with no warning that any of it is out
of date, and no way to post a followup warning that it's obsolete. In
such a case I don't expect the author to delete it, but it would be
nice if somebody would just indicate which parts are hopelessly
obsolete, compared to the main bulk of the old article which describes
some particular methodology which is still the correct description and
hence still eduational.
But in the other cases I cited, I presume you can see the problems with
the conflicting or just plain wrong text, where the Web maintainer
really ought to correct the information.
.
r norman
>Complaint: You quoted eight screens of context before you wrote your
>first new line, and I had to scroll down all that distance and eyeball
>all of it just to make sure I didn't miss the start of your reply.
Get a real news reader that automatically scrolls to the new stuff!
(Sorry, couldn't help that. Before you insult me by complaining
about my insulting you, that is what you told me in a previous post.)
I included eight screens of total nonsense because that is what you
wrote that I argued belonged to a totally alien universe.
>> Take a steady-state ecosystem including, say, one million asexually
>> reproducing organisms, whether bacteria or amoeba or whatever. Wait a
>> year after which time these organisms will have gone through many cell
>> divisions. Because the ecosystem is in a steady-state, you will end
>> up with one million organisms. Trace the ancestry of these one
>> million back to the original population and you will get a tiny
>> fraction of the original successfully reproducing.
>
>Yes, that's correct. You're talking about longterm survival of the
>various clades of a population, not about individual acts of
>reproduction. With bacterial reproductive time of 20 minutes, that
>would be equivalent to more than 26 thousand generations, which in
>human terms would be half a million years, longer than modern Homo
>sapiens has even existed.
So don't wait one year. Wait one hour. That is three rounds. Wait
20 minutes. In mitosis or binary fission, one cell produces two.
Period. End of story. In the real world, only half of the organisms
reproduce and the other die. After a number of such rounds, it turns
out that only a relatively small number of original citizens produced
any surviving offspring at all. If you don't think that is important
in evolution, then you do inhabit a different universe.
>If you're going back that far in time, and rewriting history to say a
>particular individual "never reproduced" just because none of its
>descendents exist half a million years later, you're abusing the
>English language, IMO. By that standard of language misuse, you could
>say that none of the Neandertals ever reproduced.
As I said, look at one reproductive cycle. Look at ten. Look at a
few dozen. That is realistic.
<snip>
>> The only thing that makes sense biologically in reproduction is the
>> entire life cycle, from adult organisms in one generation all the way
>> to adult organisms in the next generation.
>
><flip> But none of that matters, because it's only a single generation,
>not 26 thousand generations. </flip>
For a population in steady state, in even a single generation not
every individual produces offspring. I don't care if it is one
generation or 26,000.
>I agree that conflating all the processes all the way around one life
>cycle is the correct way to calculate effective fecundity-number,
>providing the life cycle works that way. In some kinds of life, this
>isn't possible, for example:
>- Cnidaria have both a sexual cycle and an asexual cycle, but both pass
>through a stage called "colony", so perhaps if you conflate both cycles
>and talk about how many new descent "colonies" are produced from a
>starting colony it works, just barely.
>- Apicocomlexa have an even more complicated intermix of life cycles,
>although again they all pass through a common phase, namely trophozoite
>within host cell, so again we might be able to declare that as the
>phase for "start" of life cycle where we measure fecundity, just barely
>again.
>- Oomycota, specifially Saprolegnia, again have two interlocked life
>cycles, but the single point of intersection is hypha, and I'm not sure
>you can count how many there are in order to compute fecundity from
>that point to the same point in the next generation.
>- Angeiospermophyta, specifically vines such as blackberries or grapes,
>where there is both sexual reproduction, and vegetative growth and
>happenstance splitting of a single "plant" into separate plants. How
>can you count how many "plants" there are, in order to compute
>fecundity number???? How can you even decide where to draw the line
>between "generations"??
I don't care how complex the life cycle is. One complete cycle is from
one stage to the comparable stage of the next generation. From
medusa to medusa or from sporozoite to sporozoite or from gametophyte
to gametophyte. And ecologists have ways of measuring population size
in organisms that reproduce in complex ways. Look it up.
You should learn about real populations that are in close
approximation to steady state. That is why I say that the ability for
any one specific organism to leave a copy of its alleles in the next
generation (one cycle) is stochastic, probabilistic, contains a random
component. The probabilities need not be uniformly distributed over
phenotype or genotype. Hence, the process exhibits both drift and
selection.
an...@sci.sci
> frequencies of all alleles, which you seem to have here?
Your question is nonsense because I don't have any such idea.
Under neutral descent-with-modification, whatever the initial
distribution of allelles is, so long as it's not close to either
endpoint (fixation of one or another allelle), and so long as the
population is large, the mean distribution expected after one
generation is exactly the same as the initial (parent) distribution,
and the standard deviation of that expectation is dependent on the
total population size. After N generations, the mean remains fixed at
the original distribution, while the standard distribution grows with
(the square root of?) the number of generations. If N is sufficiently
large, the chance of hitting one of the endpoints needs to be taken
into consideration.
For example (**only** an example), if the initial distribution is
50-50, then the mean remains 50-50.
If the initial distribution is very close to one of the endpoints, then
N doesn't have to be very large before there's considerable chance of
hitting an endpoint already, and with the initial distribution only one
individual different from an endpoint we get a 50% chance of fixation
after just one generation, making the mean/sd analysis wrong from the
start. That's why I chose an initial distribution, for my one example,
at 50-50, to avoid any chance of running off either end at the start
(assuming a total population of any decent size, such as a thousand).
> You understand that neutral evolution will always result in either
> fixation or extinction of the allele, right?
No. In the absense of new mutations that would be true, but if wishes
were horses we'd be hip deep in manure, as you must surely know. If the
expected time for fixation is much larger than the expected time to
next mutation, then there's very high chance of fixation before
mutation, and even if mutation happens there's another period of high
chance of fixation before the second mutation, etc., and over many
cycles of mutation the overall chance of fixation approaches 1. But if
the expected time for fixation is much larger than the espected time
between mutations, then it's very very rare that the number of allelles
would ever reduce to just 1, instead the number of allelles would
increase to 3 and then extinction would reduce the number back to 2
which doesn't yet fix either one. When considering a whole haplotype
block, the equilibrium between increase in block-allelles (due to new
mutations) and decrease in allelles (due to extinction of one of the
block-allelles) would be at a rather large number of allelles, that
number dependent on the population size. Not only wouldn't fixation
ever happen, but the number of block-allelles wouldn't even get close
to such a small number during the entire lifetime of this particular
haplotype block (until it splits into two blocks at which point we have
to stop our calculation and declare the old intact block *never* got a
single block-allelle fixed).
> Yes, we attempt to detect selection by looking at nucleotide diversity
> in a small region. But we don't therefore suppose that an entire region
> is under selection, merely that some site within that region is under
> selection.
With hitchiking, especially within a single haplotype block which
remains intact (not split, zero internal crossings) for longer than the
lifetime of a species, effectively all the sites within a block are
under group selection. If just one site is being selected (by
difference in fitness caused by allelles of *that* one site), the
entire block is effectively selected as a unit, group selection par
excellence. HItchhiking within an unsplittable haplotype block is
hitchhiking to the extreme.
So yes I agree there may be only one site that actually causes a
phenotype fitness difference hence causes the selection of the group,
but the entire group is in effect selected. It's just like in
prokaryotes where the *entire* DNA loop is selected as a unit, so if a
really good mutation happens in an individual, but an even better
mutation happened in another individual (not its ancestor or
descendent), the first mutation goes extinct eventually, because it's
the whole DNA loop that matters, not any single site on it, and it's
forever competing with that better mutation, never joining forces to
have a best-of-both-mutations effect. (I'm assuming zero horizontal
gene flow for this analysis. If one of the two fine mutations can get
laterally copied across to the other clade, then best-of-both-mutations
can occur within a single individual, saving the second-best mutation
from going extinct.) (Note: I'm assuming the two mutations are
compatible together of course. If each is good, but the two together
are worse than either alone, the above doesn't apply.)
> Or, to put it more reasonably, with present technology it's impossible
> to tell whether the allele is being selected or is hitchhiking.
Correct. With two sites close together but still occasional crossings
between them, it's difficult to tell which is causing the selection in
a single generation, takes longterm breeding, or deliberate forcing of
cross-over, to tease them apart. With them in the same haplotype block,
there isn't enough time in the species to ever tease them apart without
artificial splicing/crossing. Only if you can discover that one of them
codes for a protein, and the other doesn't, and then by artificially
manipulating the level of that protein directly (supplying it as a
"vitamin", or introducing something that cleaves it apart), can you
demonstrate easily that the protein-coding gene really is the site
responsible for the selection.
> > ... I claim that whereever there's a haplotype block
> > that persists over tens of thousands of years, due to the lack of even
> > one hot or even warm crossing site within it, for all practical
> > purposes (for evolution studies over such time spans and any shorter
> > spans) the various sites on that block are co-evolving as a single
> > unit, with selection pressure on them all merged into a single
> > effective selection pressure on the group, so it's group selection that
> > we need be concerned about here, and the actual selection on individual
> > loci within that block can't be seen by any experiment.
> Careful. You just used the term "group selection" when all you meant was
> the sum of selection.
What's the difference? Sum-of-selection-factors, among various sites
that are inseparable, is how group selection works.
> Now in actuality, we could do experiments on
> individual loci or individual sites by genetically engineering changes,
> even if in nature the blocks never recombined internally.
Yes, we're in agreement. In *nature*, some of the blocks are kept
intact longer than the lifetime of the species, so group selection
occurs in an absolute sense. In other blocks, they get split maybe once
or twice during the lifetime of the species, so they engage in group
selection for long spans of time, but not absolutely for the lifetime
of the species. In the genetic-engineering lab of course we can force
changes that would never occur in nature or would be very very rare in
nature.
> We could also
> examine the specific effects of individual differences on physiology,
> anatomy, or whatever by looking at the actual biological mechanisms more
> closely, and that would let us predict the effects on fitness of
> individual bits, even if they were never separated into different
> individuals.
Yes, such as the coded-protein vitamin-or-destroy experiment I
described a couple paragraphs back. We seem to be in agreement about
the science.
> > we might directly observe deaths of
> > individual cells/organisms, perform autopsies on each to determine
> > cause of death, and thereby attribute some allelle of some locus at
> > fault for the death, and by compiling statistics of such
> > causes-of-death we might then calculate selection pressure for each
> > non-neutral locus. But that's not yet feasible, ...
> Actually, similar sorts of studies are quite feasible with current
> technology.
Wow, that's news to me. Where'd you read about it? I haven't yet seen
anything in _Science_ that implies such technology exists yet.
> I'm also not sure what sort of point you are trying to make here.
It appears we agree on the science, we merely disagree on how
to express it in English, so that's what we've been discussing.
But this all started when I simply asked what the major aspects of
evolutionary theory were, so that I could make up a check-list for
anti-evolutionists and naive freethinkers to say where they agree and
where they dispute the theory, so that we can target our education to
precisely the places where there's a problem. My latest checklist has
three lines of evidence and three aspects of theory as the core part
for the initial/minimal checklist, and I think I'm going to stick with
that checklist and then move on to part 2, the survey of who agrees
with which parts and disputes which other parts.
> > Aha, yes in *that* sense even with hitchhiking the particular locus is
> > still neutral. But still thinking of it all as a "single site" may be
> > the best way.
> I don't think it is for most purposes. The purpose for which it's the
> best way is when you are computing a fitness, and you want to know which
> block is going to be fixed.
Well, I agree, it depends on the purpose. For all studies of natural
population statistics, what "evolution" really means, change of
statstics of allelles over time in **natural** populations, it's the
haplotype block which is the basic unit of statistics and evolution.
But in the lab, when you're trying to understand *why* a particular
block has this allelle increasing in frequency and that other allelle
decreasing, but it's drifting in the other direction in a different
population in a different part of the world, discerning the individual
effects of each difference along the block, and why each difference
makes some selection (or doesn't, is neutral), and then adding up all
these selections and checking whether your block-total matches
selection in nature, then of course you need to treat the block as
non-atomic.
> Or, more reasonably, you could say that frameshifts can change the
> selective value of different nucleotides at a given site.
Yes, that's saying the same thing in different words. You tend to
sometimes know just the right jargon to express concepts more
succinctly than I did. If I ever write a book, I'll have to ask you to
help me tighten the language so it's only 250 pages instead of another
monster like Atlas Shrugged or War and Peace. (Those are each very
long, circa 1000 pages each, right? If I'm mis-remembering, sorry
Rand and Tolstoy.)
> >>A recent study of a long stretch of avian DNA shows that sites as
> >>little as 500 bases apart are effectively unlinked over very short
> >>timespans.
> > Does that observation apply to *all* pairs of sites located at that
> > pairwise distance, or only to carefully selected sites that straddle a
> > recombination hot spot?
> As I recall, they could find no significant linkage between any two
> sites separated by 500 bases or more.
Ah, that's very different from the human genome where there are
haplotype blocks averaging 10k bases long that haven't been split since
the pre-human / human bottleneck. So haplotype blocks, if any, in
whatever species they were studying there, may be very much smaller
than in humans, probably smaller than a single gene, but possibly big
enough to include an entire active motiety of a polypeptide that is
part of a gene.
Major line of research that is needed to understand this difference:
What is the precise mechanism behind those "hot spots" of recombination
and the vast majority of other places there's hardly any recombination
whatsoever. Are there RNA sequences that bind to DNA at hotspots and
direct cleavage enzymes to attack precisely there? Or are there RNA
sequences that bind everywhere else to protect them from cleavage?
Given the difference between birds and humans, I am now leaning toward
the latter as the mechanism. Some random goings on, perhaps free
radicals, perhaps O2, perhaps nitrogen oxide (the one that was recently
discovered to be a neurotransmitter and also a hormone, I think it's
NO, not NO2 or N2O3, right?), tends to cleave DNA just about
everywhere, but human ancestors have evolved a way of protecting most
of DNA, whereas birds don't have that mechanism so they get cleaved
everywhere. Is NIH or NSF etc. funding any such research? It would seem
to me to be a very important line of research that is very timely now
that HapMap1 is done and HapMap2 is funded.
> If you have access to a scientific literature database like BIOSIS or
> Medline, you could look this up.
No, I am not aware of having any such access. Remember, I'm just a
layman in biology, not a working (funded!!) scientist.
> PubMed is freely available and probably has what you want. Pick a few
> search terms and go to it. (I commonly use PubMed through the NCBI web
> site, though I'm sure there are other ways.)
Let me do a Google search right now to see if I can find the free service...
<http://www.ncbi.nlm.nih.gov/entrez/query.fcgi>
PubMed is a service of the National Library of Medicine that includes
over 15 million citations from MEDLINE and other life science journals
for biomedical articles back to the 1950s. PubMed includes links to
full text articles and other related resources.
The citations wouldn't do me any good, because I don't have any access
to a professional library, but if the full articles are online in text
format that would be nice, although between citation and full-text I'd
really need to see an abstract. Continuing with Google search to try to
find the actual PubMed search engine access point ... Oh, at the top
of that Web page there are two lines:
Search [PubMed___________] for
_____________________________________________ Go
Gonna give it a try ... Searched for recombination mechanism, didn't
see any reasonable matches, although one abstract on a different topic
was interesting:
<http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=pubmed&dopt=Abstract&list_uids=16316003&itool=iconabstr&query_hl=1>
[The uncertainty of "fitness:" what prevents understanding of
the role of genetic exchange]
[Article in Russian]
[No authors listed]
The evolutionary development of highly organized species is
attained through an increase in average survival of
individuals, whereas the evolution of primitive species
involves only an increase in fecundity (Zavadsky, 1958, 1961).
However, in population genetics, survival (or ecological
resistance) and fecundity are regarded as components of a
single character, fitness. Employment of the notion of fitness,
which lacks a strict definition, hinders understanding of the
mechanism of progressive evolution as the process that enhances
ecological resistance of organisms. The notion of fitness also
exacerbates understanding the role of genetic exchange, since
the primary advantage of genetic recombination and sexual
reproduction apparently is producing of progeny with high
ecological resistance rather than with high genetic diversity
as such.
I tried adding meiosis as a keyword, and even haplotype, and still it
didn't produce any reasonable matches. Do you have any idea what
keywords I should use to find articles describing attempts to discover
what the mechanism is that causes recombination to occur during
meiosis?
.
an...@sci.sci
> landscape.
But they are eliminated immediately, within one generation, so in the
context of evolution they play no role whatsoever. They don't in any
way contribute to finding a high peak in "fitness landscape". So they
can be completely ignored in explaining to newcomers what causes
evolution to yield new species better than ever before.
> Selection pressure, a term that you disparage, occurs only after
> mutation has introduced the new allele into the mix.
It's an illogical composition of words. The so-called "pressure" acts
upon statistics of populations, increasing the frequency of one allelle
at the expense of another. The so-called "pressure" is caused by
differential success at full-cycle success, which is a conflation of
fecundity-number and survival-probability. (If the fecundity-number is
6, and the survival-probability is 1/3, then the full-cycle growth
factor is 6 * 1/3 = 2, i.e. twice as many individuals each generation
as the one before, which of course continues only so long as conditions
are the same, in particular conditions change almost immediately due to
limits to resources available, hence the typical exponential expansion
followed by slowing down to linear followed by slowing down to
exponential-decay-from-full-filling-of-habitat, the S-shaped growth
curve. (For full-cycle growth factor less than 1, the allelle frequency
shrinks exponentially to zero until the total number of individuals is
so small that computing exact probabilities of different sizes of
populations is better than the exponential approximation.)
A more accurate compositon of words would be something that includes
the word "change" and the phrase "population statistics", but I can't
think of any reasonably brief such composition.
So you like the misnomer "selection pressure" while I dislike it.
Big deal.
But we agree on the science, that the directional bias in the expected
change in population statistics, caused by difference in fitness,
happens *only* after there are two allelles that have different
fitness, one allelle ancestral and the other generated from it by
mutation. (Or if there's another mutation before one of the first two
allelles has gone extinct, then there will be three alleles at the same
time, with more complicated math but the same basic idea.)
> Nothing in my
> statement even suggests that I am listing all the factors behind
> evolution. I am only describing one of them.
The actual factor is a stochastic relationship between individuals and
the live/die or make-N-babies result, whose expected mean is based
purely on the two fitnesses, but whose standard deviation is based on
population size and the nature of the environment (nice and stable
gives small s.d. whereas lots of unpredictable disasters gives large
s.d.).
That one factor can be analyzed per mean and s.d. separately, thereby
separating bias from noise. Since the bias (mean) is caused by fitness
difference, I accept if you call it one factor instead of only part of
one factor.
> I was also making the extremely important point that you seem to have
> missed in earlier posts, and whose significance you snipped away,
> about the population nature of evolution.
"A half truth is a lie."
Evolution has three factors:
- Ideal replication of genome.
- Mistakes in replication, or mistakes in maintaining DNA between acts
of replication, which generate new allelles.
- Change in allelle frequencies, which is stochastic, which can be
analyzed as part bias (mean change, or "selection pressure") and part
noise (random drift).
Only that third factor deals with populations (statistics thereof).
When you say "the population nature of evolution", your words could
mean either:
- *The* (one and only) nature of evolution *is* population based.
- There are several natures of evolution, and population is just one of them.
The first is false. The second is true.
Your words could mean either of them, I have no idea which you meant.
Please state which you meant.
.
an...@sci.sci
> > just collect evidence regarding some particular kind of phenomen.
> This is very seldom done. I can't really think of an example right off.
> You almost always start with some kind of theory. That theory may be
> based on evidence collected for some other purpose, by someone else.
Fine, then tell me as best you can discern, when the very first fossils
were dug up by accident during construction or along stream beds due to
washing out of canyons, why did anybody to keep the fossils instead of
just throwing them in the cement grinder as worthless pieces of junk
irrelevant to anything they were doing at the time? Why were the
first ten fossils *ever* collected instead of just tossed away? Surely
the people who happened upon the first ten fossils didn't have a theory
they were researching where fossils would be useful to decide in favor
of or against their theory of aqueduct construction or basement
construction or riverside fishing or surfing or sandcastle buiding or
limestone mining or coal mining or swamp draining or tarpit removal
etc.
So when we're trying to explain the evidence that supports evolution,
should we collect evidence to support Noah's Flood and discard all
evidence contrary to it? That's really the only theory of burial of
bones that people might have had at the start when the first fossils
were discovered and retained to support a theory rather than discarded.
> both of your chosen forms of evidence seem to rely on identifying
> actual fossils ...
You aren't a good reader, are you. Please look at:
<http://groups.google.com/group/talk.origins/msg/b8bc89c9f330eafa>
and see where I list e1 e2 and e3. Note carefully not two but three
different forms of evidence, and only one not two are based on fossils.
> Instead you should be thinking about nested hierarchy.
Nested hierarchy is just one of the three forms of evidence.
Why do you think I should ignore the other two??
Nested hierarchy by itself doesn't imply any sort of change over time
whatsoever, except within a single individual which develops from
zygote through ball and blastula and embryo and juvenile to adult,
which has nothing to do with the evolution of species over time.
All that the nested hierarchy by itself says is that God delegated
authority to committees, one committee for each kingdom, and those
committed delegated authority to phyla-subcommittees, which delegated
authority to class-subsubcommittees, etc. all the way down to
genus-subsubsubsubsubcommittees each of which designed several species
per their particular idea how that genus should look and act.
And all the angels on those committes could have met on the head of a
pin of course, including Gabriel, the chairman of the Homo sssssc.
The only way to show any relation whatsoever between the nested
hierarchy and geological time is to have fossils, so if you omit that
form of evidence you kill any way to demonstrate or test the theory of
evolution.
> It's probably not a good idea to think of these as separate. The first
> and third especially seem like just two ways of saying the same thing.
> The middle one is a garbled form of something that paleontologists
> sometimes do, which is to test the stratigraphic fit of their
> cladograms. It usually comes out as highly non-random despite the patchy
> nature of the record.
No, first and third are not at all two ways of saying the same thing.
One nested hierarchy deals with physical characteristics, which are
presumably caused by DNA differences but we haven't yet sequenced more
than about twenty species, and we don't yet know which DNA sequences
are responsible for even half of the physical characteristics. This
form of evidence completely ignores the 99% of the genome which doesn't
code for any phenotype characters. We are just starting to understand
how DNA codings causes embryology to develop in a particular way.
Thus physical characteristics and enbryological studies are almost
completely independent of DNA sequences so-far. If we generate a nested
hierarchy (cladogram), from clustering of characters, we have no idea
whether the common ancestor of a cluster really did look like the
average of all current-day species, in fact we have good reason to
believe the common ancestor did *not* look like the centroid of
present-day cluster-members.
The other nested hierarchy deals with DNA sequences, which are 99.9%
neutral, in fact it's the neutral sequences which give the best
cladograms for discerning evolutionary relationships. In most cases we
must discard the 1% that code for anything and compute our cladograms
only on the basis of the neutral DNA segments. If we do that, the
evidence we have here is completely separate from the
physical-characteristics cladograms we get elsewhere. Furthermore, DNA
sequences allow us to actually predict much of what the common ancestor
had, with reasonble assurance we are predicting correctly.
The DNA sequences give us some reason to think we're dealing with
evolution over time rather than committees of angels working for the
Great Omphalos Designer, but really that's just a bias in our minds.
Actually for all we know, the angel committees might have been required
to include the neutral DNA sequences as a sort of copyright mark, with
each sub-committee inheriting the basic mark from its parent committee
and then mutating it randomly to prove authorship of the sub-commmittee
product while still showing responsibility to the parent committee.
So if G.O.D. is dissatisfied with some individual, he can find out the
responsible party by starting from the toplevel kingdom committees,
checking which toplevel committee has a digital signature closest to
that of this individual, then going to that committee and scanning its
sub-committees to find which sub-committee has the sub-signature
closest to the individual, etc. working His way down through the six
levels of committee hierarchy until he reaches the correct Genus
ssssscommittee, and calls those members onto the carpet to show the
paperwork for this particular species, whereupon he finds out that the
viper-snake committee had an angel named Satan do that design, which
was so grossly faulty that G.O.D. cast Satan out of the committee, to
exile in a rather unpleasant place, and replaced that member by a
different angel who then did a better job, unfortunately after Satan's
viper did something really bad that messed up humanity forever after,
but at least the talking viper was exterminated and replaced by a
non-talking viper.
You see how the YECs could, with only the e1 and e3 evidence, come up
with a perfectly valid theory with fixed species all individually
designed by angels on authority of the Great Omphalos Designer? That's
why we absolutely *must* include fossil evidence if we hope to convince
them to abandon literal Genesis and accept evolution as a chronological
fact rather than just an illusion from angel-committed Intelligent
Design. (Actually most/all YECs are too stupid to be capable of
inventing a theory that satisfies both literal Genesis and both of the
nested hierarchies, as I did just now, well after midnight when I
should have been in bed already. So now that I've given the YECs
exactly what they want, a scientific and religious theory that explains
two third of all evidence, and now they can quote-mine me, are you
going to crucify me, or tar and feather me, or ban me from the
newsgroup, or invent a time machine and go back and make me go to bed
when I should have so I wouldn't have ever posted this?)
Hey, teach the controversy! I propose that *my* theory of
angel-committee intelligent-design be taught in Kansas schools. Explain
just the first and third types of evidence, and convince the
high-school students that my angel-committee ID theory is just as valid
the theory of Darwin/Wallace. Let them go home after the first week of
classes fully convinced of that equal validness of the two theories.
Let them have a whole weekend thinking their Kansas biology teacher is
a YEC. Then first thing Monday morning show them the fossil evidence,
and watch them all turn red at their foolishnesses.
Or maybe ecosystems are irreducibly complex, so you can't just design a
bunch of individual species in different committees and hope them able
to survive together in food webs. Maybe the toplevel committes needed
to design prototypes for each kingdom, several strains of bacteria
(first the chemoautotrophs, then the sulfur photoautotrophs, then the
cyanobacteria, etc., to establish a primitive ecosystem that had all
the basic major parts functionning together, and then the second-level
committees had to introduce the various phyla into this
already-existing ecosystem, etc. until finally the modern species were
created, culminating in Adam and Eve as the very last species created.
So the hierarchial angel-committee theory, together with the ID concept
of need to create a crude ecosystem before most of the rest of the
varieties of life could survive, fits all three lines of evidence now.
Do any of you wish I'd gone to bed two hours ago?
.
an...@sci.sci
> > to ask who accepts it and who disputes it, and then for whose who agree
> > to stipulate the evidence is correct, a nice organization of the
> > theories of evolution and common descent, and ask them whether they
> > accept those conclusions. For those who accept all the evidence, and
> > evoution and common descent, then I ask them what theory, other than
> > mutation + replication + selection/drift, they have to explain the
> > cause of evolution and common descent.
> have a strange way to go about it.
No, the talk-original FAQ does not have any place where it nicely
organizes the various lines of evidence and the various factors of
theory. It merely answers a lot of stupid questions from YECs and
IDiots, giving point-by-point rebuttals, but not putting the whole
thing into any sort of coherent whole.
Likewise I've never seen anybody before me posting the list of basic
kinds of evidence and various aspects of theory. If you can cite
anybody before me who did, please find it in Google Groups and cite the
message-ID and/or GG URL. Put up or shut up.
> As I recall we began this argument
> because you were trying to use the evidence above as some kind of
> argument for natural selection, not just common descent.
Yes, that's correct. You remember the original well. It's just the
middle you seem to have failed in reading comprehension where you cited
two instead of three types of evidence. Maybe that was just a slip-up
in your otherwise fine comprehension.
Anyway, I got harassed by nitpicking so much, led off the main topic to
detailed debate over language expression and mathematical factor
analysis and subtle effects of stochastic mechanisms that for a day or
so I forgot what I was originally trying to cover.
I now think I need to break the overall checklist into four sections:
- p1 p2 p3 p4 p5 p6 p7 p8 (prerequisites: atomic theory, crystal
forming, radioactive decay, sedimentation, fossilization, plate
tectonics, biochemistry, genetics)
- e1 e2 e3 (evidence: taxonomy, fossils, DNA cladograms)
- f1 f2 f3 (facts: lines of descent, branchings apparent from fossil
evidence, and the two unrooted cladogram trees and the one forest of
rooted fossil trees fitting together nicely to yield common ancestry)
- t1 t2 t3(a,b) (theory: replication, mutation, differences in
survival factored into mean (bias, "pressure") and random (noise, "drift"))
Regarding my proposed poll I've been working on developing:
> You have a strange way of going about it.
Well I think the checklist of items to stipulate (legal sense) or
dispute is a reasonble way to find out exactly where the others don't
accept the modern synthesis of neo-Darwinism.
I think I'm settled on the items in the checklist. Next I need to word
each as a poll question. For example: The modern theory of evolution is
based on several facts about nature, one of which is that all material
is divided into atoms, and those atoms are of various types based on
the number of protons in the nucleus, these types we call "elements",
such as Hydrogen or Oxygen, and for many of these elements there is a
further division into "isotopes" based on the number of neutrons
present in the nucleus, such as normal Carbon 12 (6 protons and 6
neutrons) which is the common isotope and Carbon 14 (6 protons and 8
neutrons) which is less common. Agree [ ] / Disagree [ ]
> > I really want a check-list of evidence, accept or deny, and then a
> > check-list of theory, accept or deny, so we can pin down where exactly
> > the others disagree with modern evolutionary fact or theory.
> Good luck on that. I've had trouble getting creationists to admit what
> they believe, especially in any consistent fashion, or to accept the
> consequences of what they believe.
I'll make the first few poll questions, such as the one above on atomic
theory, so very obviously correct that nobody except a pre-school child
or a Zen buddhist would disagree. Once they get warmed up on those
first few questions, then when I get the ones they might possibly not
agree with because they never heard of any such ideas, such as magnetic
impressions in solidified ferromagnetic materials, they can learn a
little science, without yet attackig their religion, after all there's
nothing in the Bible saying magnets and magnetic compasses are
impossible or that it's impossible to create a magnet by subjecting it
to a strong fixed magnetic field, or anything like that, and Bible
thumpers might actually get a kick out of understanding why lodestone
has any intrinsic magnetism in the first place (because was liquid then
cooled in the presence of Earth's magnetic field) thereby allowing it
to be used as a crude compass by hanging it from a string.
> > Then that matrix is fed to such-and-such computer program
> > which generates the best-fit tree, if any, and shows a measure of
> > fitness
> Better call that "fit"; fitness has a different meaning.
Oops, I was thinking the mathematical term was "test of fitness" but I
did a Google search just now and see it's called "Goodness-of-Fit
Test". Thanks for the correction.
> > such as a P (confidence) value for tree-ness (and/or that
> > specific tree model) preferred over the null hypothesis of random
> > number input not belonging to any tree? It would be nice to
> > have two such examples, one very small, such as apes, only five or ten
> > species, such that the whole thing could be processed by hand or at
> > least the final result verified by hand, and one much larger to show
> > how really interesting trees can be demonstrated.
> Only the one I did myself with mtDNA sequences. Did you read that post?
> I've put it up many times.
I don't remember. Please try to remember a few keywords you had in that
article, and then do a Google Groups search to find the archive copy,
and then tell me the mesasge-ID and/or the URL, and I'll take a look
and see if it looks familiar. Don't post the whole thing again if there
are already several copies online.
> > If within a single geographic area there's an isolated species that
> > lasts a few million years, then immediately afterward there's another
> > species only slightly different from the previous but clearly showing a
> > significant change from the previous, and it's in the same area as the
> > first, we can reasonably guess that the second descends from the first.
> No, we can't. We could if we could be confident that there had been no
> movement between unsampled geographic areas, and that we had sampled all
> species in that particular area, and that we could actually recognize
> which individuals belonged to the same or different species. But we
> really can't do any of those things.
I didn't say such an observation is strong evidence for the hypothesis
being true. I merely said it's reasonable to invent that hypothesis to
explain it. Then if there are many other similar linked species pairs,
we might generate a general hypothesis that one species descending from
an earlier species is a common occurrance. Once we have that
hypothesis, that it's common, not a one-time occurrance, *then* we can
perform statistical tests to see whether that general hypothesis seems
to be validated by the data. But science works only with general rules,
not one-of exceptions, so we need to have many such apparent lines of
evolution before we can even propose anything that can be tested.
So my evidence is here's an apparent line of descent. Here's another.
Here's another. Here are a hundred. Hey, do you think there's something
going on here? What do you think that might be? Yeah, that's right,
evolution from one species to another, some kind of descent with
modification, either direct biological descent, or some Intelligent
Designer who lived in that local area and released one edition of the
genus after another and never acquired means to ship his product beyond
the local area, but in any case new species whose design is somehow
modified from design of corresponding earlier species, regardless of
whether that "design" is intelligent design or merely apparent design
due to some physical mechanism, in any case the design seems to be
evolving locally.
> > Now Darwin's finches wouldn't be like that. He observed a whole bunch
> > of species, which all seemed be be variants upon a common theme, and he
> > guessed that they all descended from a common ancestor, but his "tree"
> > had a single level wherein all modern species suddenly appeared in
> > parallel, no successive splittings over time that he could discern
> > without any fossils to look at.
> In fact Darwin didn't even propose a tree for Darwin's finches, just
> common ancestry for the lot.
That's basically what I said. He envisioned common ancestry, with no
idea how the various sub-clades might have occurred over time,
whereupon the mathematical graph, with a single massive unresolved
branching node, doesn't look at all like a tree, so Darwin would have
looked silly drawing a tree like this:
ancestral finch
|
+-----+-----+-----+-----+-----+--+--+-----+-----+-----+-----+-----+
| | | | | | | | | | | |
DF1 DF2 DF3 DF4 DF5 DF6 DF7 DF8 DF9 DF10 DF11 DF12
It probably didn't even occur to him to display the common ancestry
like that, since obviously it's wrong, obviously there was sub-clading,
it's just that the details of such sub-clading were unknown to Darwin,
so rather than draw a really stupid tree, just wait until some
sub-clade information might someday be available, and let that later
researcher draw a nice tree-looking tree showing what he discovered.
> However, we don't need fossils to determine this successive
> splitting. There are several papers on the phylogeny of Darwin's
> finches, the most recent using DNA sequences.
Sorry if I sound like a Creationist attacking evolution, but:
If we believe with all our hearts that *all* variation of species in
nature is caused by evolution over time, *not* by the result of
hierarchial committees of angels (I hope you saw the first part of my
followup I posted about 2AM PST), then whenever we compute a cladogram
based on DNA we can just take it for granted that our cladogram
represents a family tree of evolution over time. But such a cladogram
by itself does nothing to prove common ancestry if not a single fossil
had ever been discovered from a species of the past.
> I don't know of any "chains", just trees.
Every monotonic path through a tree is a "line" of descent, which is a
"chain" of successive links between adjacent species.
Two paths are concident until the point where a species-split
("speciation") event occurred, then the two paths run in parallel from
that point onward until one or the other goes extinct.
> You can turn trees into chains by ignoring some of the branches, if
> you really want a chain.
If you're trying to solve a maze, but it has multiple solutions,
if you're really smart you can find all the solutions to the maze (all
the paths from start to goal) without needing the crutch of erasing all
the other paths while looking at one of them. You can just "see"
several paths simultaneously, recognizing each of them as *a* solution
to the maze, recognizing that each is not unique, that there are others
in addition to any one you are looking at.
You don't have to ignore the other branches, you can see more than one
chain all at the same time, up to five if you are human, up to seven if
you are a bird. :-)
> > I believe such trees would show low levels of confidence for some
> > branchings, expecially pre-Cambrian, and if the program is forced to
> > exclude all low-confidence parts of the tree, it would yield a bunch of
> > separate trees instead of just a single tree. (Some programs would
> > simply refuse to give any result at all.
> Not true. I only know of one such program, and I doubt sincerely that
> you have ever heard of it.
I'm a bright guy who hates cruddy software, and often I think of ways
the software could have been better. I don't need to see an already
existing better program to suggest how the other cruddy software could
have been better. If I saw a program that broke horribly, or which
generated a grossly unsupported toplevel node in the UCA cladogram, I
would surely suggest the program be fixed to simply report separate
trees that can't be confidently joined, and thereby anticipate that
wonderful but secret program you know of. By the way, what's its name?
> The node of all eukaryotes, for example, is a very high confidence
> one.
Are you merely saying that eukaryotes cluster grossly distinctly from
all prokaryotes, so that we can be really sure eukaryotes are a single
clade? But what about the next level inside eukaryotes. Do we know with
high confidence how eukaryotes divide into exactly two sub-clades?
And when you say eukaryotes, are you considering only nuclear DNA, or
only mitochondrial DNA, or both of those conflated together? I've seen
a lot of published reports in _Science_ which say the same sort of
thing you are saying but never say which part of their DNA is being
cladogrammed. (Is that a verb??)
I can't finish tonight. I gotta go to bed now.
.
John Wilkins
...
>>In fact Darwin didn't even propose a tree for Darwin's finches, just
>>common ancestry for the lot.
>
>
> That's basically what I said. He envisioned common ancestry, with no
> idea how the various sub-clades might have occurred over time,
> whereupon the mathematical graph, with a single massive unresolved
> branching node, doesn't look at all like a tree, so Darwin would have
> looked silly drawing a tree like this:
>
> ancestral finch
> |
> +-----+-----+-----+-----+-----+--+--+-----+-----+-----+-----+-----+
> | | | | | | | | | | | |
> DF1 DF2 DF3 DF4 DF5 DF6 DF7 DF8 DF9 DF10 DF11 DF12
>
> It probably didn't even occur to him to display the common ancestry
> like that, since obviously it's wrong, obviously there was sub-clading,
> it's just that the details of such sub-clading were unknown to Darwin,
> so rather than draw a really stupid tree, just wait until some
> sub-clade information might someday be available, and let that later
> researcher draw a nice tree-looking tree showing what he discovered.
...
If you look at the Origin, you will see that he did draw a tree indicating
what he expected evolution to "look like":
http://bio.research.ucsc.edu/~barrylab/Shawn/Graphics/Biology%20175/darwin.phylogeny.jpg
He clearly expected that there would be periods of nascent speciation, and
comments in the Origin indicate that he thought that widely distributed
species would speciate more. He certainly had no problem with polytomous
trees, but the overall "shape" is dichotomous. Whether that meant anything to
him, I can't say - I have never seen anything he said that indicates one way
or the other.
--
John S. Wilkins, Postdoctoral Research Fellow, Biohumanities Project
University of Queensland - Blog: evolvethought.blogspot.com
Nihil tam absurdum quod non quidam Philosophi dixerit - adapted from Cicero
John Harshman
>>I wasn't talking about independent assortment between haplotype blocks.
>>I was talking about the very existence of haplotype blocks as stable
>>entities.
>
>
> Nothing is absolutely stable. Even protons will eventually decay.
> The question is how stable are they, stable enough or not?
> It's my understanding that among the 3.1 billion
> adjacent-base-junctions in the DNA strands in our human genome, only
> 0.01% of those have crossed-over even once in all our history since the
> major bottleneck that probably caused a founder effect to yield our
> present species.
Bottleneck? Founder event? What evidence exists for that?
[snip]
>>>... The entire chromosome is a single linkage group, by such a
>>>daisy chain of A-link-B-link-C-..., assuming markers are known
>>>sufficiently close each to the next.
>>
>>That's not the way it's generally used in my experience.
>
> What experience is that?
Just my general impression from reading the literature. The term
"linkage group" has most often, that I recall, been used for regions
that haven't experienced crossing over during some period under study.
Of course I have a biased sample of the literature, just things I'm
interested in.
[snip]
> Are we in agreement, except your particular lab is using only a few
> markers so the first part of that applies relative to your set of
> markers, but if using all known markers what I said would apply, per
> the last quote above?
Actually, the only linked loci I have ever used at all are
mitochondrial. I will agree that there seem to be different definitions
in use. One of those definitions is fairly close to the definition of a
haplotype block, and others are not.
>>><http://www.iscid.org/encyclopedia/Open_Reading_Frame>
>>> A reading frame, in biology, consists of three-nucleotide codon sets
>>> in either DNA or RNA that are contiguous and non-overlapping. An open
>>> reading frame (ORF) is a similar sequence that can be translated into
>>> a protein or a polypeptide.
>>> In any open reading frame, the start-code sequence or initiation codon
>>> that begins the protein is methionine ATG, and then stop-code sequence
>>> or termination codon ends it. The stop sequence is coded by what is
>>> termed a nonsense codon, or a codon that does not have an RNA match.
>>> There are only three nonsense codons: amber(UAG) ochre(UAA) and opal
>>> (UGA). As you can see, each one contains a "U" nucleotide, not normal
>>> to DNA.
>>>(That last one is bonkers!! Who wrote it, Archimedes Plutonium??)
>>
>>It's convoluted and contains much unnecessary information, but it's
>>really the same as all the others, which match each other.
>
>
> Well when I said it was bonkers, I was referring mostly to the claim
> that the U nucleotide appears in the DNA genome. Is that correct or
> bonkers?? If the U is in DNA, how does it get put there during
> anti-replication from the opposite strand of DNA?
It's common to represent codons as RNA, since that's what is really
translated. But yes, that last sentence reveals confusion on the part of
the writer. I hadn't read that part closely. Strangely, it gives the
start codon as ATG, not AUG. Definite weirdness.
>>Gene-estimation programs are used to identify genes, introns,
>>exons, and such, and they rely on a large set of clues.
>
> There's a difference between the following claims:
> - Our best estimate, based on all those clues, is 30k genes in human
> genome, of which only half have been proven to be genes, by means of
> showing that they either code for polypeptides or code for regulatory
> RNA or have some other known function.
> - There *are* 30k genes in human genome, of which only half have known
> function.
There is indeed. But if you have a point it's disappeared along with the
context.
>>><http://www.thedoctorwillseeyounow.com/articles/other/genome_4/index.shtml>
>>> It is believed that
>>> the human genome has between 50,000 and 100,000 genes.
>>>(Somebody needs to update their Web site!)
>>> It is estimated that the sequence of the human genome should be
>>> completely mapped by approximately the year 2005.
>>>(Yeah!)
>>
>>Ah, internet paleontology, or vestiges of the natural history of web
>>creation.
>
> Yeah. In that one case, I'm complaining about an article from years ago
> which is still online as if were current information, accessible from
> Google search (how I found it), with no warning that any of it is out
> of date, and no way to post a followup warning that it's obsolete. In
> such a case I don't expect the author to delete it, but it would be
> nice if somebody would just indicate which parts are hopelessly
> obsolete, compared to the main bulk of the old article which describes
> some particular methodology which is still the correct description and
> hence still eduational.
>
> But in the other cases I cited, I presume you can see the problems with
> the conflicting or just plain wrong text, where the Web maintainer
> really ought to correct the information.
> .
And that's why the web is a source of dubious reliability. We pay a
price for easy access.