The Living Dead
Seanpit
RobinGoodfellow wrote:
< snip >
> As for the deleterious rate being biased downword in that regard, the
> conclusion is almost certainly valid. However, the calculation of U
> has a very large margin of error attached to it. If you use their
> lower-end estimate of U=1.5, bump it up by, say, 20 percent to adjust
> for the bias, and apply their calculation, you get 12 offspring per 2
> parents per generation rather than 40. Still on the high side, but not
> insurmountable, especially if you actually start taking epistasis into
> account. The point being, N&C estimate of U=3 are not the final word.
N&C did not think that U was less than 3 due to the fact that they only
considered a relatively small portion of the genome as being under the
influence of natural selection. In fact, the odds are most likely a
whole lot higher than U=3 now that a lot more of what was thought to be
"junk" DNA, because it doesn't code for proteins and whatnot, is no
longer junk, but functional and constrained by natural selection after
all. This has lead to the suggestion that U is actually "greater than
5" (see references below). Your notion of U is in reality as low as 1.5
is highly unlikely.
References with relevant excerpts:
"Using conservative calculations of the proportion of the genome
subject to purifying selection, we estimate that the genomic
deleterious mutation rate (U) is *at least 3* . . . In fact, this range
is likely to be biased downward because we have considered only
nonsynonymous sites as potential targets for deleterious mutations."
[emphasis added]
http://www.genetics.org/cgi/content/full/156/1/297
"Mukai, 1979; Kondroshov, 1988; Crow,1993; suggest that Ud > 5 is
feasible."
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
"The genomic deleterious mutation rate is likely much larger given
our estimate that 80% of amino acid mutations are deleterious and given
that it does not include deleterious mutations in noncoding regions,
which may be quite common."
http://www.genetics.org/cgi/content/full/156/1/297
< snip >
> But why not read the Kimura and Maruyama
> article exploring the effects of epistosis, and tell me if you think
> the situation is so simple?
I have read several of Kimura's papers as well as several papers from
others dealing with this problem by suggesting some form of epistasis
(multiplicative increases in the effects of detrimental mutations). The
problem is that with detrimental mutation rates as high as 3 per person
per generation, positive epistasis only increases the death rate, but
does not clear detrimental mutations from the gene pool faster than the
fitness of the most fit individual in the population decreases. Of
course, sexual reproduction is supposed to help out with this problem,
but sexual reproduction doesn't even help unless mating is done in a
non-random way.
"Under asexual reproduction, however, the epistasis has no effect in
reducing the load." (1). What you need, then, is some sort of
non-random mating so that those individuals in the population with the
most fit genomes mate, preferentially, with others in the genome with
the most fit genomes (1). That's the only way to theoretically avoid
genetic deterioration. "Nonrandom mating can cause the standardized
selection gradient and the opportunity for selection to change in a
noncompensating manner and thereby cause the requisite load to deviate
from eUd -1" (1).
Of course, there are several problems with this non-random mating
notion. With a detrimental mutation rate of 3 per individual per
generation, only 2 of 40 individuals will have a neutral fitness
balance relative to the original parent population. What are the odds
that these 2 individuals will actually mate "preferentially" so that
the overall number of individuals with equivalent parent-level fitness
does not decrease in each generation? Sure, it may happen
occasionally, but, on average, there will be less and less individuals
with parent level fitness in each generation until the most fit
individual is less fit than the original ancestors.
The only real way to increase the odds of this fortuitous mating is by
increasing the death rate of the other members of the population with
detrimental mutations before they have a chance to reproduce with the
remaining non-mutant individuals. But, if all the mutant individuals
die off before mating, to increase the odds of mating between
non-mutant individuals, the non-mutant individuals will have to
increase their reproductive rates dramatically to keep the population
of non-mutant individuals constant in each generation. Of course, this
creates an end-result of a need for a very large reproductive rate
(i.e., 40+ offspring per female) - and we are back to square one.
1.http://mbe.oxfordjournals.org/cgi/reprint/1/1/84?ijkey=d0f8b3b833737a5cb5dcbaca416ab8cb5fb0c1eb
> > > There are multiple
> > > scenarios as to what can happen once such effects accumulate up to a
> > > certain threshold, mass extinction being one of them, but certainly not
> > > the only one.
> >
> > What are these other options?
>
> a) The fixation of some of the slightly deleterious alleles in the
> population. While the overall phenotypic fitness will go down, the
> relative frequency of the "most-fit" phenotype (i.e., the one with the
> fewest deleterious mutations) would go up, and the overall rate of
> deleterious mutations would decrease. The population will reach
> equilibrium, albeit at a slightly less fit level than it used to be.
> Life will go on.
In every generation every individual receives around 3 deleterious
mutations. The "most fit" phenotype in each generation comes in at a
ratio of about 1 in 20. With that ratio, unless very fortuitous mating
and high reproductive rates take place, the direction is still
extinction. There is no equilibrium here. The only way to avoid this
problem is by having the most fit individuals mate pretty much
exclusively with each other and still make a whole lot of children in
each generation.
> b) Even as it accumulates slightly deleterious mutations, a population
> may expand to a certain size where deleterious mutations fail to
> propagate as quickly,
Deleterious mutations happen on an individual basis - more than 3 per
individual per generation. These are then passed on to the next
generation from parents to children and do not leave the gene pool
until they die out of it with the premature death of an individual who
didn't get a chance to replicate. They don't need to "propagate" any
other way to be a problem. Their rate of propagation is not
significantly related to the size of the population, but to the
reproductive rate and death rate.
> whereas beneficial mutaitons will arise and
> spread much more rapidly (this is one of the basic theorems of
> population genetics).
Not quite true.
"In sexual populations, the combined effect of beneficial and
deleterious mutations is to favor a decreased rate of mutation and that
the indirect selection resulting from beneficial mutations is small or
negligible compared to that resulting from deleterious mutations. . .
Relative to an asexual population, increased levels of recombination
reduce the effects of beneficial mutations more rapidly than those of
deleterious mutations" (1). "The probability of fixation of a given
beneficial mutation decreases with both population size and mutation
rate" (2). "Moreover, deleterious mutations reduce the chance of
fixation of advantageous mutations, since they increase the occurrence
of such an event in a genome that already has a large number of
segregated deleterious mutations. In this way, to be successful,
mutations that have larger selective benefit must be produced (de
Oliveira and Campos 2004)" (3). "Rare reverse and compensatory
mutations can move deleterious mutations, via genetic hitchhiking,
against the flow of genetic polarization. But this is a minor
influence, analogous to water turbulence that occasionally transports a
pebble a short distance upstream. . . Whenever demographic, ecological,
and/or physiological constraints cause, R-best t to be less than eUd,
then the progenitor class will decline in size each generation and
deterministic mutation accumulation will ensue. Such mutation
accumulation will be opposed by reverse and compensatory mutations, but
if R-best is much less than eUd, then net mutation accumulation will
ensue" (4).
Especially note that the rate of fixation of beneficial mutations is
less in sexual than in asexual populations. This problem is made even
worse by the high rate of detrimental mutations sustained by a slowly
reproducing population.
1. http://www.genetics.org/cgi/content/full/151/4/1621
2. http://www.iubs.org/test/bioint/44/9.htm
4.
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
> Thus, over time, overall fitness of the
> population may increase as it grows and mutates. Small negative and
> positive epistasis effects can play an important role here, in both
> helping to eliminate deleterious mutations and in propagating
> beneficial ones.
Epistasis does not change the need for a certain death rate and
reproductive rate. Epistasis may preferentially remove mutations from
those with the most detrimental mutations in a population, but it
doesn't remove them faster than they are formed. That's the problem.
The only way to get rid of negative mutations faster than they are
introduced, without increasing the death rate, is to somehow
concentrate them in a few individuals at a higher rate than the
detrimental mutations are formed. Epistasis alone does not do this. In
fact, there is no way I can think of as to how this concentration might
occur - outside of ID that is.
> c) A population may reach a limit where a certain proportion of
> deleterious mutations will no longer be tolerated. This will cause a
> massive die-off of the organisms possessing such genotypes, thereby
> boosting the proportion of most-fit individuals in the population.
Yes, and these most fit individuals might be just above the threshold
of die-off themselves. In order to maintain their level of
just-barely-making-it fitness, their reproductive rate would have to
skyrocket.
You see, it is not the proportion of most-fit individuals that you need
to think about, it is the proportion of the fittest individuals of the
current generation relative to the fittest individuals of the previous
generation that's important here. All of the other individuals,
besides this special group, are referred to as the non-self-sustaining
"living dead" (ref).
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
> The
> population will once again slowly degrade, until it goes through the
> same die-off process. Lather, rinse, repeat ad nausem. Not a pleasant
> thought, but not extinction, either.
It will not "rise" to the level of maximum fitness of the previous
generation - that's the problem. The maximum fitness level of the
population (the level of the most fit individual in the entire
population) will always decrease when the reproductive rates are lower
than the rate of build up of deleterious mutations.
> > > 5) The accumulation of positive selective effects in the population.
> >
> > These do not accumulate nearly as fast as the negative mutations - Even
> > in literature the suggested ratio is less than 1 beneficial to 1000
> > negative.
>
> Do you mean "classic English literature"? Because, in the scientific
> literature, the ratios are considerably higher. Looking at the article
> by Bustamante et. al., they have found evidence that, of the genetic
> loci they examined, 9.0% show signs of positive selection, whereas
> 13.5% show signs of negative selection.
That is not the "rate" of beneficial vs. detrimental mutations. The
rate of detrimental mutations far outpaces the rate of beneficial
mutations.
For E. coli, the estimated value for the beneficial mutation rate
(Miralles et. al., 1999) was 6.4 × 1e-8 beneficial mutations per
genome per generation. (1) The beneficial mutation rate obtained by
Imhof and Schlötterer was 4 x 1e-9 per genome per generation (see
reference links below). (2) Compare this with the detrimental mutation
rate for E. coli "in excess of 0.0002" per genome per generation. (3)
That produces a ratio of between 1 in ~3,000 to 1 in ~50,000. But what
about eukaryotes? "In sexual populations of higher eukaryotes, there is
extensive data showing that U >> K." (4) "In general, organisms with
larger genomes appear to have a greater number of deleterious
mutations, although it does not appear that the deleterious mutation
rate is constant per base pair across these organisms." (5)
1.http://www.nap.edu/openbook/0309070996/html/73.html
2. Marianne Imhof and Christian Schlötterer, Fitness effects of
advantageous mutations in evolving Escherichia coli populations, Proc
Natl Acad Sci USA. 2001 January 30; 98(3): 1113-1117 -
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=14717
3.http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=retrieve&db=pubmed&list_uids=8649513&dopt=Abstract
4.http://www.genetics.org/cgi/content/full/151/4/1621
5.http://www.genetics.org/cgi/content/full/156/1/297
> Now, even if we were to accept
> Kimura's estimate that 86% of all functional mutations in primates are
> negative, we would still arrive at the conclusion that 0.09 * (1-0.86)
> = 1.3% percent of all functional mutations are beneficial, whereas .86
> + .135*(1-0.86)= 87.9% of all functional mutations are deleterious.
> That gives a ratio of 15 beneficial to 1000 negaitve, which is 15 times
> what you state. Of course, this ratio would be even bigger using N&C
> estimates of the fraction of deleterious mutations (about 30 to 1000).
Perhaps you've misread Kimura? It seems that the "86%" number is not
the number of all "functional" mutations, but of nonsynonymous
substitutions in a functional protein - like the hemoglobin protein
used by Kimura in this particular study. Kimura concludes that the
other 14% are *neutral* mutations. Evidently, Kimura concludes that
many non-synonymous mutations, mutations that actually change a residue
sequence, are functionally neutral.
Kimura writes:
"I can compute the fraction of *neutral mutations* with respect to
electrophoretically detectable changes in hemoglobin by the ratio (0.71
x 0.28 + 0.62/3)/(4.6 x 0.28 + 4.6/3), which gives P,,,,(Hb) = 0.14."
[emphasis added]
http://mbe.oxfordjournals.org/cgi/reprint/1/1/84?ijkey=d0f8b3b833737a5cb5dcbaca416ab8cb5fb0c1eb
N&C write:
"What proportion of nonsynonymous changes are neutral and what
proportion are deleterious? The fraction that are neutral, fo, can be
calculated by comparing the total mutation rate, µt, with the
substitution rate, vo = foµt (KIMURA 1983A, KIMURA 1983B). The
proportion that are deleterious is 1 - fo. Using this approach, KIMURA
1983B estimated that 86% of nonsynonymous substitutions are
deleterious. . . The genomic deleterious mutation rate is likely much
larger given our estimate that 80% of amino acid mutations are
deleterious and given that it does not include deleterious mutations in
noncoding regions, which may be quite common."
http://www.genetics.org/cgi/content/full/156/1/297
> > > This is not taken into account by N&C at all, but positive mutations do
> > > happen, and while more rare than their deleterious counterparts, their
> > > effects on the overall fitness of the population can be just as
> > > important, if not more so.
> >
> > They are far more rare. That's the problem.
>
> Yes, they are more rare, but they spread through populations and fix
> far more quickly than deleterious mutations. As the result, the
> balance of positive versus negative mutations can maintain or even
> increase the overall fitness of a population, despite their relative
> rarity.
Not quite true, as discussed above.
> > > Recently, Bustamante et. al. (Nature
> > > 437:1153-57) have shown that a considerable number of loci in the human
> > > genome appear to be undergoing positive selection, and simulations show
> > > that advantageous mutations tend to become fixed in a population both
> > > faster and in greater proportion than their deleterious counterparts,
> > > even under evolutionary scenarios where the overwhelming majority of
> > > non-neutral mutations happen to be deleterious. In other words, the
> > > small balance of positive mutations, due to their faster fixation
> > > times, can maintain or increase the overall fitness of a population,
> > > even as it keeps accumulating slightly deleterious mutations.
> >
> > There is no reason to suggest that a slight positive mutation is fixed
> > any faster in a population than a negative mutation of equal degree can
> > be eliminated from a population.
>
> Of course there is. The longer a slightly positive mutation stays in
> the population, the more likely it is to spread. Unless such a
> mutation is lost quickly, it will almost certainly achieve fixation.
Not exactly true, as discussed above.
> On the other hand, slightly deleterious mutations are still subject to
> negative selection in every generation, and the chances of their
> continued presence in the gene pool are not improved with each passing
> generations, at least in steady-state populations.
But the only way to get rid of these negative mutations faster than
they are being formed is by preferentially concentrating them in some
members of a population, or by increasing the reproductive rate
dramatically. Epistasis doesn't solve this problem - as discussed
above.
> So, while they may
> enter the gene pool more quickly, they will also be eliminated more
> quickly, especially if they accumulate and the deleterious effect is
> amplified.
They will not be eliminated more quickly than they are produced in a
slowly reproducing population - that's the problem.
< snip >
Cheers,
Leonid.
Sean Pitman
www.DetectingDesign.com
John Harshman
> In fact, the odds are most likely a
> whole lot higher than U=3 now that a lot more of what was thought to be
> "junk" DNA, because it doesn't code for proteins and whatnot, is no
> longer junk, but functional and constrained by natural selection after
> all.
I could use some evidence for the assertion that the proportion of
sequences under selection is much higher than previously thought. Your
references don't seem to provide any.
By the way, are you ever going to respond to any of the common descent
or flood geology questions?
Richard Forrest
I predict that the answer to that question is a resounding "NO".
RF
Ron O
Does Loma Linda have a psychiatric division? Sean really does need
some help. Someone that he trusts ought to get him to seek it out. He
really went over the deep end trying to defend the ID scam. I wonder
if he still thinks that he could have done better than the IDiots in
Ohio after the Dover fiasco?
He can blather on about U=3 and can't get a grip on why his own junk
doesn't measure up. Wasn't there someone else that went on about
selection pressure and a populations ability to remove detrimentals?
ReMine used to go on about something like it, but I recall someone else
tried just what Sean is trying now. It may have been over at ARN or
something. What anyone that wants to use this argument has to do is
explain the greater than 3:1 silent to replacement amino acid ratio
found in genes between closely related species. I've seen it as high
as 5:1. Sean or anyone else has to figure out where all the
replacement substitutions went if selection can't remove the
detrimentals as he claims. We are talking about differences between
species like horses and donkeys that Sean would probably acknowledge
share a common ancestor. The ratio should be closer to 1:2 if mutation
was arbitrary, as we observe it to be, but a boat load of replacement
substitutions have been removed from these populations and they didn't
go extinct. How is this possible if Sean is correct. Sean doesn't
seem to understand what epistasis is in population genetics. The short
answer is that sets of genes can be selected against at the same time
due to their interactions. You don't have to select against one at a
time.
The same thing applies to his bogus protein sequence probability
estimates when he can't explain how antibodies work in less than 10^12
trials. How could functional sequences be as rare has he demands if
the fraction of sequence space that has to be searched is so close to
zero?
In science we understand that we don't know everything. Scientists
accept that limitation, but guys like Sean can't seem to come to grips
with reality.
I'd like to see his junk that he considers to be better than the
scientific evidence for common descent that he claims isn't good
enough. He can come up with junk about U=3 that he probably doesn't
understand, but he can't demonstrate that he is able to rationally
weigh evidence for or against his own beliefs. Pretty sad, and Sean is
about the best of the creationist lot even if he is mentally
incompetent. Just think how easy it would be for someone like Sean to
destroy his own creationist claptrap if he ever bothered to try. It
should be a lot easier to do if you can use bogus arguments that you
can't verify to do it.
Ron Okimoto
Seanpit
to deterimental mutation has indeed increased now that we know of other
functional regions of the genome besides protein coding regions.
Sean Pitman
www.DetectingDesign.com
Seanpit
Ron O wrote:
> What anyone that wants to use this argument has to do is
> explain the greater than 3:1 silent to replacement amino acid ratio
> found in genes between closely related species. I've seen it as high
> as 5:1. Sean or anyone else has to figure out where all the
> replacement substitutions went if selection can't remove the
> detrimentals as he claims. We are talking about differences between
> species like horses and donkeys that Sean would probably acknowledge
> share a common ancestor. The ratio should be closer to 1:2 if mutation
> was arbitrary, as we observe it to be, but a boat load of replacement
> substitutions have been removed from these populations and they didn't
> go extinct.
Section can remove detrimental mutations, just not fast enough to keep
up with the steady decline in the maximum fitness of a population that
is reproducing slowly. Species like donkeys and horses and mules and
apes and us humans have not gone extinct - yet. But, we are all still
headed for extinction and always have been since original creation.
> How is this possible if Sean is correct.
They haven't been around very long - only a few thousand years.
> Sean doesn't
> seem to understand what epistasis is in population genetics. The short
> answer is that sets of genes can be selected against at the same time
> due to their interactions. You don't have to select against one at a
> time.
That's exactly right. But, this doesn't help get remove the
detrimental mutations from the gene pool as fast as they enter it.
Certainly many mutations can be removed at the same time, but many more
enter the pool than are removed by epistasis. The only way to keep up
is by increasing the reproductive rate.
For example, say that you have a steady state population of 1 billion
with a generation time of 20 years and a detrimental mutation rate of 3
per individual per generation. In the first generation each female
gives birth to 10 children. The next generation now has a population
of 5 billion. Of these, how many have no detrimental mutations? - Only
1 in 20 or 250,000,000. These survive, along with 750,000,000 of their
less fit pears. In the next generation each woman again gives birth to
an average of 10 children. How many, on average, will have a neutral
mutational balance with respect to the original population? - Only
62,500,000 out of 1 billion.
Do you see what is happening here? The number of those that are at
least as fit as the original population decreases even though the
population itself my not decrease. This steady decrease in maximal
fitness is not significantly helped by positive epistasis of any kind.
Many in the population may have multiple mutations built up and
epistasis will remove these individuals, but it will not make up for
the fact that the numbers of most fit individuals is steadily
decreasing.
Oh, but what if you preferentially mate those with the highest-level
fitness only with those with the highest-level fitness? Well, it might
slow things down a bit, but without an increase in the reproductive
rate above 40 per woman, the decline will continue. Also, how are you
going to guarantee that a woman with the fewest detrimental mutations
only has offspring with a man with the fewest detrimental mutations?
> The same thing applies to his bogus protein sequence probability
> estimates when he can't explain how antibodies work in less than 10^12
> trials.
They only work for functions, like enzymes, at very low levels of
complexity - functions that require no more than a few hundred fairly
specified residues at minimum. Note that "fairly specified" is equal to
~1 in 1e30 per 100aa.
> How could functional sequences be as rare has he demands if
> the fraction of sequence space that has to be searched is so close to
> zero?
The density of sequence space at these low levels is relatively high.
This is just not so easy to do for functions that require a minimum of
several thousand fairly specified residues.
< snip all the other personal remarks >
> Ron Okimoto
Sean Pitman
www.DetectingDesign.com
Ron O
Seanpit wrote:
> Ron O wrote:
>
> > What anyone that wants to use this argument has to do is
> > explain the greater than 3:1 silent to replacement amino acid ratio
> > found in genes between closely related species. I've seen it as high
> > as 5:1. Sean or anyone else has to figure out where all the
> > replacement substitutions went if selection can't remove the
> > detrimentals as he claims. We are talking about differences between
> > species like horses and donkeys that Sean would probably acknowledge
> > share a common ancestor. The ratio should be closer to 1:2 if mutation
> > was arbitrary, as we observe it to be, but a boat load of replacement
> > substitutions have been removed from these populations and they didn't
> > go extinct.
>
> Section can remove detrimental mutations, just not fast enough to keep
> up with the steady decline in the maximum fitness of a population that
> is reproducing slowly. Species like donkeys and horses and mules and
> apes and us humans have not gone extinct - yet. But, we are all still
> headed for extinction and always have been since original creation.
>
> > How is this possible if Sean is correct.
>
> They haven't been around very long - only a few thousand years.
Unfortunately this is bogus. If they have been around for only a few
thousand years how do you account for all the missing mutations that
have been selected out of the population already? Remember you have to
account for the bettern than 3:1 ratio. If that many mutations can be
selected out since they got off the ark how can you say that it is not
enough?
This is just another one of your assertions that you can't back up.
Just because you say so doesn't make it so.
>
> > Sean doesn't
> > seem to understand what epistasis is in population genetics. The short
> > answer is that sets of genes can be selected against at the same time
> > due to their interactions. You don't have to select against one at a
> > time.
>
> That's exactly right. But, this doesn't help get remove the
> detrimental mutations from the gene pool as fast as they enter it.
> Certainly many mutations can be removed at the same time, but many more
> enter the pool than are removed by epistasis. The only way to keep up
> is by increasing the reproductive rate.
How do you know that? All the data indicates that species have been
around for a lot longer than you claim and their DNA says that they
have gotten rid of a boat load of detrimental mutations. This is just
another of your bogus assertions that you can't back up. Prove that
the reproductive rate hasn't been high enough when the masses of data
tell us that you are wrong. Just think of all the data that you have
to ignore to make your assertion make sense. The age of the earth and
the fact that life has been around for over 3 billion years on this
planet is just a couple of the data points that you have to deal with,
but you can't.
What credible person thinks that existing species have only been around
for a few thousand years. Just the tree ring data tells you that you
are off on that one too. When did the last ice age end? It was only
the last one. When did the previous ones end?
>
> For example, say that you have a steady state population of 1 billion
> with a generation time of 20 years and a detrimental mutation rate of 3
> per individual per generation. In the first generation each female
> gives birth to 10 children. The next generation now has a population
> of 5 billion. Of these, how many have no detrimental mutations? - Only
> 1 in 20 or 250,000,000. These survive, along with 750,000,000 of their
> less fit pears. In the next generation each woman again gives birth to
> an average of 10 children. How many, on average, will have a neutral
> mutational balance with respect to the original population? - Only
> 62,500,000 out of 1 billion.
Sean, humans have a relatively low genetic load. It is estiimated to
be around 2.5. This just means that you have the equivalent of 5
recessive lethals in your genome. You are still alive and you still
reproduce. This isn't just 5 mutations. Most of the detrimental
mutations are probably less than 10% lethal. That just means that it
would take around 10 combining together to kill you, and it would take
all 10 of them out with one death.
Some species have much higher genetic loads and they still reproduce
and do just fine. There is a wood rat that is an obligate outcrosser
(it doesn't inbreed) and it has the highest genetic load that I've
heard of at around 14. As long as the recessive detrimentals do not
become homozygous you can accumulate quite a few in a population.
Humans have a low genetic load because they think that we practiced
inbreeding in small bands in the relatively recent past along with the
genetic bottle neck that we went through that seems to have missed most
other species.
>
> Do you see what is happening here? The number of those that are at
> least as fit as the original population decreases even though the
> population itself my not decrease. This steady decrease in maximal
> fitness is not significantly helped by positive epistasis of any kind.
> Many in the population may have multiple mutations built up and
> epistasis will remove these individuals, but it will not make up for
> the fact that the numbers of most fit individuals is steadily
> decreasing.
You might as well have pulled those numbers out of your butt for all
the good they do you.
>
> Oh, but what if you preferentially mate those with the highest-level
> fitness only with those with the highest-level fitness? Well, it might
> slow things down a bit, but without an increase in the reproductive
> rate above 40 per woman, the decline will continue. Also, how are you
> going to guarantee that a woman with the fewest detrimental mutations
> only has offspring with a man with the fewest detrimental mutations?
Actually no, selection works best if you actually inbreed so that
recessive detrimentals can become homozygous and be selected against.
You can't identify the most fit to breed together when you are dealing
with the majority of detrimental mutations.
>
> > The same thing applies to his bogus protein sequence probability
> > estimates when he can't explain how antibodies work in less than 10^12
> > trials.
>
> They only work for functions, like enzymes, at very low levels of
> complexity - functions that require no more than a few hundred fairly
> specified residues at minimum. Note that "fairly specified" is equal to
> ~1 in 1e30 per 100aa.
This still doesn't cut it 10^12 or in your notation 10e12 is a maximum
number for antibody production. The real number is likely lower.
Above that you reach a limit of the number of cells in the human body.
You still come up way too low in your estimate for functional
sequences. 10e18 is not a number to just shrug off. That you try and
shrug it off is just a reflection on your boneheaded willful ignorance
on this subject. Somehow you have to get your estimate to reflect
reality. 18 orders of magnitude off isn't a number to sneaze at. Your
above bogus arguement about detrimental mutations depends on estimates
that could be two fold off, that isn't even 1 order of magnitude. Why
do you think that you are onto something when you are still so far out
of the ball park on this bogus estimate?
>
> > How could functional sequences be as rare has he demands if
> > the fraction of sequence space that has to be searched is so close to
> > zero?
>
> The density of sequence space at these low levels is relatively high.
> This is just not so easy to do for functions that require a minimum of
> several thousand fairly specified residues.
Another assertion that you can't back up. What is low? Selection only
cares about if it is high enough of an effect to notice.
>
> < snip all the other personal remarks >
>
> > Ron Okimoto
>
> Sean Pitman
> www.DetectingDesign.com
Instead of this bogus junk, put up your evidence for your beliefs and
compare it to what you don't think is good enough. That is what
science does every day. Why do you think that you refuse to do
something so simple and direct? Get help. Find someone you trust and
get them to help you. You know that I used to think that you were just
dishonest, but I don't make that claim anymore. Do you still think
that ID is good enough science to teach in the public schools? Have
you, at least, come to grips with that reality, or are you still
deluded enough to think that you have the answers that no one else
seems to have? Have you compared your evidence to the evidence for
common descent that you claim isn't good enough? Just imagine that you
were reading about someone else that refused to do some simple things
in order to keep believing some bogus notions. Why do you have to do
it?
Ron Okimoto
Seanpit
Ron O wrote:
< snip >
> > > Sean doesn't
> > > seem to understand what epistasis is in population genetics. The short
> > > answer is that sets of genes can be selected against at the same time
> > > due to their interactions. You don't have to select against one at a
> > > time.
> >
> > That's exactly right. But, this doesn't help get remove the
> > detrimental mutations from the gene pool as fast as they enter it.
> > Certainly many mutations can be removed at the same time, but many more
> > enter the pool than are removed by epistasis. The only way to keep up
> > is by increasing the reproductive rate.
>
> How do you know that?
Did you read any of the references I provided?
> All the data indicates that species have been
> around for a lot longer than you claim and their DNA says that they
> have gotten rid of a boat load of detrimental mutations.
But you have no logical explanation for how to get rid of a boatload of
detrimental mutations faster than the many boatloads that are added in
every generation.
> This is just
> another of your bogus assertions that you can't back up. Prove that
> the reproductive rate hasn't been high enough when the masses of data
> tell us that you are wrong.
Tell me, where do you say any woman having over 40 pregnancies? And,
that's just with U equal to 3. Many suggest that U is actually over 5.
Do you think it possible for the average human woman to have over a
hundred pregnancies?
> Just think of all the data that you have
> to ignore to make your assertion make sense. The age of the earth and
> the fact that life has been around for over 3 billion years on this
> planet is just a couple of the data points that you have to deal with,
> but you can't.
Think of all the data, like this data, that you have to ignore to
believe that life has been around on this planet for billions of years
and that slowly reproducing life has been around for many millions of
years.
> What credible person thinks that existing species have only been around
> for a few thousand years. Just the tree ring data tells you that you
> are off on that one too. When did the last ice age end? It was only
> the last one. When did the previous ones end?
I've written essays about both ice core dating as well as radiocarbon
dating - listed on my website. They don't help you out of your
problems. You may not think these arguments of mine "credible", but
hey, you can believe whatever crazy theories you want. Just don't
expect me to go along with this ToE nonsense.
> > For example, say that you have a steady state population of 1 billion
> > with a generation time of 20 years and a detrimental mutation rate of 3
> > per individual per generation. In the first generation each female
> > gives birth to 10 children. The next generation now has a population
> > of 5 billion. Of these, how many have no detrimental mutations? - Only
> > 1 in 20 or 250,000,000. These survive, along with 750,000,000 of their
> > less fit pears. In the next generation each woman again gives birth to
> > an average of 10 children. How many, on average, will have a neutral
> > mutational balance with respect to the original population? - Only
> > 62,500,000 out of 1 billion.
>
> Sean, humans have a relatively low genetic load. It is estiimated to
> be around 2.5. This just means that you have the equivalent of 5
> recessive lethals in your genome. You are still alive and you still
> reproduce. This isn't just 5 mutations. Most of the detrimental
> mutations are probably less than 10% lethal. That just means that it
> would take around 10 combining together to kill you, and it would take
> all 10 of them out with one death.
I'm not talking about just recessive lethal mutations. I'm talking
about all types of detrimental mutations. They don't have to be
recessively "lethal" to count as detrimental mutations. These
detrimental mutations do in fact build up much faster than they can be
eliminated by epistasis of any kind. Only a dramatic increase in the
reproductive rate can really solve the problem of the "Living Dead".
The living dead can reproduce just fine. They can even reproduce a
whole lot. However, they are not as fit as their parents were. They
do indeed have more detrimental mutations and fewer individuals with
equivalently fit genomes as the previous generation had.
> Some species have much higher genetic loads and they still reproduce
> and do just fine.
Sure they do. That's not the question. The question is if there are as
many members of a particular generation that are as least as fit as the
most fit members of the previous generation. The answer to this
question for humans and other slowly reproducing creatures, is no.
>There is a wood rat that is an obligate outcrosser
> (it doesn't inbreed) and it has the highest genetic load that I've
> heard of at around 14. As long as the recessive detrimentals do not
> become homozygous you can accumulate quite a few in a population.
> Humans have a low genetic load because they think that we practiced
> inbreeding in small bands in the relatively recent past along with the
> genetic bottle neck that we went through that seems to have missed most
> other species.
This is not the question or the problem I'm talking about Ron. You
obviously need to read up more on this topic.
> > Do you see what is happening here? The number of those that are at
> > least as fit as the original population decreases even though the
> > population itself my not decrease. This steady decrease in maximal
> > fitness is not significantly helped by positive epistasis of any kind.
> > Many in the population may have multiple mutations built up and
> > epistasis will remove these individuals, but it will not make up for
> > the fact that the numbers of most fit individuals is steadily
> > decreasing.
>
> You might as well have pulled those numbers out of your butt for all
> the good they do you.
These aren't my numbers. These are numbers have been published. This
idea comes trait from papers taking about this problem, such as this
one who calls most in each generation "The Living Dead":
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
> > Oh, but what if you preferentially mate those with the highest-level
> > fitness only with those with the highest-level fitness? Well, it might
> > slow things down a bit, but without an increase in the reproductive
> > rate above 40 per woman, the decline will continue. Also, how are you
> > going to guarantee that a woman with the fewest detrimental mutations
> > only has offspring with a man with the fewest detrimental mutations?
>
> Actually no, selection works best if you actually inbreed so that
> recessive detrimentals can become homozygous and be selected against.
> You can't identify the most fit to breed together when you are dealing
> with the majority of detrimental mutations.
You're wrong as far as the problem at hand is concerned. Selective
breeding of only the fittest with the fittest is one way to slow down
the decline. Random breeding or inbreeding, only speeds up the overall
decline of a population and hastens either extinction or a dramatic
increase in reproductive rates to overcome the problem.
> > > The same thing applies to his bogus protein sequence probability
> > > estimates when he can't explain how antibodies work in less than 10^12
> > > trials.
> >
> > They only work for functions, like enzymes, at very low levels of
> > complexity - functions that require no more than a few hundred fairly
> > specified residues at minimum. Note that "fairly specified" is equal to
> > ~1 in 1e30 per 100aa.
>
> This still doesn't cut it 10^12 or in your notation 10e12 is a maximum
> number for antibody production. The real number is likely lower.
> Above that you reach a limit of the number of cells in the human body.
> You still come up way too low in your estimate for functional
> sequences. 10e18 is not a number to just shrug off.
I don't shrug such numbers off. There certainly are far less specified
functions in low level sequence space. Some functions may have ratios
as low as 1 in 5 or 1 in 2 or even 1 in 1. However, these are not what
I would call "fairly specified" functions. And, there certainly are
functions that do indeed show ratios lower than 1 in 1e30 per 100aa.
That's the whole point. This level of minimum functional complexity is
much higher than the levels you are talking about - which are even
easier to evolve.
> That you try and
> shrug it off is just a reflection on your boneheaded willful ignorance
> on this subject. Somehow you have to get your estimate to reflect
> reality. 18 orders of magnitude off isn't a number to sneaze at. Your
> above bogus arguement about detrimental mutations depends on estimates
> that could be two fold off, that isn't even 1 order of magnitude. Why
> do you think that you are onto something when you are still so far out
> of the ball park on this bogus estimate?
I'm not out of the ballpark at all. There are functions that do indeed
require such higher levels of specificity as well as sequence sizes
that go into the thousands of residues - all of which are specified to
at least 1 in 1e30 per 100aa.
> > > How could functional sequences be as rare has he demands if
> > > the fraction of sequence space that has to be searched is so close to
> > > zero?
> >
> > The density of sequence space at these low levels is relatively high.
> > This is just not so easy to do for functions that require a minimum of
> > several thousand fairly specified residues.
>
> Another assertion that you can't back up. What is low? Selection only
> cares about if it is high enough of an effect to notice.
Very low level functions include those that you've just listed,
functions that have short minimum size requirements and little
specificity - specificities much greater than 1 in 1e30 per 100aa.
These types of functions are relatively easy to evolve.
< snip more irrelevant personal comments >
Ron O
Seanpit wrote:
> Ron O wrote:
>
> < snip >
>
> > > > Sean doesn't
> > > > seem to understand what epistasis is in population genetics. The short
> > > > answer is that sets of genes can be selected against at the same time
> > > > due to their interactions. You don't have to select against one at a
> > > > time.
> > >
> > > That's exactly right. But, this doesn't help get remove the
> > > detrimental mutations from the gene pool as fast as they enter it.
> > > Certainly many mutations can be removed at the same time, but many more
> > > enter the pool than are removed by epistasis. The only way to keep up
> > > is by increasing the reproductive rate.
> >
> > How do you know that?
>
> Did you read any of the references I provided?
How do you know that? Answer the question. It isn't in the papers
that you cite, because I know for a fact that Crow admits that he
doesn't understand the mechanism by which these mutations are removed.
We don't know everything, and Crow is too good of a scientists to claim
otherwise.
Have you ever been honest in the evaluation of your own beliefs?
Demonstrate it. Put up the evidence you claim exists for your beliefs
that is better than the evidence that you claim isn't good enough for
common descent.
>
> > All the data indicates that species have been
> > around for a lot longer than you claim and their DNA says that they
> > have gotten rid of a boat load of detrimental mutations.
>
> But you have no logical explanation for how to get rid of a boatload of
> detrimental mutations faster than the many boatloads that are added in
> every generation.
What have you got? Really, we have data that tells us that these
mutations have been removed from the population. What have you got
that tells you differently? Why ignore what we know because we don't
know something else?
>
> > This is just
> > another of your bogus assertions that you can't back up. Prove that
> > the reproductive rate hasn't been high enough when the masses of data
> > tell us that you are wrong.
>
> Tell me, where do you say any woman having over 40 pregnancies? And,
> that's just with U equal to 3. Many suggest that U is actually over 5.
> Do you think it possible for the average human woman to have over a
> hundred pregnancies?
Demonstrate that this many pregnancies are required of human females.
There is that two fold estimate again. What about 18 orders of
magnitude that you can't deal with? How can you say that 40
pregnancies are needed when you admit that you don't know the mechanism
by which the detrimental mutations are removed? How have the
detrimentals been removed from the existing populations to get the 3:1
ratio in just the few thousand years you have to work with? You
require rates so much greater than real science that you should give up
on this line of argument. The 3:1 is an average for every gene.
Science claims that it has literally millions of years to produce that
ratio, you are around 3 orders of magnitude worse off to make it fit
your model. You have to make these arguments fit your model, you can't
just claim that it is a problem for science.
>
> > Just think of all the data that you have
> > to ignore to make your assertion make sense. The age of the earth and
> > the fact that life has been around for over 3 billion years on this
> > planet is just a couple of the data points that you have to deal with,
> > but you can't.
>
> Think of all the data, like this data, that you have to ignore to
> believe that life has been around on this planet for billions of years
> and that slowly reproducing life has been around for many millions of
> years.
We don't ignore it. You wouldn't be able to misuse the scientific
papers if it wasn't discussed and people weren't researching the
issues. What a blockhead.
>
> > What credible person thinks that existing species have only been around
> > for a few thousand years. Just the tree ring data tells you that you
> > are off on that one too. When did the last ice age end? It was only
> > the last one. When did the previous ones end?
>
> I've written essays about both ice core dating as well as radiocarbon
> dating - listed on my website. They don't help you out of your
> problems. You may not think these arguments of mine "credible", but
> hey, you can believe whatever crazy theories you want. Just don't
> expect me to go along with this ToE nonsense.
Who cares? You have to take all of the data into consideration. So
what if you are a few thousand years off for ice cores and radiocarbon
has to be calibrated using methods like dendro chronology. What about
Isochron dating that blows these estimates of less than a few hundred
thousand years (for radiocarbon less than 50,000) out of the water. It
isn't just the recent dating technology that you have to deal with.
Just reread what you just wrote, and think about time related questions
in science, and then think about flat earthism and how your world view
compares with it. Just pretend that it was some other incompetent
writing junk about ice cores as if it meant anything. You have some
inkling about what you have to ignore. It is crazy isn't it? If you
were ignorant that would be one thing, but you are literally claiming
that the earth is flat when you know that we have so much evidence
contrary to that position that it is ridiculous. We have literally
retrieved rocks from the moon and dated them. We can date meteorites,
and we can date the oldest existing rocks on earth, and you are trying
to claim that ice cores and radiocarbon dating mean anything except a
more accurate recent timeliine for events of the recent past.
They don't have to be, but most of them are. How many dominant lethals
are there in the population above mutation frequency (about zero, is
just a hint). You know that just about every site in the human genome
is hit by mutation in just the existing population. The only ones that
we couldn't find if we sequenced the genomes of everyone would be the
dominant lethals. There are defects that turn lethal later in life
like Huntingtons, but you have to admit that they are pretty rare.
If they are additive or not competely recessive it is easier to get
them out of the population, but it would take combinations of them to
make problems, and the combinations would be removed, wouldn't they?
They are not removed one at a time.
>
> > Some species have much higher genetic loads and they still reproduce
> > and do just fine.
>
> Sure they do. That's not the question. The question is if there are as
> many members of a particular generation that are as least as fit as the
> most fit members of the previous generation. The answer to this
> question for humans and other slowly reproducing creatures, is no.
This is bogus and you can't demonstrate it because the populations
haven't been tested that you need to test. I'm certain that Crow has
something in the discussion about the fact that the human population is
expanding. There are models where expanding populations can accumulate
more detrimental mutations. What you have to worry about is if our
population was shrinking over the last 1000 years. More individuals
are surviving. Humans can have fewer pregnancies and still increase
the population. What happens to selection pressure?
>
> >There is a wood rat that is an obligate outcrosser
> > (it doesn't inbreed) and it has the highest genetic load that I've
> > heard of at around 14. As long as the recessive detrimentals do not
> > become homozygous you can accumulate quite a few in a population.
> > Humans have a low genetic load because they think that we practiced
> > inbreeding in small bands in the relatively recent past along with the
> > genetic bottle neck that we went through that seems to have missed most
> > other species.
>
> This is not the question or the problem I'm talking about Ron. You
> obviously need to read up more on this topic.
You need to. Humans aren't even close to a limit for their genetic
load. How did these other species accumulate so many? How long did it
take them to do it? The wood rats probably have large litters and have
multiple generations each year. Remember humans like other species
have that 3:1 ratio so how did they get rid of all those replacement
substitutions? You know the prediction for the Wood rat is that even
with it's genetic load that it still has the 3:1 ratio when compared to
related species. This is data that you have to deal with. You just
don't want to because it means that you are probably wrong.
>
> > > Do you see what is happening here? The number of those that are at
> > > least as fit as the original population decreases even though the
> > > population itself my not decrease. This steady decrease in maximal
> > > fitness is not significantly helped by positive epistasis of any kind.
> > > Many in the population may have multiple mutations built up and
> > > epistasis will remove these individuals, but it will not make up for
> > > the fact that the numbers of most fit individuals is steadily
> > > decreasing.
> >
> > You might as well have pulled those numbers out of your butt for all
> > the good they do you.
>
> These aren't my numbers. These are numbers have been published. This
> idea comes trait from papers taking about this problem, such as this
> one who calls most in each generation "The Living Dead":
>
> http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
The vast majority of species that have ever existed are extinct. How
many species of great apes are there including humans? Where are all
the Australopithicines? We don't understand everything about
speciation, but we can study extant species, and what do we find? The
ones that made it have fewer replacement substitutions between their
closest relatives than the current genetic variation within their
population. Hey, this could be a way to figure out how many species
were on the Ark. Every species that has the same ratios as chimps and
humans would have to be created kinds because you can't figure out a
way to get rid of the replacement substitutions that are obviously
missing. The number of species on the Ark just went through the roof.
>
> > > Oh, but what if you preferentially mate those with the highest-level
> > > fitness only with those with the highest-level fitness? Well, it might
> > > slow things down a bit, but without an increase in the reproductive
> > > rate above 40 per woman, the decline will continue. Also, how are you
> > > going to guarantee that a woman with the fewest detrimental mutations
> > > only has offspring with a man with the fewest detrimental mutations?
> >
> > Actually no, selection works best if you actually inbreed so that
> > recessive detrimentals can become homozygous and be selected against.
> > You can't identify the most fit to breed together when you are dealing
> > with the majority of detrimental mutations.
>
> You're wrong as far as the problem at hand is concerned. Selective
> breeding of only the fittest with the fittest is one way to slow down
> the decline. Random breeding or inbreeding, only speeds up the overall
> decline of a population and hastens either extinction or a dramatic
> increase in reproductive rates to overcome the problem.
Where did you read this, cite the paper. I've never seen this
conclusion, ever.
Just think for a moment, you can't determine the fittest unless you
inbreed.
>
> > > > The same thing applies to his bogus protein sequence probability
> > > > estimates when he can't explain how antibodies work in less than 10^12
> > > > trials.
> > >
> > > They only work for functions, like enzymes, at very low levels of
> > > complexity - functions that require no more than a few hundred fairly
> > > specified residues at minimum. Note that "fairly specified" is equal to
> > > ~1 in 1e30 per 100aa.
> >
> > This still doesn't cut it 10^12 or in your notation 10e12 is a maximum
> > number for antibody production. The real number is likely lower.
> > Above that you reach a limit of the number of cells in the human body.
> > You still come up way too low in your estimate for functional
> > sequences. 10e18 is not a number to just shrug off.
>
> I don't shrug such numbers off. There certainly are far less specified
> functions in low level sequence space. Some functions may have ratios
> as low as 1 in 5 or 1 in 2 or even 1 in 1. However, these are not what
> I would call "fairly specified" functions. And, there certainly are
> functions that do indeed show ratios lower than 1 in 1e30 per 100aa.
> That's the whole point. This level of minimum functional complexity is
> much higher than the levels you are talking about - which are even
> easier to evolve.
What is unspecified about new enzymatic function? Demonstrate that the
functions that you are interested in are more "specified" than the
antibody functions. You are out of the ball park by 18 orders of
magnitude and you still claim that you can make viable conclusions?
How?
>
> > That you try and
> > shrug it off is just a reflection on your boneheaded willful ignorance
> > on this subject. Somehow you have to get your estimate to reflect
> > reality. 18 orders of magnitude off isn't a number to sneaze at. Your
> > above bogus arguement about detrimental mutations depends on estimates
> > that could be two fold off, that isn't even 1 order of magnitude. Why
> > do you think that you are onto something when you are still so far out
> > of the ball park on this bogus estimate?
>
> I'm not out of the ballpark at all. There are functions that do indeed
> require such higher levels of specificity as well as sequence sizes
> that go into the thousands of residues - all of which are specified to
> at least 1 in 1e30 per 100aa.
What are they? Demonstrate that they exist.
>
> > > > How could functional sequences be as rare has he demands if
> > > > the fraction of sequence space that has to be searched is so close to
> > > > zero?
> > >
> > > The density of sequence space at these low levels is relatively high.
> > > This is just not so easy to do for functions that require a minimum of
> > > several thousand fairly specified residues.
> >
> > Another assertion that you can't back up. What is low? Selection only
> > cares about if it is high enough of an effect to notice.
>
> Very low level functions include those that you've just listed,
> functions that have short minimum size requirements and little
> specificity - specificities much greater than 1 in 1e30 per 100aa.
> These types of functions are relatively easy to evolve.
You keep repeating this, but where are the counter examples to compare
with the antibody examples, what makes the antibody examples so much
less specified? Antibodies bind specific amino acid sequences with a
high degree of specificity. The minimum antigen is something like 7
amino acids. You can produce antibodies that will only bind a specific
7 in a specific order. How is this a lower level of specificity?
>
> < snip more irrelevant personal comments >
>
> > Ron Okimoto
>
> Sean Pitman
> www.DetectingDesign.com
You snipped the parts that you can't deal with. Just put up what you
claim to have. It isn't asking the moon. You are the one that claimed
that you had junk better than the evidence for common descent so put it
forward. If the evidence for common descent is so bad, what does that
say about your evidence? Snipping reality doesn't do you any good.
Get help. Get someone you trust to walk you through just how badly off
you are. You have to have some inkling that something is wrong.
Really, are you still deluding yourself that you have more on the ball
about ID than the scam artists that have gone belly up? You know how
badly you did in defense of those delusions. Do you still claim to
have something scientific to teach to kids?
Ron Okimoto
John Harshman
> Read the actual papers. The amount of the genome known to be subject
> to deterimental mutation has indeed increased now that we know of other
> functional regions of the genome besides protein coding regions.
I looked at the papers you cited. They don't seem to have anything on
the subject. We have known about functional regions that aren't
protein-coding for quite a long time, and they amount to only a percent
or two more than the protein-coding regions alone, unless you have some
new evidence that I don't know about. Your proposed criterion of
conservation over evolutionary time is a good one (though I don't see
how you can possiblly use it, since it must assume common descent). But
using that criterion, only a few percent of the genome would appear to
be functional (at the sequence level -- not counting spacer regions of
various sorts whose sequence is irrelevant).
So I ask you again. What references do you have that the amount of the
genome under selection is much greater than we think, i.e. a few percent?
And as I expected, you just ignored my questions about common descent
and flood geology.
Seanpit
> How do you know that? Answer the question. It isn't in the papers
> that you cite, because I know for a fact that Crow admits that he
> doesn't understand the mechanism by which these mutations are removed.
> We don't know everything, and Crow is too good of a scientists to claim
> otherwise.
Perhaps Crow doesn't know how the detrimental mutations are removed
because there isn't any good way to remove them? How have you helped
solve the problem that is in fact quite a mystery indeed?
< snip irrelevant personal comments>
> > > This is just
> > > another of your bogus assertions that you can't back up. Prove that
> > > the reproductive rate hasn't been high enough when the masses of data
> > > tell us that you are wrong.
> >
> > Tell me, where do you see any woman having over 40 pregnancies? And,
> > that's just with U equal to 3. Many suggest that U is actually over 5.
> > Do you think it possible for the average human woman to have over a
> > hundred pregnancies?
>
> Demonstrate that this many pregnancies are required of human females.
Read the papers Ron. This estimate is not based on my calculations.
Don't take my word for it. I've directly quoted several relevant
sections to these papers and you still don't understand. Look again at
the following quote:
"The reduction in fitness (i.e., the genetic load) due to
deleterious mutations with multiplicative effects is given by 1 - e^-U
(Kimura and Moruyama 1966). For U = 3, the average fitness is reduced
to 0.05, or put differently, each female would need to produce 40
offspring for 2 to survive and maintain the population at constant
size. This assumes that all mortality is due to selection and so the
actual number of offspring required to maintain a constant population
size is probably higher." - Nuchman, Michael W., Crowell, Susan L.,
Estimate of the Mutation Rate per Nucleotide in Humans, Genetics,
September 2000, 156: 297-304 (
http://www.genetics.org/cgi/content/full/156/1/297? )
>From this calculation, the number of births per female is 2 * (e^U).
When U = 3, the number of required births is 2 * (2.72^3) = ~ 40 births
And, this is felt to be a very low estimate. Consider the following
excerpt, which I've already referenced in my original post to this
thread:
"Accurate estimates of Ud for most metazoans with large genomes
(e.g., vertebrates) are not yet available but extrapolations from
studies of humans and Drosophila (Mukai, 1979; Kondroshov, 1988; Crow,
1993) suggest that Ud > 5 is feasible. In this case an asexual
population would have to have Rbest > e5 =148, a value far beyond the
physiological/ecological capabilities of most vertebrates."
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
For a *sexual* population, U=5 means that the average woman would have
to have over 296 pregnancies.
> There is that two fold estimate again.
It's not my estimate. Read the papers.
> What about 18 orders of
> magnitude that you can't deal with? How can you say that 40
> pregnancies are needed when you admit that you don't know the mechanism
> by which the detrimental mutations are removed?
I don't think they are removed. I think that we have always been headed
downhill since our very recent creation just a few thousand years ago.
> How have the
> detrimentals been removed from the existing populations to get the 3:1
> ratio in just the few thousand years you have to work with?
You're not dealing with all types of detrimental mutations in this
estimate of yours.
> You
> require rates so much greater than real science that you should give up
> on this line of argument. The 3:1 is an average for every gene.
> Science claims that it has literally millions of years to produce that
> ratio, you are around 3 orders of magnitude worse off to make it fit
> your model. You have to make these arguments fit your model, you can't
> just claim that it is a problem for science.
Not if you consider other types of detrimental mutations. I'm not just
talking about "recessive lethals" here - many of which are quite common
actually.
< snip >
> > > Some species have much higher genetic loads and they still reproduce
> > > and do just fine.
> >
> > Sure they do. That's not the question. The question is if there are as
> > many members of a particular generation that are as least as fit as the
> > most fit members of the previous generation. The answer to this
> > question for humans and other slowly reproducing creatures, is no.
>
> This is bogus and you can't demonstrate it because the populations
> haven't been tested that you need to test. I'm certain that Crow has
> something in the discussion about the fact that the human population is
> expanding. There are models where expanding populations can accumulate
> more detrimental mutations. What you have to worry about is if our
> population was shrinking over the last 1000 years. More individuals
> are surviving. Humans can have fewer pregnancies and still increase
> the population. What happens to selection pressure?
Yes, the population can be booming and still the numbers in the
population with at least the level of fitness of the previous
generation continues to shrink in each generation. That is what is
called the "Living Dead" problem. There can be quite a number of
living dead and this number can actually grow - i.e., the population of
"living dead" can increase for a time. However, the direction of the
overall population is still toward extinction.
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
< snip >
> > > Actually no, selection works best if you actually inbreed so that
> > > recessive detrimentals can become homozygous and be selected against.
> > > You can't identify the most fit to breed together when you are dealing
> > > with the majority of detrimental mutations.
> >
> > You're wrong as far as the problem at hand is concerned. Selective
> > breeding of only the fittest with the fittest is one way to slow down
> > the decline. Random breeding or inbreeding, only speeds up the overall
> > decline of a population and hastens either extinction or a dramatic
> > increase in reproductive rates to overcome the problem.
>
> Where did you read this, cite the paper. I've never seen this
> conclusion, ever. Just think for a moment, you can't
> determine the fittest unless you inbreed.
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
> > I don't shrug such numbers off. There certainly are far less specified
> > functions in low level sequence space. Some functions may have ratios
> > as low as 1 in 5 or 1 in 2 or even 1 in 1. However, these are not what
> > I would call "fairly specified" functions. And, there certainly are
> > functions that do indeed show ratios lower than 1 in 1e30 per 100aa.
> > That's the whole point. This level of minimum functional complexity is
> > much higher than the levels you are talking about - which are even
> > easier to evolve.
>
> What is unspecified about new enzymatic function? Demonstrate that the
> functions that you are interested in are more "specified" than the
> antibody functions. You are out of the ball park by 18 orders of
> magnitude and you still claim that you can make viable conclusions?
> How?
You need to read the papers by Sauer et al., which I've referenced
several times recently in this forum, which describe protein-based
functions with ratios as low as 1 in 1e62 per 100aa. Estimated ratios
for other types of protein-based functions, like cytochrome c are at
least as low as 1 in 1e30 per 100aa (most recent Yockey estimate).
John F. Reidhaar-Olson and Robert T. Sauer, "Functionally Acceptable
Substitutions in Two [alpha]- helical Regions of [lambda] Repressor,"
Proteins: Structure, Function, and Genetics, 7:315, 1990 p. 315
> > > That you try and
> > > shrug it off is just a reflection on your boneheaded willful ignorance
> > > on this subject. Somehow you have to get your estimate to reflect
> > > reality. 18 orders of magnitude off isn't a number to sneaze at. Your
> > > above bogus arguement about detrimental mutations depends on estimates
> > > that could be two fold off, that isn't even 1 order of magnitude. Why
> > > do you think that you are onto something when you are still so far out
> > > of the ball park on this bogus estimate?
> >
> > I'm not out of the ballpark at all. There are functions that do indeed
> > require such higher levels of specificity as well as sequence sizes
> > that go into the thousands of residues - all of which are specified to
> > at least 1 in 1e30 per 100aa.
>
> What are they? Demonstrate that they exist.
"Extrapolating to the rest of the protein indicates that there should
be about 10^57 different allowed sequences for the entire 92-residue
domain. Clearly, this is an extraordinarily rough calculation, and we
do not intend to suggest that we can accurately determine how many
sequences would actually adopt a structure resempling the N-terminal
domain of [lambda] repressor. However, the calculation does indicate in
a qualitative way the tremendous degeneracy in the information that
specifies a particular protein fold."~John F. Reidhaar-Olson and Robert
T. Sauer, "Functionally Acceptable Substitutions in Two [alpha]-
helical Regions of [lambda] Repressor," Proteins: Structure, Function,
and Genetics, 7:315, 1990 p. 315
< snip >
RobinGoodfellow
> RobinGoodfellow wrote:
Scary title, Sean. And quite apt for the fantastic yarn you spin
below.
> < snip >
No comment on how your acceptance of deleterious mutation rates
conflicts with your belief in recent separate creation of all life,
since these rates are computed based on the premise that humans and
other primates diverged from a common ancestor? OK, then.
> > As for the deleterious rate being biased downword in that regard, the
> > conclusion is almost certainly valid. However, the calculation of U
> > has a very large margin of error attached to it. If you use their
> > lower-end estimate of U=1.5, bump it up by, say, 20 percent to adjust
> > for the bias, and apply their calculation, you get 12 offspring per 2
> > parents per generation rather than 40. Still on the high side, but not
> > insurmountable, especially if you actually start taking epistasis into
> > account. The point being, N&C estimate of U=3 are not the final word.
>
> N&C did not think that U was less than 3 due to the fact that they only
> considered a relatively small portion of the genome as being under the
> influence of natural selection.
You seem to give a lot of weight to the fact that N&C seem to think
that the deleterious mutation rate is too high, and not at all to the
fact that they seem to think that humans and chimps diverged from a
common ancestor - even though they use the latter as a premise to
arrive at the former conclusion (the Ka/Ks ratios). Not too big on
consistency, are we?
But, to answer your point, N&C had no way to assess the impact of
non-coding DNA on U. They could only compute U for coding DNA, and
those estimates vary considerably depending on their different
estimates of overall mutation rates. N&C's lower bound estimate of U
is 1.5, and they site another estimate of 1.6 (from 1999, By
Eire-Walker and Keightley) that agrees with that lower bound. I'm not
saying that is necessarily the correct estimate either, but your desire
to accept U>=3 suggests that you've already made up your that we are
devolving towards extinction, and are cherry-picking data to support
this position.
> In fact, the odds are most likely a
> whole lot higher than U=3 now that a lot more of what was thought to be
> "junk" DNA, because it doesn't code for proteins and whatnot, is no
> longer junk, but functional and constrained by natural selection after
> all. This has lead to the suggestion that U is actually "greater than
> 5" (see references below).
The paper you reference by Rice you are referring to simply mentions
this figure in passing, citing three other studies, the latest of which
came of 1993. This pre-dates the U=1.6 estimate of Eire-Walker and
Keightley by 6 years, and N&C's estimate of U=3 by 7 years. Of course,
the ironic part is that Rice's paper proposes specfically how epistasis
effects can compensate for such high deleterious mutation rates.
> Your notion of U is in reality as low as 1.5 is highly unlikely.
It's not my notion at all. I am arguing that in order for your idea to
have some merit, you need to be able to shown not only that U>=3 (Which
is by no means clear), but that epistasis effects aren't real or
sufficient, beneficial mutations can't maintain a an overall level of
fitness, deleterious effects are not, on average, so small as to have
extremely little effect on the overall fitness of the population in
practical terms, and some other things besides. I am certainly willing
to beleive that U>=3, but a) it is far from certain, and b) it hardly
consitutes evidence that we are devolving toward extinction.
[snip]
> < snip >
>
> > But why not read the Kimura and Maruyama
> > article exploring the effects of epistosis, and tell me if you think
> > the situation is so simple?
>
> I have read several of Kimura's papers as well as several papers from
> others dealing with this problem by suggesting some form of epistasis
> (multiplicative increases in the effects of detrimental mutations). The
> problem is that with detrimental mutation rates as high as 3 per person
> per generation, positive epistasis only increases the death rate, but
> does not clear detrimental mutations from the gene pool faster than the
> fitness of the most fit individual in the population decreases.
Apparently, you haven't read enough papers. What you describe is
reinforcing epistasis, which is only one several possible forms of
positive epistasis. Rice, for instance also discusses the possible
effects of buffering and pathway epistasis, both of which can have a
significant effect on reducing genetic load in sexually reproducing
populations according to his simulations.
> Of course, sexual reproduction is supposed to help out with this problem,
> but sexual reproduction doesn't even help unless mating is done in a
> non-random way.
And of course, in real life, mating happens in a purely random fashion.
I don't know if you're married, but if you are, I hope you didn't just
ask the first girl you saw on the street to be your wife - and if you
did, I hope for both your sakes she didn't say "yes". More seriously,
don't you think that an organism suffering many deleterious mutations
will have much lower chances of mating successfully, especially with
organisms of the "most fit" genotype? There is fierce competition for
mates in both the human and the larger animal kingdom: as the result,
genetically fit individuals have a higher propensity of mating with one
another, and individuals of lower fitness are forced to mate with one
another, assuming they can mate successfully at all. This phenomenon
is called positive assortive mating, and Rice shows that it can
drastically reduce mutational load by eliminating whole swaths of
deleterious mutations at once, even in the absence of all other
epistaisis effects.
[snip]
> Of course, there are several problems with this non-random mating
> notion. With a detrimental mutation rate of 3 per individual per
> generation, only 2 of 40 individuals will have a neutral fitness
> balance relative to the original parent population
> What are the odds that these 2 individuals will actually mate "preferentially" so that
> the overall number of individuals with equivalent parent-level fitness
> does not decrease in each generation?
Non-random mating does not mean "only best fit individuals mate with
each other". In most fit individuals do not derive any benefits from
recombination or epistatic effects, since they do not suffer any
deleterious mutations in the first place, relative to the rest of the
population. Non-random mating could mean any number of things,
including the positive assortative mating scenario described above:
that is, individuals within the same fitness classes tend to mate with
one another.
[snip]
> > > > There are multiple
> > > > scenarios as to what can happen once such effects accumulate up to a
> > > > certain threshold, mass extinction being one of them, but certainly not
> > > > the only one.
> > >
> > > What are these other options?
> >
> > a) The fixation of some of the slightly deleterious alleles in the
> > population. While the overall phenotypic fitness will go down, the
> > relative frequency of the "most-fit" phenotype (i.e., the one with the
> > fewest deleterious mutations) would go up, and the overall rate of
> > deleterious mutations would decrease. The population will reach
> > equilibrium, albeit at a slightly less fit level than it used to be.
> > Life will go on.
>
> In every generation every individual receives around 3 deleterious
> mutations. The "most fit" phenotype in each generation comes in at a
> ratio of about 1 in 20. With that ratio, unless very fortuitous mating
> and high reproductive rates take place, the direction is still
> extinction. There is no equilibrium here. The only way to avoid this
> problem is by having the most fit individuals mate pretty much
> exclusively with each other and still make a whole lot of children in
> each generation.
I will concede this point, more or less. My argument was that unlike
in the infinite allele/infinite pool of deleterious mutations models
that are used to derive genetic loads, the potential for deleterious
mutations in a population will go down once a particular slightly
deleterious allele becomes fixed - and in fact, the potential for a
reverse beneficial mutation opens up (more on that below). In other
words, the deleterious mutation rate is not a constant thing. However,
if the number of loci that can undergo such mutations is high enough
and U decreases sufficiently slowly, a population will likely undergo a
very significant reduction in fitness - though not necessarily
extinction - the genetic load approximations start to break down.
> > b) Even as it accumulates slightly deleterious mutations, a population
> > may expand to a certain size where deleterious mutations fail to
> > propagate as quickly,
>
> Deleterious mutations happen on an individual basis - more than 3 per
> individual per generation. These are then passed on to the next
> generation from parents to children and do not leave the gene pool
> until they die out of it with the premature death of an individual who
> didn't get a chance to replicate. They don't need to "propagate" any
> other way to be a problem. Their rate of propagation is not
> significantly related to the size of the population, but to the
> reproductive rate and death rate.
As the population size grows, you will see a broad spectrum of mutants
with accumulation slightly deleterious mutations, but any one specific
deleterious mutation at a specific locus will have a much harder time
achieving fixation in the population. For sexual populations, this is
where recombination can act to eliminate multiple deleterious mutations
in fell swoops, especially in the context of non-random mating and
epistasis.
> > whereas beneficial mutaitons will arise and
> > spread much more rapidly (this is one of the basic theorems of
> > population genetics).
>
> Not quite true.
Ah, Quote-Mining, the discerning creationist's way of doing science
(TM)! None of those quotes actually say what you want them to say when
placed in proper context, Sean. Let's examine closer, shall we?
The first paper, "Beneficial Mutations, Hitchhiking and the Evolution
of Mutation Rates in Sexual Populations" by Toby Johnson (Genetics,
Vol. 151, 1621-1631, April 1999), says nothing about the fixation rates
of beneficial mutations, but rather their effect on modifiers of
mutation rate. They concude that, according to their model,
deleterious mutations have more of effect on the overall mutation rate
due to stronger indirect selection of hitckhiking rate modifiers. It
has nothing to do with whether or not beneficial mutations fix faster
in a population or have more of an effect on the overall fitness of the
population compared to their deleterious counterparts.
> "In sexual populations, the combined effect of beneficial and
> deleterious mutations is to favor a decreased rate of mutation and that
> the indirect selection resulting from beneficial mutations is small or
> negligible compared to that resulting from deleterious mutations. . .
Notice the words "indirect selection", referring to the effect on the
propagation of hitckhiking elements. Also, the following sentences
which are so conveniently replaced by the ellipsis are "However, this
does not necessarily mean that removing the beneficial mutation effect
altogether would result in only a small change in the ESS
[evolutionarily stable mutation rate]. In the absence of any
information about the cost function, a general argument is presented to
explain why this is so."
> Relative to an asexual population, increased levels of recombination
> reduce the effects of beneficial mutations more rapidly than those of
> deleterious mutations" (1).
And, in proper context:
"This article models the long-term effect of a series of such
hitchhiking events and determines the resulting strength of indirect
selection on the modifier. This is compared to the indirect selection
due to deleterious mutations, when both types of mutations are randomly
scattered over a given genetic map. Relative to an asexual population,
increased levels of recombination reduce the effects of beneficial
mutations more rapidly than those of deleterious mutations. However,
the role of beneficial mutations in determining the evolutionarily
stable mutation rate may still be significant if the function
describing the cost of high-fidelity replication has a shallow
gradient."
I am surprised Sean. I would have thought such blatant quote-mining
was beneath you.
> "The probability of fixation of a given beneficial mutation decreases
> with both population size and mutation rate" (2).
This paper, "Adaptive dynamics during experimental evolution of RNA
viruses" by Elena et. al., talks about the adaptive stage in the
evolution of a viruses, when multiple beneficial mutations are in
competition with one another, a phenomenon known as clonal
interference. In this stage, due to the presence of a large variety of
beneficial mutations, "the probability of fixation of a given
beneficial mutation decreases with both population size and mutation
rate", as you so aptly mis-quote. Of course, the "winning" beneficial
mutations are the ones that contribute to the greatest increase in
overall population fitness. Again, this paper says nothing about the
effects on fitness from beneficial mutations versus the effects from
deleterious ones, or the differences in their fixation rates.
> "Moreover, deleterious mutations reduce the chance of
> fixation of advantageous mutations, since they increase the occurrence
> of such an event in a genome that already has a large number of
> segregated deleterious mutations. In this way, to be successful,
> mutations that have larger selective benefit must be produced (de
> Oliveira and Campos 2004)" (3).
Can't access the article, but the paper in question is "Mutational
Effects on the Clonal Interference Phenomenon", by Campos and Oliveira.
The abstract reads:
"We study the process of fixation of beneficial mutations in an asexual
population by means of a theoretical model. Particularly, we wish to
investigate how the supply of deleterious and beneficial mutations
influences the dynamics of the adaptive process of an evolving
population. It is well known that the deleterious mutations drastically
affect the fate of beneficial mutations. In addition, an increasing
supply of favorable mutations, to compensate the decay of the fitness
due to the accumulation of deleterious mutations, produces the clonal
interference phenomenon where advantageous mutations in distinct
lineages compete to reach fixation. This competition imposes a limit to
the speed of adaptation of the population. Intuitively, we would expect
that the interplay of the two mechanisms would conspire to ensure
fixation of only large-effect beneficial mutations. ***Our results,
however, show that beneficial mutations of small effect have an
increased probability of fixation when both beneficial and deleterious
mutations rates are increased***." [Emphasis mine]
> "Rare reverse and compensatory
> mutations can move deleterious mutations, via genetic hitchhiking,
> against the flow of genetic polarization. But this is a minor
> influence, analogous to water turbulence that occasionally transports a
> pebble a short distance upstream. . .
We are now back to Rice's article. First, notice here here that he is
talking about reverse and compensating mutations - i.e. mutations that
specifically undo the effects of a preceding deleterious mutation - and
not beneficial mutation in general. Secondly, at this point, he is
talking about asexually reproducing populations specifically. Later
on, Rice goes on to say:
"Overall, epistasis and nonrandom mating can cause recombination to
build the best class faster than its own net reproductive rate.
Clearly, when R_best(realized) << R_best(req) [the best realized
reproductive rate is much smaller than the required rate], then
deterministic mutation accumulation will lead to extinction. But when
the increment (R_best(req) - R_best(realized)) is smaller, a mutational
Red-Queen may ensue with mutation accumulation being accommodated by
perpetual compensating adaptation. Recombination, by reducing
R_best(req) via epistasis and nonrandom mating, extends the permissible
range of phenotypic complexity (large Ud [rate of deleterios mutation])
that can potentially evolve."
> Whenever demographic, ecological,
> and/or physiological constraints cause, R-best t to be less than eUd,
> then the progenitor class will decline in size each generation and
> deterministic mutation accumulation will ensue. Such mutation
> accumulation will be opposed by reverse and compensatory mutations, but
> if R-best is much less than eUd, then net mutation accumulation will
> ensue" (4).
Notice the constant use of qualifiers such as "whenever" and "if" in
this parpagraph. And again, he is talking about reverse and
compensatory mutations specifically: not about adaptations in general.
> Especially note that the rate of fixation of beneficial mutations is
> less in sexual than in asexual populations.
None of the links you gave actually suggest this, unless you grossly
take them out of context. And even if it were true, it does not mean
that the rate of fixation of beneficial mutations in sexual (or
asexual, for that matter) populations is smaller than the rate of
fixation of deleterious mutations, or, more importantly that the net
effects on fitness of slightly deleterious fixations is greater than
the net effect on fitness of their slighly and not-so-slightly
beneficial counterparts.
[snip]
> > Thus, over time, overall fitness of the
> > population may increase as it grows and mutates. Small negative and
> > positive epistasis effects can play an important role here, in both
> > helping to eliminate deleterious mutations and in propagating
> > beneficial ones.
>
> Epistasis does not change the need for a certain death rate and
> reproductive rate. Epistasis may preferentially remove mutations from
> those with the most detrimental mutations in a population, but it
> doesn't remove them faster than they are formed. That's the problem.
The problem appears to be that you didn't so much read Rice's paper as
scanned it for quotes to mine. Go back to the paper, look at his
figure 3, read his analysis and conclusions. It certainly disagrees
with your unsupported assertions.
> The only way to get rid of negative mutations faster than they are
> introduced, without increasing the death rate, is to somehow
> concentrate them in a few individuals at a higher rate than the
> detrimental mutations are formed. Epistasis alone does not do this. In
> fact, there is no way I can think of as to how this concentration might
> occur - outside of ID that is.
I have no idea what this means. As far as I can parse this, you are
saying that the "Intelligent" Designer created many species with far
higher mutation rates than populations could tolerate, and then, upon
realizing his screw up and so as to avoid extincition, he is directing
the bulk of negative mutations into a relatively small subset of
individuals so that they die off! I knew that accepting ID required
one to accept that the designer is either incompetent or malicious, but
here you seem to be saying that he (or she) is both! Please tell me I
misunderstood you!
> > c) A population may reach a limit where a certain proportion of
> > deleterious mutations will no longer be tolerated. This will cause a
> > massive die-off of the organisms possessing such genotypes, thereby
> > boosting the proportion of most-fit individuals in the population.
>
> Yes, and these most fit individuals might be just above the threshold
> of die-off themselves. In order to maintain their level of
> just-barely-making-it fitness, their reproductive rate would have to
> skyrocket.
I was trying (admittedly, very poorly) to describe a scenario where
truncation selection takes place, the extereme case of which is the
accumulation of deleterious mutations with no measureable decline in
fitness until a certain threshold is reached, which would cause genetic
death of the organism. Kimura and Maruyama (Genetics 54:1337-1351)
show that under such a scenario, the mutation load can be several times
lower than in the default case for an asexual population. Whether or
not truncastion selection actually exists in nature is another story.
It is a purely theoretical construction, but so are most of the
population genetic models.
[snip]
> > > > 5) The accumulation of positive selective effects in the population.
> > >
> > > These do not accumulate nearly as fast as the negative mutations - Even
> > > in literature the suggested ratio is less than 1 beneficial to 1000
> > > negative.
> >
> > Do you mean "classic English literature"? Because, in the scientific
> > literature, the ratios are considerably higher. Looking at the article
> > by Bustamante et. al., they have found evidence that, of the genetic
> > loci they examined, 9.0% show signs of positive selection, whereas
> > 13.5% show signs of negative selection.
>
> That is not the "rate" of beneficial vs. detrimental mutations. The
> rate of detrimental mutations far outpaces the rate of beneficial
> mutations.
I acknowledge that, and below, I have shown how to compute the ratio of
beneficial to determinetal mutations from Bustamante et. al. Using
Nachmann and Crowell's estimate of 73% of non-synonymous subsitituions
being deleterious, we can compute the rate as 30 beneficial mutations
to 1000 deleterious.
> For E. coli, the estimated value for the beneficial mutation rate
> (Miralles et. al., 1999) was 6.4 × 1e-8 beneficial mutations per
> genome per generation. (1) The beneficial mutation rate obtained by
> Imhof and Schlötterer was 4 x 1e-9 per genome per generation (see
> reference links below). (2) Compare this with the detrimental mutation
> rate for E. coli "in excess of 0.0002" per genome per generation. (3)
Why quote beneficial mutation rates in viruses and bacteria, when I'm
giving you the most recent estimates for humans? Rates of both
beneficial and deleterious mutations differ greatly between these
organisms. It's quite simple to compute these ratios in human coding
regions using the estimates of Bustamante et. al., coupled with N&C's
or Kimura's estimate of the proportion of deleterious non-synonymous
substitutions.
> That produces a ratio of between 1 in ~3,000 to 1 in ~50,000. But what
> about eukaryotes? "In sexual populations of higher eukaryotes, there is
> extensive data showing that U >> K." (4)
Certainly, taking K/U = 3/100, U >> K. But that K/U ratio is still
much greater than 1/1000 that you were claiming, with no justification
from human genomic data.
> "In general, organisms with
> larger genomes appear to have a greater number of deleterious
> mutations, although it does not appear that the deleterious mutation
> rate is constant per base pair across these organisms." (5)
Quoting rampantly again? This only means that the rates of deleterious
mutations are higher in organisms with larger genomes, not that the
rates of beneficial mutations don't increase in the same or even
greater proportions.
[snip]
> > Now, even if we were to accept
> > Kimura's estimate that 86% of all functional mutations in primates are
> > negative, we would still arrive at the conclusion that 0.09 * (1-0.86)
> > = 1.3% percent of all functional mutations are beneficial, whereas .86
> > + .135*(1-0.86)= 87.9% of all functional mutations are deleterious.
> > That gives a ratio of 15 beneficial to 1000 negaitve, which is 15 times
> > what you state. Of course, this ratio would be even bigger using N&C
> > estimates of the fraction of deleterious mutations (about 30 to 1000).
>
> Perhaps you've misread Kimura?
I haven't read that particular paper by Kimura - I took the figure from
N&C.
> It seems that the "86%" number is not the number of all "functional" mutations,
> but of nonsynonymous substitutions in a functional protein - like the hemoglobin
> protein used by Kimura in this particular study.
My bad. I should have written "non-synonymous" instead of
"functional".
> N&C write:
>
> "What proportion of nonsynonymous changes are neutral and what
> proportion are deleterious? The fraction that are neutral, fo, can be
> calculated by comparing the total mutation rate, µt, with the
> substitution rate, vo = foµt (KIMURA 1983A, KIMURA 1983B). The
> proportion that are deleterious is 1 - fo. Using this approach, KIMURA
> 1983B estimated that 86% of nonsynonymous substitutions are
> deleterious. . .
Yep, that's where I got the 86% percent figure.
> The genomic deleterious mutation rate is likely much
> larger given our estimate that 80% of amino acid mutations are
> deleterious and given that it does not include deleterious mutations in
> noncoding regions, which may be quite common."
>
> http://www.genetics.org/cgi/content/full/156/1/297
This phrase does not appear anywhere in N&C's 2000 Genetics paper you
link to. Are you sure you have your quotes mined correctly?
Also, note that while we cannot estimate the rate of potential
deleterious substitutions in the non-coding regions, we cannot estimate
the rate of potential beneficial substitutions in those regions,
either, but there's no reason to believe that beneficial mutations
would be more uncommon, relatively speaking, in those regions, than
deleterious mutations. Regardless, estimates of beneficial to
deleterious ratios for the coding region of the human genome is all
we've got at the moment.
>
> > > > This is not taken into account by N&C at all, but positive mutations do
> > > > happen, and while more rare than their deleterious counterparts, their
> > > > effects on the overall fitness of the population can be just as
> > > > important, if not more so.
> > >
> > > They are far more rare. That's the problem.
> >
> > Yes, they are more rare, but they spread through populations and fix
> > far more quickly than deleterious mutations. As the result, the
> > balance of positive versus negative mutations can maintain or even
> > increase the overall fitness of a population, despite their relative
> > rarity.
>
> Not quite true, as discussed above.
Yes, if by "discussed" you mean cherry-picking a set of quotes from a
few papers, removing all context, and grossly misrepresenting what the
authors were trying to say.
> > > > Recently, Bustamante et. al. (Nature
> > > > 437:1153-57) have shown that a considerable number of loci in the human
> > > > genome appear to be undergoing positive selection, and simulations show
> > > > that advantageous mutations tend to become fixed in a population both
> > > > faster and in greater proportion than their deleterious counterparts,
> > > > even under evolutionary scenarios where the overwhelming majority of
> > > > non-neutral mutations happen to be deleterious. In other words, the
> > > > small balance of positive mutations, due to their faster fixation
> > > > times, can maintain or increase the overall fitness of a population,
> > > > even as it keeps accumulating slightly deleterious mutations.
> > >
> > > There is no reason to suggest that a slight positive mutation is fixed
> > > any faster in a population than a negative mutation of equal degree can
> > > be eliminated from a population.
> >
> > Of course there is. The longer a slightly positive mutation stays in
> > the population, the more likely it is to spread. Unless such a
> > mutation is lost quickly, it will almost certainly achieve fixation.
>
> Not exactly true, as discussed above.
Yes, if by "discussed" you mean cherry-picking a set of quotes from a
few papers, removing all context, and grossly misrepresenting what the
authors were trying to say. See, kids - copying and pasting is easy,
fun, and saves much thought and effort!
> > On the other hand, slightly deleterious mutations are still subject to
> > negative selection in every generation, and the chances of their
> > continued presence in the gene pool are not improved with each passing
> > generations, at least in steady-state populations.
>
> But the only way to get rid of these negative mutations faster than
> they are being formed is by preferentially concentrating them in some
> members of a population, or by increasing the reproductive rate
> dramatically. Epistasis doesn't solve this problem - as discussed
> above.
All the above discussion reveals is that you are don't really
understand what epistasis is, or the way in which it may work in
conjunction with recombinations to reduce mutational loads.
> > So, while they may
> > enter the gene pool more quickly, they will also be eliminated more
> > quickly, especially if they accumulate and the deleterious effect is
> > amplified.
>
> They will not be eliminated more quickly than they are produced in a
> slowly reproducing population - that's the problem.
They can be, at least reduced to the extent where the balance of
beneficial mutations can offset the difference and even improve overall
fitness.
Ron O
Seanpit wrote:
> Ron O wrote:
>
> > How do you know that? Answer the question. It isn't in the papers
> > that you cite, because I know for a fact that Crow admits that he
> > doesn't understand the mechanism by which these mutations are removed.
> > We don't know everything, and Crow is too good of a scientists to claim
> > otherwise.
>
> Perhaps Crow doesn't know how the detrimental mutations are removed
> because there isn't any good way to remove them? How have you helped
> solve the problem that is in fact quite a mystery indeed?
This is the same stupidity that ID is based on. Some idiot can't
imagine something so it is impossible. Since anyone with half a brain
knows that science doesn't know everything, what is your excuse? This
is such a lame argument that you should be ashamed to put it forward.
>
> < snip irrelevant personal comments>
>
> > > > This is just
> > > > another of your bogus assertions that you can't back up. Prove that
> > > > the reproductive rate hasn't been high enough when the masses of data
> > > > tell us that you are wrong.
> > >
> > > Tell me, where do you see any woman having over 40 pregnancies? And,
> > > that's just with U equal to 3. Many suggest that U is actually over 5.
> > > Do you think it possible for the average human woman to have over a
> > > hundred pregnancies?
> >
> > Demonstrate that this many pregnancies are required of human females.
>
> Read the papers Ron. This estimate is not based on my calculations.
> Don't take my word for it. I've directly quoted several relevant
> sections to these papers and you still don't understand. Look again at
> the following quote:
I would bother to read the papers, but it would be a waste of time
since I am not currently interested in that line of research. Why
would it be a waste of time, because it doesn't matter if I read them
or not. Your arguments are bogus and you know it. The simple reason
that I can say that is that you will not put forward your own junk for
evaluation. The reason why you will not do this is because you know
that your junk is worse off than the real science that you try to
obfuscate and don't like.
Demonstrate that it is worth going to read the papers. Put up your
evidence for your beliefs that is as good as the evidence for common
descent that you claim isn't good enough.
That is simple and direct, since you have claimed to have such
evidence, all you have to do is put it forward.
>
> "The reduction in fitness (i.e., the genetic load) due to
> deleterious mutations with multiplicative effects is given by 1 - e^-U
> (Kimura and Moruyama 1966). For U = 3, the average fitness is reduced
> to 0.05, or put differently, each female would need to produce 40
> offspring for 2 to survive and maintain the population at constant
> size. This assumes that all mortality is due to selection and so the
> actual number of offspring required to maintain a constant population
> size is probably higher." - Nuchman, Michael W., Crowell, Susan L.,
> Estimate of the Mutation Rate per Nucleotide in Humans, Genetics,
> September 2000, 156: 297-304 (
> http://www.genetics.org/cgi/content/full/156/1/297? )
Bumble bees can't fly because some guy calculated that it was
impossible. The fact is that we see bumble bees flying and we observe
the after effects of removal of all these detrimental mutations. How
do you remove them all in just the few thousand years since the ark?
You can't just use some lame argument if it means that your own
explanation is bogus.
>
> >From this calculation, the number of births per female is 2 * (e^U).
>
> When U = 3, the number of required births is 2 * (2.72^3) = ~ 40 births
>
> And, this is felt to be a very low estimate. Consider the following
> excerpt, which I've already referenced in my original post to this
> thread:
>
> "Accurate estimates of Ud for most metazoans with large genomes
> (e.g., vertebrates) are not yet available but extrapolations from
> studies of humans and Drosophila (Mukai, 1979; Kondroshov, 1988; Crow,
> 1993) suggest that Ud > 5 is feasible. In this case an asexual
> population would have to have Rbest > e5 =148, a value far beyond the
> physiological/ecological capabilities of most vertebrates."
>
> http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
>
> For a *sexual* population, U=5 means that the average woman would have
> to have over 296 pregnancies.
>
> > There is that two fold estimate again.
>
> It's not my estimate. Read the papers.
Again, why? We observe that the mutations have been removed. What is
your explanation for those observations?
>
> > What about 18 orders of
> > magnitude that you can't deal with? How can you say that 40
> > pregnancies are needed when you admit that you don't know the mechanism
> > by which the detrimental mutations are removed?
>
> I don't think they are removed. I think that we have always been headed
> downhill since our very recent creation just a few thousand years ago.
Then explain how you get the current data that tells us that they
existed and that they have been removed. How can you get a 3:1 ratio
when the average for going down hill should be 1:2 or if you are
correct about the selectable sequence space it would be much higher if
we could figure out what sequences you were talking about.
>
> > How have the
> > detrimentals been removed from the existing populations to get the 3:1
> > ratio in just the few thousand years you have to work with?
>
> You're not dealing with all types of detrimental mutations in this
> estimate of yours.
That is correct, so you probably have even more to figure out how they
were removed. When we find control regions we find that they are well
conserved but the sequence around them changes to randomness. You have
to figure out how all the mutations that should have made the conserved
sequence random compared to other species got removed from the
population for every such sequence in the genome. The problem is that
we don't know what sequences you are talking about so we can't do the
comparisons. Your sequences might not even exist in the numbers that
you claim. How much of the genome is under such selection?
>
> > You
> > require rates so much greater than real science that you should give up
> > on this line of argument. The 3:1 is an average for every gene.
> > Science claims that it has literally millions of years to produce that
> > ratio, you are around 3 orders of magnitude worse off to make it fit
> > your model. You have to make these arguments fit your model, you can't
> > just claim that it is a problem for science.
>
> Not if you consider other types of detrimental mutations. I'm not just
> talking about "recessive lethals" here - many of which are quite common
> actually.
No matter what types of mutations you are talking about, just in these
types you have to deal with the 3 orders of magnitude difference.
Adding other types of mutations just makes it worse for you. Just
think about it for a second. You are just digging a deeper hole for
yourself.
>
> < snip >
>
> > > > Some species have much higher genetic loads and they still reproduce
> > > > and do just fine.
> > >
> > > Sure they do. That's not the question. The question is if there are as
> > > many members of a particular generation that are as least as fit as the
> > > most fit members of the previous generation. The answer to this
> > > question for humans and other slowly reproducing creatures, is no.
> >
> > This is bogus and you can't demonstrate it because the populations
> > haven't been tested that you need to test. I'm certain that Crow has
> > something in the discussion about the fact that the human population is
> > expanding. There are models where expanding populations can accumulate
> > more detrimental mutations. What you have to worry about is if our
> > population was shrinking over the last 1000 years. More individuals
> > are surviving. Humans can have fewer pregnancies and still increase
> > the population. What happens to selection pressure?
>
> Yes, the population can be booming and still the numbers in the
> population with at least the level of fitness of the previous
> generation continues to shrink in each generation. That is what is
> called the "Living Dead" problem. There can be quite a number of
> living dead and this number can actually grow - i.e., the population of
> "living dead" can increase for a time. However, the direction of the
> overall population is still toward extinction.
>
> http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
So what does that say about your current estimate. Is it biased?
Could it be misleading you. Just think about the direction of the bias
for your inferences. The bias is a problem isn't it?
>
> < snip >
>
> > > > Actually no, selection works best if you actually inbreed so that
> > > > recessive detrimentals can become homozygous and be selected against.
> > > > You can't identify the most fit to breed together when you are dealing
> > > > with the majority of detrimental mutations.
> > >
> > > You're wrong as far as the problem at hand is concerned. Selective
> > > breeding of only the fittest with the fittest is one way to slow down
> > > the decline. Random breeding or inbreeding, only speeds up the overall
> > > decline of a population and hastens either extinction or a dramatic
> > > increase in reproductive rates to overcome the problem.
> >
> > Where did you read this, cite the paper. I've never seen this
> > conclusion, ever. Just think for a moment, you can't
> > determine the fittest unless you inbreed.
>
> http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
Bullshit. Put up the quote. Really I'm not going to bother to go read
anything that you claim because it isn't worth the time. Until you can
demostrate that you can actually reason, why should I? Heck, I only
respond to your quackery when you respond to one of my posts, your
posts aren't even worth reading.
What is your evidence for your beliefs and why is that bogus evidence
better than what science has? If you can't do that why should anyone
listen to you?
>
> > > I don't shrug such numbers off. There certainly are far less specified
> > > functions in low level sequence space. Some functions may have ratios
> > > as low as 1 in 5 or 1 in 2 or even 1 in 1. However, these are not what
> > > I would call "fairly specified" functions. And, there certainly are
> > > functions that do indeed show ratios lower than 1 in 1e30 per 100aa.
> > > That's the whole point. This level of minimum functional complexity is
> > > much higher than the levels you are talking about - which are even
> > > easier to evolve.
> >
> > What is unspecified about new enzymatic function? Demonstrate that the
> > functions that you are interested in are more "specified" than the
> > antibody functions. You are out of the ball park by 18 orders of
> > magnitude and you still claim that you can make viable conclusions?
> > How?
>
> You need to read the papers by Sauer et al., which I've referenced
> several times recently in this forum, which describe protein-based
> functions with ratios as low as 1 in 1e62 per 100aa. Estimated ratios
> for other types of protein-based functions, like cytochrome c are at
> least as low as 1 in 1e30 per 100aa (most recent Yockey estimate).
>
> John F. Reidhaar-Olson and Robert T. Sauer, "Functionally Acceptable
> Substitutions in Two [alpha]- helical Regions of [lambda] Repressor,"
> Proteins: Structure, Function, and Genetics, 7:315, 1990 p. 315
Who the heck cares? No matter what, you stil have to figure out how
antibodies work if these guys estimates mean anything. Really, we have
reality and it conflicts with these estimates, what would you go with,
your bogus inferences from these estimates or reality? How do
antibodies work if they only have a maximum of 10e12 sequences tested?
Why should anyone believe what you claim about some 10e62 claim when it
seems to be 50 orders of magnitude wrong?
>
> > > > That you try and
> > > > shrug it off is just a reflection on your boneheaded willful ignorance
> > > > on this subject. Somehow you have to get your estimate to reflect
> > > > reality. 18 orders of magnitude off isn't a number to sneaze at. Your
> > > > above bogus arguement about detrimental mutations depends on estimates
> > > > that could be two fold off, that isn't even 1 order of magnitude. Why
> > > > do you think that you are onto something when you are still so far out
> > > > of the ball park on this bogus estimate?
> > >
> > > I'm not out of the ballpark at all. There are functions that do indeed
> > > require such higher levels of specificity as well as sequence sizes
> > > that go into the thousands of residues - all of which are specified to
> > > at least 1 in 1e30 per 100aa.
> >
> > What are they? Demonstrate that they exist.
>
> "Extrapolating to the rest of the protein indicates that there should
> be about 10^57 different allowed sequences for the entire 92-residue
> domain. Clearly, this is an extraordinarily rough calculation, and we
> do not intend to suggest that we can accurately determine how many
> sequences would actually adopt a structure resempling the N-terminal
> domain of [lambda] repressor. However, the calculation does indicate in
> a qualitative way the tremendous degeneracy in the information that
> specifies a particular protein fold."~John F. Reidhaar-Olson and Robert
> T. Sauer, "Functionally Acceptable Substitutions in Two [alpha]-
> helical Regions of [lambda] Repressor," Proteins: Structure, Function,
> and Genetics, 7:315, 1990 p. 315
This just claims that there is a 92 residue domain involved, but it
also claims that their estimate is that 10^57 sequences will do the
same function just in that limitation, what about a different protein
or a dimer or tetramer? What do you think "degeneracy" means in terms
of the antibody example? What about matching another DNA sequence that
would have worked just as well as a repressor site? How many of those
are around if they all have around a 10^57 degeneracy for proteins that
would interact with them?
Snipping the junk that you are too incompetent to deal with. Ignoring
your mental problems doesn't make them go away. If you want to claim
simple dishonesty, than do it. Just state that you refuse to present
your wonderful evidence for your beliefs because it is too wonderful
for unwashed eyes to see, and everyone will know that you are either
lying or crazy. Simply pretending that the problem doesn't exist is
crazy. It is the stupidest dishonesty that you could perpetrate. You
have to be crazy or desperate to do it.
Ron Okimoto
>
Seanpit
Ron O wrote:
> Seanpit wrote:
> > Ron O wrote:
> >
> > > How do you know that? Answer the question. It isn't in the papers
> > > that you cite, because I know for a fact that Crow admits that he
> > > doesn't understand the mechanism by which these mutations are removed.
> > > We don't know everything, and Crow is too good of a scientists to claim
> > > otherwise.
> >
> > Perhaps Crow doesn't know how the detrimental mutations are removed
> > because there isn't any good way to remove them? How have you helped
> > solve the problem that is in fact quite a mystery indeed?
>
> This is the same stupidity that ID is based on. Some idiot can't
> imagine something so it is impossible. Since anyone with half a brain
> knows that science doesn't know everything, what is your excuse? This
> is such a lame argument that you should be ashamed to put it forward.
And so, if you can't present anything besides ID to explain a given
phenomenon, what makes to think that you have to search for a non-ID
explanation until you find one? Where in science is that a
requirement? How long will you search before you given up and accept
that ID might actually be a viable explanation for some things that
humans aren't actually responsible for? How long would you wait before
suggesting ID yourself if you saw the same person winning the
California Lottery over and over again? It is possible - right? But
is it likely? ; )
> > > Demonstrate that this many pregnancies are required of human females.
> >
> > Read the papers Ron. This estimate is not based on my calculations.
> > Don't take my word for it. I've directly quoted several relevant
> > sections to these papers and you still don't understand. Look again at
> > the following quote:
>
> I would bother to read the papers, but it would be a waste of time
> since I am not currently interested in that line of research.
Then you have nothing to say here now do you? You haven't read the
papers and you're not about to read them because you are not interested
in them. Yet, somehow you feel qualified to tell me I'm wrong when I
only present what the authors of these papers are saying themselves.
>Why
> would it be a waste of time, because it doesn't matter if I read them
> or not. Your arguments are bogus and you know it.
LOL - That's just too funny Ron. These particular arguments, arguments
that you are calling "bogus", such as the high required birth rates,
are not mine. The authors themselves present them.
> The simple reason
> that I can say that is that you will not put forward your own junk for
> evaluation.
What?
> The reason why you will not do this is because you know
> that your junk is worse off than the real science that you try to
> obfuscate and don't like.
I'm putting forward what the authors themselves are saying Ron.
> Demonstrate that it is worth going to read the papers. Put up your
> evidence for your beliefs that is as good as the evidence for common
> descent that you claim isn't good enough.
>
> That is simple and direct, since you have claimed to have such
> evidence, all you have to do is put it forward.
You're one of the biggest self-deluded people here if you keep talking
this way. You think "I don't have to read the evidence you're talking
about because I already know that whatever you say is already wrong. No
need to look into it any further."
> > "The reduction in fitness (i.e., the genetic load) due to
> > deleterious mutations with multiplicative effects is given by 1 - e^-U
> > (Kimura and Moruyama 1966). For U = 3, the average fitness is reduced
> > to 0.05, or put differently, each female would need to produce 40
> > offspring for 2 to survive and maintain the population at constant
> > size. This assumes that all mortality is due to selection and so the
> > actual number of offspring required to maintain a constant population
> > size is probably higher." - Nuchman, Michael W., Crowell, Susan L.,
> > Estimate of the Mutation Rate per Nucleotide in Humans, Genetics,
> > September 2000, 156: 297-304 (
> > http://www.genetics.org/cgi/content/full/156/1/297? )
>
> Bumble bees can't fly because some guy calculated that it was
> impossible.
That was for a fixed-wing model. This is an urban legend Ron.
Do you actually disagree with the numbers listed by these authors? If
so, upon what basis do you disagree?
> The fact is that we see bumble bees flying and we observe
> the after effects of removal of all these detrimental mutations. How
> do you remove them all in just the few thousand years since the ark?
> You can't just use some lame argument if it means that your own
> explanation is bogus.
You can remove many detrimental mutations in a relatively short time.
That's not the problem. The problem is restoring the level of maximum
fitness in a population to the original level of maximum fitness that
the previous generation had. That's a whole different problem. That
problem is what creates the notion of the "Living Dead".
> > >From this calculation, the number of births per female is 2 * (e^U).
> >
> > When U = 3, the number of required births is 2 * (2.72^3) = ~ 40 births
> >
> > And, this is felt to be a very low estimate. Consider the following
> > excerpt, which I've already referenced in my original post to this
> > thread:
> >
> > "Accurate estimates of Ud for most metazoans with large genomes
> > (e.g., vertebrates) are not yet available but extrapolations from
> > studies of humans and Drosophila (Mukai, 1979; Kondroshov, 1988; Crow,
> > 1993) suggest that Ud > 5 is feasible. In this case an asexual
> > population would have to have Rbest > e5 =148, a value far beyond the
> > physiological/ecological capabilities of most vertebrates."
> >
> > http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
> >
> > For a *sexual* population, U=5 means that the average woman would have
> > to have over 296 pregnancies.
> >
> > > There is that two fold estimate again.
> >
> > It's not my estimate. Read the papers.
>
> Again, why? We observe that the mutations have been removed. What is
> your explanation for those observations?
Many detrimental mutations have been removed, but not all of them. Not
even close. Not all detrimental mutations are "recessively lethal"
Ron. Very small detrimental effects build up over time in a population
in a way in which the population as a whole can survive and even grow
and thrive for quite some time. However, all the while the total
maximum level of fitness of this slowly reproducing population is
declining. That is why the vast majority of a population is called,
"The Living Dead". Only a relatively few individuals in each
generation are as fit as the maximally fit individuals in the previous
generation.
< snip >
> > > > You're wrong as far as the problem at hand is concerned. Selective
> > > > breeding of only the fittest with the fittest is one way to slow down
> > > > the decline. Random breeding or inbreeding, only speeds up the overall
> > > > decline of a population and hastens either extinction or a dramatic
> > > > increase in reproductive rates to overcome the problem.
> > >
> > > Where did you read this, cite the paper. I've never seen this
> > > conclusion, ever. Just think for a moment, you can't
> > > determine the fittest unless you inbreed.
> >
> > http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
>
> Bullshit. Put up the quote. Really I'm not going to bother to go read
> anything that you claim because it isn't worth the time. Until you can
> demostrate that you can actually reason, why should I? Heck, I only
> respond to your quackery when you respond to one of my posts, your
> posts aren't even worth reading.
The entire paper in the reference is about this problem Ron. If you
won't read it, why should I bother. You've been wrong on all of your
other challenges in this thread so far Ron. You were wrong about the
minimum birth requirements not going into the hundreds. You also
called that number "BS" if I recall. Just read the paper Ron. It isn't
that long of a paper or all that difficult to read.
I've never seen someone who claims to be a scientist argue that someone
else is wrong about what a paper says while refusing to actually read
the paper in question.
> What is your evidence for your beliefs and why is that bogus evidence
> better than what science has? If you can't do that why should anyone
> listen to you?
I've just given you the reference Ron - a reference to a real article
in a real scientific journal. What more can I do if you won't read it?
"You can take a horse to water . . . "
> > > > I don't shrug such numbers off. There certainly are far less specified
> > > > functions in low level sequence space. Some functions may have ratios
> > > > as low as 1 in 5 or 1 in 2 or even 1 in 1. However, these are not what
> > > > I would call "fairly specified" functions. And, there certainly are
> > > > functions that do indeed show ratios lower than 1 in 1e30 per 100aa.
> > > > That's the whole point. This level of minimum functional complexity is
> > > > much higher than the levels you are talking about - which are even
> > > > easier to evolve.
> > >
> > > What is unspecified about new enzymatic function? Demonstrate that the
> > > functions that you are interested in are more "specified" than the
> > > antibody functions. You are out of the ball park by 18 orders of
> > > magnitude and you still claim that you can make viable conclusions?
> > > How?
> >
> > You need to read the papers by Sauer et al., which I've referenced
> > several times recently in this forum, which describe protein-based
> > functions with ratios as low as 1 in 1e62 per 100aa. Estimated ratios
> > for other types of protein-based functions, like cytochrome c are at
> > least as low as 1 in 1e30 per 100aa (most recent Yockey estimate).
> >
> > John F. Reidhaar-Olson and Robert T. Sauer, "Functionally Acceptable
> > Substitutions in Two [alpha]- helical Regions of [lambda] Repressor,"
> > Proteins: Structure, Function, and Genetics, 7:315, 1990 p. 315
>
> Who the heck cares? No matter what, you stil have to figure out how
> antibodies work if these guys estimates mean anything.
They only work at low levels of functional complexity.
> Really, we have
> reality and it conflicts with these estimates, what would you go with,
> your bogus inferences from these estimates or reality? How do
> antibodies work if they only have a maximum of 10e12 sequences tested?
> Why should anyone believe what you claim about some 10e62 claim when it
> seems to be 50 orders of magnitude wrong?
The functions that these antibody tests have discovered have all been
lower level functions having much higher ratios in sequence space.
None of these discovered functions have been of the types discussed in
the papers written by Yockey and Sauer. If you think this is wrong,
why hasn't anyone challenged the estimates of Yockey and Sauer in the
literature? If you think this is wrong, then please do provide your
own references.
A dimer or tetramer having more than 92 residues would not produce the
same ratio. The authors themselves quote Yockey as supporting their
estimate for about 1 in 1e62 sequences with this particular type of
function in all of sequence space for 92aa proteins.
> What do you think "degeneracy" means in terms
> of the antibody example? What about matching another DNA sequence that
> would have worked just as well as a repressor site? How many of those
> are around if they all have around a 10^57 degeneracy for proteins that
> would interact with them?
That doesn't matter. What matters is the ratio for this particular
function.
Read the papers Ron. You'll be better off.
< snip more personal comments >
Seanpit
What do you think the following means?:
"Using conservative calculations of the proportion of the genome
subject to purifying selection, we estimate that the genomic
deleterious mutation rate (U) is *at least 3* . . . In fact, this range
is likely to be biased downward because we have considered only
nonsynonymous sites as potential targets for deleterious mutations."
[emphasis added]
http://www.genetics.org/cgi/content/full/156/1/297
> And as I expected, you just ignored my questions about common descent
> and flood geology.
Look, I've discussed common descent with you extensively. I've also
discussed flood geology with many others extensively. I just don't have
enough time to do it all the time. I am currently interested in the
problem of neutral gaps and random walks and random sampling. I'm
spending what little time I have on that right now. Is that Ok with
you? Remember, there are a lot more of you than there are of me. Your
personal interests, while certainly interesting to me, are not always
my primary interest at the moment.
Sean Pitman
www.DetectingDesign.com
Seanpit
Ron O wrote:
> How do you know that? Answer the question. It isn't in the papers
> that you cite, because I know for a fact that Crow admits that he
> doesn't understand the mechanism by which these mutations are removed.
> We don't know everything, and Crow is too good of a scientists to claim
> otherwise.
Oh, but you don't actually have to read the papers? Interesting . . .
See the rest of my response in the link to a lower thread:
http://groups.google.com/group/talk.origins/msg/40677f70ae8740f2?dmode=source
Sean Pitman
www.DetectingDesign.com
John Harshman
Not much that's relevant to the question I asked. If that's the best you
have, it would be wise to stop making the claim.
> http://www.genetics.org/cgi/content/full/156/1/297
>
>
>>And as I expected, you just ignored my questions about common descent
>>and flood geology.
>
> Look, I've discussed common descent with you extensively.
No, you've evaded discussing it extensively. Your repertoire has been
limited to citation of irrelevant papers and a network of facile denials
without any meat to them. If that's all you have, might as well not talk
about it at all. You refuse to engage seriously with this question.
> I've also
> discussed flood geology with many others extensively. I just don't have
> enough time to do it all the time. I am currently interested in the
> problem of neutral gaps and random walks and random sampling. I'm
> spending what little time I have on that right now. Is that Ok with
> you? Remember, there are a lot more of you than there are of me. Your
> personal interests, while certainly interesting to me, are not always
> my primary interest at the moment.
Our primary interests do not coincide. I'll thank you not to make claims
that depend on the particular unwarranted assumptions that fall into my
area of interest. Stop claiming, for example, that your "neutral gaps"
mean that common descent isn't possible.
Seanpit
> > What do you think the following means?:
> >
> > "Using conservative calculations of the proportion of the genome
> > subject to purifying selection, we estimate that the genomic
> > deleterious mutation rate (U) is *at least 3* . . . In fact, this range
> > is likely to be biased downward because we have considered only
> > nonsynonymous sites as potential targets for deleterious mutations."
> > [emphasis added]
>
> Not much that's relevant to the question I asked. If that's the best you
> have, it would be wise to stop making the claim.
It is very relevant to your question. Why would you think otherwise if
the authors themselves suggest that their numbers are biased downward
by this? - That there are probably more potential targets that they are
not considering? Crow himself went on to suggest that U=5 is not at
all unreasonable. Beyond this, much of what was once thought to be
"junk" DNA is now being found to have selectable function.
> > http://www.genetics.org/cgi/content/full/156/1/297
> >
> >
> >>And as I expected, you just ignored my questions about common descent
> >>and flood geology.
> >
> > Look, I've discussed common descent with you extensively.
>
> No, you've evaded discussing it extensively. Your repertoire has been
> limited to citation of irrelevant papers and a network of facile denials
> without any meat to them. If that's all you have, might as well not talk
> about it at all. You refuse to engage seriously with this question.
That's certainly your opinion. The fact of the matter is, I'm not
interesting in discussing or "evading" this topic with you right now.
If you're not interested in what I'm actually discussing with others
right now, no one is forcing you to be here. But, of course, I've told
you this before and you keep coming back anyway . . . go figure?
> > I've also
> > discussed flood geology with many others extensively. I just don't have
> > enough time to do it all the time. I am currently interested in the
> > problem of neutral gaps and random walks and random sampling. I'm
> > spending what little time I have on that right now. Is that Ok with
> > you? Remember, there are a lot more of you than there are of me. Your
> > personal interests, while certainly interesting to me, are not always
> > my primary interest at the moment.
>
> Our primary interests do not coincide. I'll thank you not to make claims
> that depend on the particular unwarranted assumptions that fall into my
> area of interest. Stop claiming, for example, that your "neutral gaps"
> mean that common descent isn't possible.
Large neutral gaps do in fact make common descent impossible - except
of you invoke ID along the way. I suppose you can invoke ID to help
explain the common descent notion of yours (though I doubt you really
want to do that), but my neutral gaps argument is against the notion of
*mindless* evolutionary processes creating what we see in higher-level
systems of function. I'm arguing that ID had to have been involved
either in the beginning or along the way. Mindless evolutionary
mechanisms simply aren't high-level enough to explain what we see in
all living things.
Sean Pitman
www.DetectingDesign.com
CreateThis
> ... if you can't present anything besides ID to explain a given
> phenomenon, what makes to think that you have to search for a non-ID
> explanation until you find one?
Because ID is just another word for "beats me". You haven't found
anything with ID; you've given up the search.
CT
John Harshman
> John Harshman wrote:
>
>
>>>What do you think the following means?:
>>>
>>>"Using conservative calculations of the proportion of the genome
>>>subject to purifying selection, we estimate that the genomic
>>>deleterious mutation rate (U) is *at least 3* . . . In fact, this range
>>>is likely to be biased downward because we have considered only
>>>nonsynonymous sites as potential targets for deleterious mutations."
>>>[emphasis added]
>>
>>Not much that's relevant to the question I asked. If that's the best you
>>have, it would be wise to stop making the claim.
>
> It is very relevant to your question. Why would you think otherwise if
> the authors themselves suggest that their numbers are biased downward
> by this? - That there are probably more potential targets that they are
> not considering? Crow himself went on to suggest that U=5 is not at
> all unreasonable. Beyond this, much of what was once thought to be
> "junk" DNA is now being found to have selectable function.
I asked for evidence, not statements. If you think a mere statement (in
passing at that) made by a scientist counts as evidence, you are
mistaken. (I will have to admit that many creationists seem to think
this way, if they're allowed to pick the statements.)
>>>http://www.genetics.org/cgi/content/full/156/1/297
>>>
>>>
>>>
>>>>And as I expected, you just ignored my questions about common descent
>>>>and flood geology.
>>>
>>>Look, I've discussed common descent with you extensively.
>>
>>No, you've evaded discussing it extensively. Your repertoire has been
>>limited to citation of irrelevant papers and a network of facile denials
>>without any meat to them. If that's all you have, might as well not talk
>>about it at all. You refuse to engage seriously with this question.
>
> That's certainly your opinion. The fact of the matter is, I'm not
> interesting in discussing or "evading" this topic with you right now.
> If you're not interested in what I'm actually discussing with others
> right now, no one is forcing you to be here. But, of course, I've told
> you this before and you keep coming back anyway . . . go figure?
Because you keep making claims that your favorite theory disproves
common descent, and because you make other claims that depend on matters
that I'm interested in. It's mostly your conflation of independent
concepts that brings me back, in an attempt to correct your misconception.
In the immediate case, I'm just trying to get a handle on your notion
that a much larger fraction of the genome is under selection than we
currently think.
>>>I've also
>>>discussed flood geology with many others extensively. I just don't have
>>>enough time to do it all the time. I am currently interested in the
>>>problem of neutral gaps and random walks and random sampling. I'm
>>>spending what little time I have on that right now. Is that Ok with
>>>you? Remember, there are a lot more of you than there are of me. Your
>>>personal interests, while certainly interesting to me, are not always
>>>my primary interest at the moment.
>>
>>Our primary interests do not coincide. I'll thank you not to make claims
>>that depend on the particular unwarranted assumptions that fall into my
>>area of interest. Stop claiming, for example, that your "neutral gaps"
>>mean that common descent isn't possible.
>
> Large neutral gaps do in fact make common descent impossible - except
> of you invoke ID along the way.
That may be. It's your attempt to redefine common descent as "common
descent under known natural processes only" that gets you into trouble
here. You may or may not have to invoke ID, and we could argue about
that. But my main point is that the presence or absence of ID says
nothing about the presence or absence of common descent. Thus your
neutral gap theory gives you no reason to reject common descent.
> I suppose you can invoke ID to help
> explain the common descent notion of yours (though I doubt you really
> want to do that),
I see no reason to. But the point is that it's irrelevant. I don't have
to invoke any mechanism, natural, supernatural, or whatever, in order to
talk about common descent itself.
> but my neutral gaps argument is against the notion of
> *mindless* evolutionary processes creating what we see in higher-level
> systems of function. I'm arguing that ID had to have been involved
> either in the beginning or along the way. Mindless evolutionary
> mechanisms simply aren't high-level enough to explain what we see in
> all living things.
If you remembered that's what you are talking about, we would not be
having this particular argument. So, are you prepared to accept that
there is good evidence for common descent, now that we're agreed that
the evidence applies whether the mechanism involved is natural or
supernatural?
Glenn
"Seanpit" <seanpi...@naturalselection.0catch.com> wrote in message
news:1134501764.9...@z14g2000cwz.googlegroups.com...
> John Harshman wrote:
>
> > > What do you think the following means?:
> > >
> > > "Using conservative calculations of the proportion of the genome
> > > subject to purifying selection, we estimate that the genomic
> > > deleterious mutation rate (U) is *at least 3* . . . In fact,
this range
> > > is likely to be biased downward because we have considered only
> > > nonsynonymous sites as potential targets for deleterious
mutations."
> > > [emphasis added]
> >
> > Not much that's relevant to the question I asked. If that's the
best you
> > have, it would be wise to stop making the claim.
>
> It is very relevant to your question. Why would you think otherwise
if
> the authors themselves suggest that their numbers are biased
downward
> by this? - That there are probably more potential targets that they
are
> not considering? Crow himself went on to suggest that U=5 is not at
> all unreasonable. Beyond this, much of what was once thought to be
> "junk" DNA is now being found to have selectable function.
>
Will this help?
http://www.entelechon.com/index.php?lang=eng&id=biotechheadlines&blogid=2005-11-18
"This means, according to his calculations, that around fifty percent
of the non-coding sequences are subject to stronger negative selection
than the coding sequences. Furthermore, he assessed that many of the
nucleotide exchanges taking place in the non-coding stretches are
influenced by positive selection. From this follows that a big part of
the non-translated genome is subjected to positive selection and
adaptive evolution, at least in this fly species, and thus has to be
of functional importance. This underlines the relevance of this
regulatory sequences for evolution."
John Harshman
Slightly. It seems garbled and one would have to go back to the original
Nature paper being discussed in order to figure out what the author
really says. But we might interpret it to mean that most (or perhaps
half, which is still quite a bit) of the Drosophila genome is conserved
by selection. However, note that the author (Peter Andolfatto)
discovered this by comparing genomes of two species of Drosophila. If
you compare the human and chimp genomes, the equivalent test, you do not
find anything near that level of sequence conservation, so by that test
the human genome is mostly junk. In fact the web article notes that
difference towards the end.
I bet you would find around the same level of conservation if you
compared the fugu genome to some other puffer. Some genomes have much
less junk than the human genome.
Noone Inparticular
Further on in the same article we see;
"Alexey S. Kondrashov, in his comment on the article published in the
same issue of Nature, constrains Andolfatto's statements and postulates
the existence of two classes of eukaryotic genomes: on the one hand,
there are species, described by Andolfatto using the fruit fly as an
example, whose genomes are kept lean by strong selective pressure and
which contain almost no non-functional parts at all. But on the other
hand, there are other species like Homo sapiens and, more generally
speaking, vertebrates having smaller population sizes and longer
generation times. In these species, negative selection works only
ineffectively and they do accumulate long stretches of non-functional
DNA - which is to say, junk - in their genomes."
Ron O
Name a single instance where your intelligent designer has been found
to be responsible for anything that we can study in nature and I'll
grant you that you can consider the option. 100% failure means what?
ID has never won the lottery. ID has never been a verifiable
explanation. Never means just that, never. All that has ever happened
is that ID retreats to the next untestable level. That isn't
verification that is failure. Why should ID be a viable option in
science when it has a 100% failure rate as an explanation?
>
> > > > Demonstrate that this many pregnancies are required of human females.
> > >
> > > Read the papers Ron. This estimate is not based on my calculations.
> > > Don't take my word for it. I've directly quoted several relevant
> > > sections to these papers and you still don't understand. Look again at
> > > the following quote:
> >
> > I would bother to read the papers, but it would be a waste of time
> > since I am not currently interested in that line of research.
>
> Then you have nothing to say here now do you? You haven't read the
> papers and you're not about to read them because you are not interested
> in them. Yet, somehow you feel qualified to tell me I'm wrong when I
> only present what the authors of these papers are saying themselves.
This is so brain dead that you should be embarrassed. I probably have
read more population genetics papers than you will ever read. Heck, I
majored in genetics. Why do you think that I can cut through your crap
and tell you why your bogus junk doesn't matter?
>
> >Why
> > would it be a waste of time, because it doesn't matter if I read them
> > or not. Your arguments are bogus and you know it.
>
> LOL - That's just too funny Ron. These particular arguments, arguments
> that you are calling "bogus", such as the high required birth rates,
> are not mine. The authors themselves present them.
Not funny just true. Demonstrate otherwise.
>
> > The simple reason
> > that I can say that is that you will not put forward your own junk for
> > evaluation.
>
> What?
Put your own evidence that you claim is better than the evidence for
common descent. You claim to have it, but you never put it forward for
evaluation.
>
> > The reason why you will not do this is because you know
> > that your junk is worse off than the real science that you try to
> > obfuscate and don't like.
>
> I'm putting forward what the authors themselves are saying Ron.
This is just real science that you can obfuscate and bullshit about.
What you need to do is put up the evidence that you claim that you have
that supports your beliefs that is better than the evidence that you
claim isn't good enough for common descent. Simple, but you are just
being stupid about it. It is called denial, and in cases like yours it
seems to be pathological.
>
> > Demonstrate that it is worth going to read the papers. Put up your
> > evidence for your beliefs that is as good as the evidence for common
> > descent that you claim isn't good enough.
> >
> > That is simple and direct, since you have claimed to have such
> > evidence, all you have to do is put it forward.
>
> You're one of the biggest self-deluded people here if you keep talking
> this way. You think "I don't have to read the evidence you're talking
> about because I already know that whatever you say is already wrong. No
> need to look into it any further."
Just prove it. Put up your evidence. Let's see it. Who is deluded?
Look in the mirror. Why can't you do something as simple as is being
requested? Don't you think that there is something mentally wrong with
a person such as yourself. Really reread this series of posts and
pretend that it is someone like Pagano making your lame excuses.
>
> > > "The reduction in fitness (i.e., the genetic load) due to
> > > deleterious mutations with multiplicative effects is given by 1 - e^-U
> > > (Kimura and Moruyama 1966). For U = 3, the average fitness is reduced
> > > to 0.05, or put differently, each female would need to produce 40
> > > offspring for 2 to survive and maintain the population at constant
> > > size. This assumes that all mortality is due to selection and so the
> > > actual number of offspring required to maintain a constant population
> > > size is probably higher." - Nuchman, Michael W., Crowell, Susan L.,
> > > Estimate of the Mutation Rate per Nucleotide in Humans, Genetics,
> > > September 2000, 156: 297-304 (
> > > http://www.genetics.org/cgi/content/full/156/1/297? )
> >
> > Bumble bees can't fly because some guy calculated that it was
> > impossible.
>
> That was for a fixed-wing model. This is an urban legend Ron.
>
> Do you actually disagree with the numbers listed by these authors? If
> so, upon what basis do you disagree?
Urban legend? What does the data say? It says that there is something
interesting that we have to figure out. It doesn't say that life is
impossible or that all species are heading to extinction. We already
know that over 99% of the species that have ever existed didn't make
it. You obviously are not looking at the average to determine what
populations or subpopulations are going to give rise to the next
species.
What does the data tell you? It tells you that no matter what you
claim the species that make it are able to get rid of the detrimental
mutations for the simple reason that we can see the evidence in their
DNA. Explain the 3:1 ratio. It is that simple. The replacement
subsitutions once existed but they were removed by some mechanism.
This is consistent across all the species that I've seen data on.
So it looks like you are claiming bumble bees can't fly when you know
that they can.
>
> > The fact is that we see bumble bees flying and we observe
> > the after effects of removal of all these detrimental mutations. How
> > do you remove them all in just the few thousand years since the ark?
> > You can't just use some lame argument if it means that your own
> > explanation is bogus.
>
> You can remove many detrimental mutations in a relatively short time.
> That's not the problem. The problem is restoring the level of maximum
> fitness in a population to the original level of maximum fitness that
> the previous generation had. That's a whole different problem. That
> problem is what creates the notion of the "Living Dead".
Restoring what? How do you know that they ever lose it? And if they
do lose it how do you know that they aren't just going to go extinct
like greater than 99% of the previous species.
Not all of them because it is obviously a continuous process that is
still going on. We can observe the intermediate stages in any
population that you want to study.
Mutations happen constantly. No big whoop.
>
> < snip >
>
> > > > > You're wrong as far as the problem at hand is concerned. Selective
> > > > > breeding of only the fittest with the fittest is one way to slow down
> > > > > the decline. Random breeding or inbreeding, only speeds up the overall
> > > > > decline of a population and hastens either extinction or a dramatic
> > > > > increase in reproductive rates to overcome the problem.
> > > >
> > > > Where did you read this, cite the paper. I've never seen this
> > > > conclusion, ever. Just think for a moment, you can't
> > > > determine the fittest unless you inbreed.
> > >
> > > http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
> >
> > Bullshit. Put up the quote. Really I'm not going to bother to go read
> > anything that you claim because it isn't worth the time. Until you can
> > demostrate that you can actually reason, why should I? Heck, I only
> > respond to your quackery when you respond to one of my posts, your
> > posts aren't even worth reading.
>
> The entire paper in the reference is about this problem Ron. If you
> won't read it, why should I bother. You've been wrong on all of your
> other challenges in this thread so far Ron. You were wrong about the
> minimum birth requirements not going into the hundreds. You also
> called that number "BS" if I recall. Just read the paper Ron. It isn't
> that long of a paper or all that difficult to read.
The entire paper. Gee, you know what that means, it means that if I
tilt my monitor at the right angle and read it with the right rose
colored glasses I might get the same meaning out of the paper that you
do.
I stand by my statement that I have never seen such a conclusion in a
genetics paper on this topic.
>
> I've never seen someone who claims to be a scientist argue that someone
> else is wrong about what a paper says while refusing to actually read
> the paper in question.
Because I know how mentally incompetent you are. No one should trust
anything that you conclude from a paper. Look at how you avoid
demonstrating that you can be rational. Why not present your evidence
for your beliefs? If you really had some, you would present it.
Pretendiing like you are is just pathological.
>
> > What is your evidence for your beliefs and why is that bogus evidence
> > better than what science has? If you can't do that why should anyone
> > listen to you?
>
> I've just given you the reference Ron - a reference to a real article
> in a real scientific journal. What more can I do if you won't read it?
You know what I am talking about. The evidence for your beliefs that
you claim is better than the evidence for common descent that you claim
isn't good enough. You know, evidence for your version of what
happened. This act is just pathetic.
>
>
> "You can take a horse to water . . . "
If you only understood what you write yourself.
>
> > > > > I don't shrug such numbers off. There certainly are far less specified
> > > > > functions in low level sequence space. Some functions may have ratios
> > > > > as low as 1 in 5 or 1 in 2 or even 1 in 1. However, these are not what
> > > > > I would call "fairly specified" functions. And, there certainly are
> > > > > functions that do indeed show ratios lower than 1 in 1e30 per 100aa.
> > > > > That's the whole point. This level of minimum functional complexity is
> > > > > much higher than the levels you are talking about - which are even
> > > > > easier to evolve.
> > > >
> > > > What is unspecified about new enzymatic function? Demonstrate that the
> > > > functions that you are interested in are more "specified" than the
> > > > antibody functions. You are out of the ball park by 18 orders of
> > > > magnitude and you still claim that you can make viable conclusions?
> > > > How?
> > >
> > > You need to read the papers by Sauer et al., which I've referenced
> > > several times recently in this forum, which describe protein-based
> > > functions with ratios as low as 1 in 1e62 per 100aa. Estimated ratios
> > > for other types of protein-based functions, like cytochrome c are at
> > > least as low as 1 in 1e30 per 100aa (most recent Yockey estimate).
> > >
> > > John F. Reidhaar-Olson and Robert T. Sauer, "Functionally Acceptable
> > > Substitutions in Two [alpha]- helical Regions of [lambda] Repressor,"
> > > Proteins: Structure, Function, and Genetics, 7:315, 1990 p. 315
> >
> > Who the heck cares? No matter what, you stil have to figure out how
> > antibodies work if these guys estimates mean anything.
>
> They only work at low levels of functional complexity.
Define low levels and differentiate them from what you claim to be high
levels. What is low level about specific binding of an antigen. Is it
the number of amino acids involved. The antigenic residues can be
spread across the surface of the antigen, they don't have to be
consecuative. The number of antibody residues that interact with the
antigen is probably just as high as any enzymatic function.
>
> > Really, we have
> > reality and it conflicts with these estimates, what would you go with,
> > your bogus inferences from these estimates or reality? How do
> > antibodies work if they only have a maximum of 10e12 sequences tested?
> > Why should anyone believe what you claim about some 10e62 claim when it
> > seems to be 50 orders of magnitude wrong?
>
> The functions that these antibody tests have discovered have all been
> lower level functions having much higher ratios in sequence space.
> None of these discovered functions have been of the types discussed in
> the papers written by Yockey and Sauer. If you think this is wrong,
> why hasn't anyone challenged the estimates of Yockey and Sauer in the
> literature? If you think this is wrong, then please do provide your
> own references.
You keep repeating that, but I haven't seen you demonstrate that "lower
level" is any different from "higher level." Make the distinction or
admit that you can't.
So what. Who says that you can't get the same function using two
peptides, or a larger protein? You are still orders of magnitude off
and you can't deal with it. Just look at your other junk, you claim
that a 2 fold difference is significant and you can't explain 18 orders
of magnitude, but you still claim that you must be right. How bogus is
that?
>
> > What do you think "degeneracy" means in terms
> > of the antibody example? What about matching another DNA sequence that
> > would have worked just as well as a repressor site? How many of those
> > are around if they all have around a 10^57 degeneracy for proteins that
> > would interact with them?
>
> That doesn't matter. What matters is the ratio for this particular
> function.
Who cares? Explain how antibodies work. What is that ratio?
>
> Read the papers Ron. You'll be better off.
Deal with reality and you'd be better off. Got any evidence that you
can deal with this topic rationally? I've never seen any. Present the
evidence that you have been pathetically dodging. Evaluate it with the
same level of scrutiny and what do you come up with?
I don't have to read these genetics papers to tell that you are just a
mental case, that should get some help. Really, just get someone that
you trust to walk you through your denial, and inablitity to deal with
reality. Either that or put up your evidence and show us how good it
is compared to the science that you don't like. Why do you think that
you have to dodge presenting the evidence that you claim that you have?
I didn't make the claim, you did. You said that you had evidence that
was better than the evidence for common descent that you claimed isn't
good enough. So what is your idea of what happened and what is the
evidence backing it up?
This so simple that even you should be able to do it, so why do you
avoid doing it?
Get some help. I only respond to your posts if you respond to one of
mine. You know that I don't chase you around the net. I think that
you have a problem. I can't help you deal with it. Someone that you
trust has to help you.
Ron Okimoto
Glenn
"John Harshman" <jharshman....@pacbell.net> wrote in message
news:_LFmf.39290$6e1....@newssvr14.news.prodigy.com...
> Seanpit wrote:
>
>
> > In fact, the odds are most likely a
> > whole lot higher than U=3 now that a lot more of what was thought
to be
> > "junk" DNA, because it doesn't code for proteins and whatnot, is
no
> > longer junk, but functional and constrained by natural selection
after
> > all.
>
> I could use some evidence for the assertion that the proportion of
> sequences under selection is much higher than previously thought.
Your
> references don't seem to provide any.
>
"...We try here to evaluate the amount of sequences in the human
genome that are under selective pressure to be transcribed. We built a
prediction tool based on the generalized linear model, using
transposable elements densities and other sequence compositional
variables. We show that these features are informative enough to
predict whether a sequence is transcribed or not, and - if
transcribed - in which orientation. We estimate that functional
transcripts constitute at least 50% of the genome, and that about one
third of these transcripts do not encode proteins."
Richard Forrest
Seanpit wrote:
<snipped>
> Look, I've discussed common descent with you extensively. I've also
> discussed flood geology with many others extensively.
Sean, you haven't.
You have made assertions which have been shown to be false, refused to
read the references which demonstrate that, refused to make any changes
to the content of your essay on your web site even when it has been
shown to contain demonstrable falsehoods, and ignored or dismissed out
of hand any criticisms of that site.
This is not "discussing" flood geology. It is simply a blanket denial
of any contrary evidence.
RF
Ron O
This is something different from what Sean is talking about. This is
selection against transposable elements and transcriptional units, like
retrovirus, insertion into regions that can influence transcription.
These types of events can insert in "junk" DNA with no function except
to hold the gene together and disrupt transcription of that gene. A
lot of this DNA can be totally deleted and you still get functional
genes. Just look at the Fugu. Around 90% of this sequence has been
deleted and the genes that are the same between fugu and humans still
function. In fact fugu has a lot fewer transposable elements probably
because it has also gotten rid of the sequence that these types of
elements can insert into and get carried around in the genome without
much effect.
What you should be looking into are the genes in fugu that aren't as
short as the DNA difference average. Genes like Duchennes muscular
dystrophy gene in fugu is only 60-70% shorter than the human gene and
they think that it has to maintain a larger size for some selectable
reason, but they haven't figured out why. The gene in humans is over a
million base-pairs in length with the final transcript only 14,000
base-pairs (over 98% of the transcript is cut out and isn't part of the
final functional transcript), but in fugu the gene is around half a
million with a final transcript about the same size as human. It could
be just chance that some genes retain more "junk" than others, but
there could be some selective advantage to some of the sequence.
Ron Okimoto
Glenn
"Ron O" <roki...@cox.net> wrote in message
news:1134565676.6...@g44g2000cwa.googlegroups.com...
(FTUs) are "necessary for the proper functioning of the genome" and
comprise perhaps more than 50% of the human genome. They weren't
looking at Fugu, and I haven't seen a convincing argument for why
removal of part of a genome neccesarily translates into evidence that
the removed part was not under selection - or even unnecessary for
that matter.
Glenn
>
> > > > "This means, according to his calculations, that around fifty
percent
> > of the non-coding sequences are subject to stronger negative
selection
> > than the coding sequences. Furthermore, he assessed that many of
the
> > nucleotide exchanges taking place in the non-coding stretches are
> > influenced by positive selection. From this follows that a big
part of
> > the non-translated genome is subjected to positive selection and
> > adaptive evolution, at least in this fly species, and thus has to
be
> > of functional importance. This underlines the relevance of this
> > regulatory sequences for evolution."
>
original
> Nature paper being discussed in order to figure out what the author
> really says. But we might interpret it to mean that most (or perhaps
> half, which is still quite a bit) of the Drosophila genome is
conserved
> by selection. However, note that the author (Peter Andolfatto)
> discovered this by comparing genomes of two species of Drosophila.
If
> you compare the human and chimp genomes, the equivalent test, you do
not
> find anything near that level of sequence conservation, so by that
test
> the human genome is mostly junk. In fact the web article notes that
> difference towards the end.
>
although any discussion here, like news articles of research papers,
might be garbled themselves.
>
http://www.the-scientist.com/news/20051020/01
"Much of the "junk" DNA in Drosophila shows signs of either negative
or positive selection, according to a study in this week's Nature. An
analysis by Peter Andolfatto of the University of California, San
Diego, reveals that around half of non-coding Drosophila DNA is
evolutionarily constrained and that much of the remaining divergent
DNA has undergone adaptive evolution. Both types of selection show
that "this non-coding DNA actually has functional importance to the
organism," said Andolfatto."
[...]
"According to an accompanying News & Views article by Alexey S.
Kondrashov of the NIH's National Center for Biotechnology Information
in Bethesda, Md., Andolfatto's findings do not completely demolish the
neutral theory of genome evolution. While this theory, which says that
most non-coding DNA is functionless and ignored by evolution, may not
be relevant to Drosophila, it still works for mammals and other
vertebrates, Kondrashov writes."
>
"It could well be that large parts of the mammalian genome are truly
junk" and therefore undergo neutral evolution, Andolfatto said.
Mammalian non-coding DNA may contain positively selected sites, he
said, but, "if your positively selected sites are buried in junk,
they're more difficult to find."
>
"Since most animals possess far more non-coding than coding DNA, even
a small amount of non-coding adaptation may contribute greatly to
genome function, Andolfatto said. "The signature of positive selection
is weaker in non-coding DNA, but there's much more non-coding DNA than
coding DNA," Andolfatto said. "So if you sum over the whole genome.you
realize that perhaps protein evolution isn't so important after all."
>
It appears to me that Andolfatto is not convinced large parts of the
mamallian genome are junk, that do not posess sequences under positive
selection. "Difficult to find" seems to indicate that he has some
doubts about other's inferences to the contrary. The argument that
removing parts of DNA and seeing no ill effects provides some evidence
that those parts were not under selective pressure is not convincing
to me. Are there others?
>
snip
John Harshman
If it's necessary, how have fugu managed to get along without it? And
how could they have evolved so as to lose it?
John Harshman
I don't think so. If there are a few positively-selected sites buried in
a lot of junk, they can't compose a large part of the genome. Hard to
find in this case means rare. If they weren't, they wouldn't be hard to
find.
> The argument that
> removing parts of DNA and seeing no ill effects provides some evidence
> that those parts were not under selective pressure is not convincing
> to me. Are there others?
Yes. Andolfatto's findings are based on testing rates of divergence
between species. If the rate is much slower than the neutral rate, he
supposes that stabilizing selection accounts for this. If it's much
higher, he supposes that positive selection accounts for this. So if you
accept his conclusions you need to accept his methods. If you try these
methods on the human genome you find no such departures from neutrality.
(The news article implies that his tests are more sensitive than
previous ones, but on the surface it looks like a form of the HKA test,
and Andolfatto was a student of the K in that test, so I wonder. HKA and
similar tests have certainly been performed on human sequences.)
Let's not confuse "junk DNA" with "non-coding DNA", though. It's been
known for a long time that many non-coding sequences are functional.
They just do not, in the human genome at least, compose a large fraction
of the whole. That's based on the same sorts of test that Andolfatto used.
I can think of a number of potential objections, though I don't know if
he already thought of them. If recombination rates are low enough,
selection on coding or other regions under obvious selection can
influence frequencies at sites far from the directly affected site. It's
also possible that the sites he used to estimate the neutral rate are
themselves under positive selection, though I have trouble seeing how
that could be. However, it's more likely that the sites used to estimate
that rate are under weak negative selection, and that might explain some
of the positive selection results; i.e. those are really the neutral
sites, and his clock is slow. There are previous papers that show silent
sites in exons to be under weak negative selection.
Interesting result, though, and I wonder what it means. How are the
conserved sequences distributed, and what do they do? Ditto for the
positively selected ones.
Seanpit
> Sean, you haven't.
> You have made assertions which have been shown to be false, refused to
> read the references which demonstrate that, refused to make any changes
> to the content of your essay on your web site even when it has been
> shown to contain demonstrable falsehoods, and ignored or dismissed out
> of hand any criticisms of that site.
>
> This is not "discussing" flood geology. It is simply a blanket denial
> of any contrary evidence.
>
> RF
Again, that is certainly your opinion. I find it most interesting then
that you continue to want to talk to someone who you think incapable of
true "discussion" ; )
Regardless, I'm not interested in talking to you about your notions
about what you think I've gotten wrong on my website. So far I haven't
seen anything from you that I find personally convincing. That doesn't
mean you're wrong or that many if not most others find you very
convincing - but I don't. Beyond that, I'm busy with other topics
right now. I just don't have to time to discuss everything with
everybody who wants to discuss his or her own favorite topic with me.
Until I am, no one is forcing you to be here. You can always go
elsewhere and join some other thread. You're certainly not contributing
much of anything to the topic at hand in this thread.
Sean Pitman
www.DetectingDesign.com
John Harshman
> Richard Forrest wrote:
>
>
>>Sean, you haven't.
>>You have made assertions which have been shown to be false, refused to
>>read the references which demonstrate that, refused to make any changes
>>to the content of your essay on your web site even when it has been
>>shown to contain demonstrable falsehoods, and ignored or dismissed out
>>of hand any criticisms of that site.
>>
>>This is not "discussing" flood geology. It is simply a blanket denial
>>of any contrary evidence.
>
> Again, that is certainly your opinion. I find it most interesting then
> that you continue to want to talk to someone who you think incapable of
> true "discussion" ; )
Nobody says you are incapable, just that you haven't done it so far. We
retain hopes that some day you will.
> Regardless, I'm not interested in talking to you about your notions
> about what you think I've gotten wrong on my website.
You aren't concerned with getting your claims right? That seems odd.
[snip]
Ron O
They aren't needed in the way that you think. They can be disrupted by
transposable elements and retroviruses because these elements have
their own transcriptional control signals that can mess up
transcriptional units. You can mess up a transcriptional unit like the
greater than million base-pair muscular dystrophy gene by putting a
retrovirus with alternate splice sites just about anywhere in the huge
introns and it will disrupt the transcriptional unit. That a lot of
sequence isn't really needed is a simple inference from the fact that
many of these genes do the exact same thing in the various species.
Enzymes like glyceraldehyde 3 phosphate (house keeping genes) are used
in things like glucose metabolism in eukaryotes and prokaryotes in all
cells. Fugu has removed quite a lot from the transcriptional units.
As the paper tells you coding sequence is only a small fraction of
transcriptional units. Some noncoding sequence that is transcribed is
regulatory, but species like fugu have gotten rid of just about
everything that they can. This turns out to be around 90% of the DNA
compared to humans.
Ron Okimoto
Seanpit
RobinGoodfellow wrote:
< snip >
> You seem to give a lot of weight to the fact that N&C seem to think
> that the deleterious mutation rate is too high, and not at all to the
> fact that they seem to think that humans and chimps diverged from a
> common ancestor - even though they use the latter as a premise to
> arrive at the former conclusion (the Ka/Ks ratios). Not too big on
> consistency, are we?
If U really is greater than 3, all slowly producing animals, to include
both humans and apes, are headed for extinction. If U is significantly
less than 3, then there is something wrong with the assumptions of
human-ape common ancestry. Either way, something's not adding up. How
is this consistency from your own side?
Beyond this, there is good real time evidence that humans do indeed
sustain upwards of at least 175 mutations per person per generation.
Oh, and by the way, my use of "29 mitotic divisions" before
reproduction of the next generation is a male to female *average*. The
average for woman is about 23, while the average for men is a good bit
higher, creating an overall average of about "29". Was this really all
that hard to understand from the way I wrote my essay?
Now, if each human individual sustains 175 mutations per generation,
how many of these will hit the "functional" area of the genome?
Consider the following discussion by Fay et al published in the July
2001 issue of Genetics:
"The genomic deleterious mutation rate in humans was previously
estimated to be at least 1.6 on the basis of an estimate that 38% of
amino acid mutations are deleterious. The genomic deleterious mutation
rate is likely much larger given our estimate that 80% of amino acid
mutations are deleterious and given that it does not include
deleterious mutations in noncoding regions, which may be quite common
(SHABALINA and KONDRASHOV 1999). The combined NC*/S* ratio of common
SNPs from both surveys indicates 50% of the noncoding sites are
constrained and must serve some function. Because an equal number of
noncoding and amino acid-altering sites were surveyed, noncoding
mutation should contribute at least 60% (0.50/0.80) as much as coding
mutations to the total genomic deleterious mutation rate."(1)
So, where is your notion that U=1.6 now? These figures put U well over
5. As far as the beneficial mutation rate, Fay et al go on to say:
"The slightly deleterious fraction, f1, cannot exceed 0.80 since 20% of
mutations were estimated to be neutral . . . Extrapolating this
proportion to the total amount of coding DNA in the genome (~5 x 10^7
bp) yields an estimate of up to 1 advantageous substitution every ~200
years since humans separated from old world monkeys 30 million years
ago."
And you think that is an adequate enough fixation rate to compensate
for a U of over 5?
1. Justin C. Fay, Gerald J. Wyckoff, and Chung-I Wu, Positive and
Negative Selection on the Human Genome, Genetics, Vol. 158, 1227-1234,
July 2001 -http://www.genetics.org/cgi/content/full/158/3/1227
> But, to answer your point, N&C had no way to assess the impact of
> non-coding DNA on U. They could only compute U for coding DNA, and
> those estimates vary considerably depending on their different
> estimates of overall mutation rates.
Ah, but N&C did comment that their estimate of U was likely biased
downward because of this problem. You just choose to overlook that?
> N&C's lower bound estimate of U
> is 1.5, and they site another estimate of 1.6 (from 1999, By
> Eire-Walker and Keightley) that agrees with that lower bound. I'm not
> saying that is necessarily the correct estimate either, but your desire
> to accept U>=3 suggests that you've already made up your that we are
> devolving towards extinction, and are cherry-picking data to support
> this position.
I'm not cherry picking to support my position. I'm just stating what
the author's themselves believed to be the most likely estimate for U.
They are the ones who said that U is most likely greater than 3, even
greater than 5, not me. Rather, it seems like you are the one who is
"cherry-picking" the extremes of the author's data set in an effort to
support your views. The most recent data, according to the authors,
not me, clearly supports U values of well over 5.
> > In fact, the odds are most likely a
> > whole lot higher than U=3 now that a lot more of what was thought to be
> > "junk" DNA, because it doesn't code for proteins and whatnot, is no
> > longer junk, but functional and constrained by natural selection after
> > all. This has lead to the suggestion that U is actually "greater than
> > 5" (see references below).
>
> The paper you reference by Rice you are referring to simply mentions
> this figure in passing, citing three other studies, the latest of which
> came of 1993. This pre-dates the U=1.6 estimate of Eire-Walker and
> Keightley by 6 years, and N&C's estimate of U=3 by 7 years.
And why do you think Rice chose to print the estimate of U = 5? Again,
what do you think about the Fay paper will estimates of well over 5?
> Of course,
> the ironic part is that Rice's paper proposes specfically how epistasis
> effects can compensate for such high deleterious mutation rates.
Rice does point out several ways to slow down the effects of
deleterious mutation rates, but he doesn't show how a high U, like a U
above 3, can really be turned around. He does list off several kinds
of epistasis as well as the positive effects of biased breeding, but
none of these really compensate for a detrimental mutational load of U
greater than 3. The net effect, even with buffering epistasis combined
with biased mating is not enough to counteract a very high detrimental
mutation rate given a low reproductive rate.
For example, take a steady state population of 1 billion, half women,
half men, with a reproductive rate of 10 children per woman and U of 3.
All of these people in this initial population start out with equally
good genomes. In other words, all of them are at "R-best". Now, each
woman has 10 pregnancies to populate the next generation. Out of these
5 billion 1 in 20 are still at R-best. The rest have at least 1
detrimental mutation with an average of 3 detrimental mutations. That
means the next generation will start out with at most 250 million
R-best individuals. The rest will be R-not as good. Even if these 250
million R-best mate exclusively with each other, regardless of the
presence of buffering epistasis or not, the number of R-best in the
next generation will be a mere 62 million in a steady state population
of 1 billion.
How is this decline reversed without dramatically increasing the
reproductive rate?
> > Your notion of U is in reality as low as 1.5 is highly unlikely.
>
> It's not my notion at all. I am arguing that in order for your idea to
> have some merit, you need to be able to shown not only that U>=3 (Which
> is by no means clear),
It is quite clear that most on your own side feel that U is well over 3
- even over 5.
> but that epistasis effects aren't real or
> sufficient, beneficial mutations can't maintain a an overall level of
> fitness, deleterious effects are not, on average, so small as to have
> extremely little effect on the overall fitness of the population in
> practical terms, and some other things besides. I am certainly willing
> to beleive that U>=3, but a) it is far from certain, and b) it hardly
> consitutes evidence that we are devolving toward extinction.
Sure it does. If U is over 3 and the beneficial mutation rate isn't
anywhere near what you say it is, what's left to reverse the trend and
make more R-bests in each generation than are lost as a ratio of the
current vs. the previous number of R-bests?
Not only is the beneficial rate nowhere near as high as you say, even
if it were 3 in 100 mutations, how would this help? Even if every one
of these 3 beneficial mutations became fixed, you'd have to get rid of
100 detrimental ones for each of these beneficial mutations to tilt the
balance. How would you do that? The only way I can think of is by
biased breeding and the use of genetic translocations that somehow know
just what to trade for what so that the good mutations could be
gathered into clusters and the bad mutations could be gathered into
clusters for batch elimination in greater quantities than they enter
the gene pool. Sure, epistasis can eliminate great quantities of bad
mutations all at once, far more than 3 at a time. The problem is that
the large numbers that are eliminated by epistasis represent build up
over many generations. Epistasis may remove 500 bad mutations via
taking out the most unfit members of a population, but these
individuals gained these mutations over many generations. Their
removal prior to reproduction does not remove bad mutations faster than
they are entering the gene pool. It this does not reduce the decline
in relative numbers of R-best over time.
In order to get more R-best, a population must somehow batch good and
bad mutations via some sort of translocation mechanism that
concentrates bad mutations into a few individuals and then gets those
individuals to reproduce and then concentrate their combined badness
into just one or two of their 10 kids. Hmmmmm . . . Doesn't sound
likely to me.
> > I have read several of Kimura's papers as well as several papers from
> > others dealing with this problem by suggesting some form of epistasis
> > (multiplicative increases in the effects of detrimental mutations). The
> > problem is that with detrimental mutation rates as high as 3 per person
> > per generation, positive epistasis only increases the death rate, but
> > does not clear detrimental mutations from the gene pool faster than the
> > fitness of the most fit individual in the population decreases.
>
> Apparently, you haven't read enough papers. What you describe is
> reinforcing epistasis, which is only one several possible forms of
> positive epistasis. Rice, for instance also discusses the possible
> effects of buffering and pathway epistasis, both of which can have a
> significant effect on reducing genetic load in sexually reproducing
> populations according to his simulations.
Yes, but his simulations make some very interesting assumptions, such
as biased mating and biased translocations. How does this happen in
the real world?
> > Of course, sexual reproduction is supposed to help out with this problem,
> > but sexual reproduction doesn't even help unless mating is done in a
> > non-random way.
>
> And of course, in real life, mating happens in a purely random fashion.
Do you know how many detrimental mutations your wife has? Does your
wife know how many you have?
> I don't know if you're married, but if you are, I hope you didn't just
> ask the first girl you saw on the street to be your wife - and if you
> did, I hope for both your sakes she didn't say "yes".
My wife is downright gorgeous, and looks quite "fit" to me. But, this
doesn't mean she has no more detrimental mutations than her parents.
The odds are that she has at least 3 detrimental mutations that they
did not have. The same is true for me. That's just the way the cookie
crumbles. And, our kids will most likely have 3 more detrimental
mutations than we have . . . and so on.
> More seriously,
> don't you think that an organism suffering many deleterious mutations
> will have much lower chances of mating successfully, especially with
> organisms of the "most fit" genotype?
Not necessarily. Studies have shown that even with a marked increase
in detrimental mutations a population can continue to grow for some
time. The odds are very good that many in the R-fittest group will
actually mate with those of less fit groups. There just aren't very
many R-fittest to go around and some of those less fit babes look fit
enough - you know?
> There is fierce competition for
> mates in both the human and the larger animal kingdom: as the result,
> genetically fit individuals have a higher propensity of mating with one
> another, and individuals of lower fitness are forced to mate with one
> another, assuming they can mate successfully at all. This phenomenon
> is called positive assortive mating, and Rice shows that it can
> drastically reduce mutational load by eliminating whole swaths of
> deleterious mutations at once, even in the absence of all other
> epistaisis effects.
I'm not saying that biased mating does not occur. I'm just saying that
it isn't nearly complete and that even if it was, it really wouldn't
solve the problem. What you need is biased mating and biased
translocations as well.
> > Of course, there are several problems with this non-random mating
> > notion. With a detrimental mutation rate of 3 per individual per
> > generation, only 2 of 40 individuals will have a neutral fitness
> > balance relative to the original parent population
> > What are the odds that these 2 individuals will actually mate "preferentially" so that
> > the overall number of individuals with equivalent parent-level fitness
> > does not decrease in each generation?
>
> Non-random mating does not mean "only best fit individuals mate with
> each other". In most fit individuals do not derive any benefits from
> recombination or epistatic effects, since they do not suffer any
> deleterious mutations in the first place, relative to the rest of the
> population. Non-random mating could mean any number of things,
> including the positive assortative mating scenario described above:
> that is, individuals within the same fitness classes tend to mate with
> one another.
There's a lot of mixing and matching you know. The fact is that we are
only concerned about the R-fittest class here. How do we get more
R-fittest in each generation? You don't do that by mating your
R-fittest with those that are less than R-fittest.
> > In every generation every individual receives around 3 deleterious
> > mutations. The "most fit" phenotype in each generation comes in at a
> > ratio of about 1 in 20. With that ratio, unless very fortuitous mating
> > and high reproductive rates take place, the direction is still
> > extinction. There is no equilibrium here. The only way to avoid this
> > problem is by having the most fit individuals mate pretty much
> > exclusively with each other and still make a whole lot of children in
> > each generation.
>
> I will concede this point, more or less.
This is the whole point! This is where the rubber meets the road!
> My argument was that unlike
> in the infinite allele/infinite pool of deleterious mutations models
> that are used to derive genetic loads, the potential for deleterious
> mutations in a population will go down once a particular slightly
> deleterious allele becomes fixed - and in fact, the potential for a
> reverse beneficial mutation opens up (more on that below). In other
> words, the deleterious mutation rate is not a constant thing.
U is not a constant thing? Where do you get that idea? Have a
reference for this one? Reverse mutations do happen, but not nearly
fast enough to keep up in slowly reproducing populations.
> However,
> if the number of loci that can undergo such mutations is high enough
> and U decreases sufficiently slowly, a population will likely undergo a
> very significant reduction in fitness - though not necessarily
> extinction - the genetic load approximations start to break down.
Why not extinction?
> > Deleterious mutations happen on an individual basis - more than 3 per
> > individual per generation. These are then passed on to the next
> > generation from parents to children and do not leave the gene pool
> > until they die out of it with the premature death of an individual who
> > didn't get a chance to replicate. They don't need to "propagate" any
> > other way to be a problem. Their rate of propagation is not
> > significantly related to the size of the population, but to the
> > reproductive rate and death rate.
>
> As the population size grows, you will see a broad spectrum of mutants
> with accumulation slightly deleterious mutations, but any one specific
> deleterious mutation at a specific locus will have a much harder time
> achieving fixation in the population. For sexual populations, this is
> where recombination can act to eliminate multiple deleterious mutations
> in fell swoops, especially in the context of non-random mating and
> epistasis.
You don't need to have fixation for deleterious mutations to be a
problem. Each individual sustains 3 detrimental mutations. The next
generation will have 6, the next 9, etc. None of these needs to become
fixed for the overall fitness level of the population to decline. How
is recombination going to help? How does random recombination know how
to cluster the bad together preferentially for batch elimination? It
is quite clear that if such clustering could actually be realized that
non-random mating and epistasis would most certainly take care of the
rest, but I don't understand how recombination is going to help here.
> > > whereas beneficial mutaitons will arise and
> > > spread much more rapidly (this is one of the basic theorems of
> > > population genetics).
> >
> > Not quite true.
>
> Ah, Quote-Mining, the discerning creationist's way of doing science
> (TM)! None of those quotes actually say what you want them to say when
> placed in proper context, Sean. Let's examine closer, shall we?
Lets . . .
> The first paper, "Beneficial Mutations, Hitchhiking and the Evolution
> of Mutation Rates in Sexual Populations" by Toby Johnson (Genetics,
> Vol. 151, 1621-1631, April 1999), says nothing about the fixation rates
> of beneficial mutations, but rather their effect on modifiers of
> mutation rate. They concude that, according to their model,
> deleterious mutations have more of effect on the overall mutation rate
> due to stronger indirect selection of hitckhiking rate modifiers. It
> has nothing to do with whether or not beneficial mutations fix faster
> in a population or have more of an effect on the overall fitness of the
> population compared to their deleterious counterparts.
It seems to me that what the authors do say does play into how
beneficial mutations affect fitness vs. detrimental mutations. The
authors write, "Relative to an asexual population, increased levels of
recombination reduce the effects of beneficial mutations more rapidly
than those of deleterious mutations."
Now, I'm not saying that beneficial mutations do not fix faster than
detrimental mutations. Of course they fix faster. I'm just saying that
beneficial mutations are so outnumbered by detrimental mutations that
they just can't keep up. And, this is made even worse by the fact that
recombination reduces the effects of beneficial mutations over and
above what they are able to do in asexual populations.
There's really no quote-mining here. I'm not intending to indicate
anything other than what the authors true intent really is in their
paper. You just misinterpret my intent - or so it seems.
> > "In sexual populations, the combined effect of beneficial and
> > deleterious mutations is to favor a decreased rate of mutation and that
> > the indirect selection resulting from beneficial mutations is small or
> > negligible compared to that resulting from deleterious mutations. . .
>
> Notice the words "indirect selection", referring to the effect on the
> propagation of hitckhiking elements. Also, the following sentences
> which are so conveniently replaced by the ellipsis are "However, this
> does not necessarily mean that removing the beneficial mutation effect
> altogether would result in only a small change in the ESS
> [evolutionarily stable mutation rate]. In the absence of any
> information about the cost function, a general argument is presented to
> explain why this is so."
Exactly. The effect of beneficial mutations in determining the ESS may
be substantial if the cost function has shallow gradient in the region
around the ESS. The problem here is that this doesn't help increase the
numbers of R-best in the population. Regardless of the fact that
beneficial mutations can play a significant part in fitness, this part
just isn't significant enough if it is so outnumbered by detrimental
mutations.
> > Relative to an asexual population, increased levels of recombination
> > reduce the effects of beneficial mutations more rapidly than those of
> > deleterious mutations" (1).
>
> And, in proper context:
>
> "This article models the long-term effect of a series of such
> hitchhiking events and determines the resulting strength of indirect
> selection on the modifier. This is compared to the indirect selection
> due to deleterious mutations, when both types of mutations are randomly
> scattered over a given genetic map. Relative to an asexual population,
> increased levels of recombination reduce the effects of beneficial
> mutations more rapidly than those of deleterious mutations. However,
> the role of beneficial mutations in determining the evolutionarily
> stable mutation rate may still be significant if the function
> describing the cost of high-fidelity replication has a shallow
> gradient."
>
> I am surprised Sean. I would have thought such blatant quote-mining
> was beneath you.
How is this blatant quote mining? This last part actually helps make
my point - as noted above. If the function describing the cost of
high-fidelity replication has a shallow gradient, then the role of
beneficial mutations may be significant. Hmmmm . . . The fact of the
matter is that "significant" here, even when it is "significant", is
not anywhere significant enough to counter the detrimental mutation
rate.
> > "The probability of fixation of a given beneficial mutation decreases
> > with both population size and mutation rate" (2).
>
> This paper, "Adaptive dynamics during experimental evolution of RNA
> viruses" by Elena et. al., talks about the adaptive stage in the
> evolution of a viruses, when multiple beneficial mutations are in
> competition with one another, a phenomenon known as clonal
> interference. In this stage, due to the presence of a large variety of
> beneficial mutations, "the probability of fixation of a given
> beneficial mutation decreases with both population size and mutation
> rate", as you so aptly mis-quote.
It's not a mis-quote at all. That's exactly what I quoted - word for
word. There is no misquote here.
> Of course, the "winning" beneficial
> mutations are the ones that contribute to the greatest increase in
> overall population fitness. Again, this paper says nothing about the
> effects on fitness from beneficial mutations versus the effects from
> deleterious ones, or the differences in their fixation rates.
The point is that even given multiple beneficial mutations, they will
not all become fixed because of the interference problem. Obviously,
the most beneficial mutation will most likely make it, but not all
beneficial mutations will. This reduces the effective beneficial
mutation rate even more. The most beneficial mutations and actually
become fixed certainly do not counteract the hoard of detrimental
mutations with which the R-fittest must contend.
Exactly! The problem is that deleterious mutation rates are far more
likely to be increased than beneficial mutation rates. Try exposing an
organism to radiation and see what happens. Even an normal rates,
beneficial mutations are far outnumbered by detrimental mutations and
these detrimental mutations do play a significant role in limiting the
effectiveness of beneficial mutations as well as the fixation of
beneficial mutations in the gene pool.
> > "Rare reverse and compensatory
> > mutations can move deleterious mutations, via genetic hitchhiking,
> > against the flow of genetic polarization. But this is a minor
> > influence, analogous to water turbulence that occasionally transports a
> > pebble a short distance upstream. . .
>
> We are now back to Rice's article. First, notice here here that he is
> talking about reverse and compensating mutations - i.e. mutations that
> specifically undo the effects of a preceding deleterious mutation - and
> not beneficial mutation in general.
Oh, come now. Compensatory mutations are about the most common type of
beneficial mutation there is. You just can't get much better than this
as far as rate of beneficial mutations is concerned.
> Secondly, at this point, he is
> talking about asexually reproducing populations specifically. Later
> on, Rice goes on to say:
>
> "Overall, epistasis and nonrandom mating can cause recombination to
> build the best class faster than its own net reproductive rate.
> Clearly, when R_best(realized) << R_best(req) [the best realized
> reproductive rate is much smaller than the required rate], then
> deterministic mutation accumulation will lead to extinction. But when
> the increment (R_best(req) - R_best(realized)) is smaller, a mutational
> Red-Queen may ensue with mutation accumulation being accommodated by
> perpetual compensating adaptation. Recombination, by reducing
> R_best(req) via epistasis and nonrandom mating, extends the permissible
> range of phenotypic complexity (large Ud [rate of deleterios mutation])
> that can potentially evolve."
Exactly. Asexual populations cannot benefit by positive epistasis and
sexual populations can only benefit if, somehow, recombinations happen
in a biased manner. But how does a random process achieve this bias?
> > Whenever demographic, ecological,
> > and/or physiological constraints cause, R-best t to be less than eUd,
> > then the progenitor class will decline in size each generation and
> > deterministic mutation accumulation will ensue. Such mutation
> > accumulation will be opposed by reverse and compensatory mutations, but
> > if R-best is much less than eUd, then net mutation accumulation will
> > ensue" (4).
>
> Notice the constant use of qualifiers such as "whenever" and "if" in
> this parpagraph. And again, he is talking about reverse and
> compensatory mutations specifically: not about adaptations in general.
Not true. In this case R-best will decline regardless of the type of
"beneficial" mutation in question given R-best t to be less than eUd.
Reverse and compensatory mutations are just about the best you've got.
I wouldn't just throw them out if I were you. With a compensatory
mutation, you almost there anyway. These are by far the most likely
beneficial mutations a genome can come across.
< snip >
> [snip]
>
> > > > > 5) The accumulation of positive selective effects in the population.
> > > >
> > > > These do not accumulate nearly as fast as the negative mutations - Even
> > > > in literature the suggested ratio is less than 1 beneficial to 1000
> > > > negative.
> > >
> > > Do you mean "classic English literature"? Because, in the scientific
> > > literature, the ratios are considerably higher. Looking at the article
> > > by Bustamante et. al., they have found evidence that, of the genetic
> > > loci they examined, 9.0% show signs of positive selection, whereas
> > > 13.5% show signs of negative selection.
> >
> > That is not the "rate" of beneficial vs. detrimental mutations. The
> > rate of detrimental mutations far outpaces the rate of beneficial
> > mutations.
>
> I acknowledge that, and below, I have shown how to compute the ratio of
> beneficial to determinetal mutations from Bustamante et. al. Using
> Nachmann and Crowell's estimate of 73% of non-synonymous subsitituions
> being deleterious, we can compute the rate as 30 beneficial mutations
> to 1000 deleterious.
I don't think so. Refer to the Fay et al reference above an not that
they suggest an 80% detrimental mutation rate for non-synonymous
substitutions being deleterious - with 20% neutral with respect to
function. They go on to propose a very low beneficial mutation rate
that is far lower than 3%. How in the world did you make this
calculation anyway? Don't you have to know two of the three (both the
detrimental and neutral rates) to calculate the third (the beneficial
rate)? I don't see where you made this calculation?
> > For E. coli, the estimated value for the beneficial mutation rate
> > (Miralles et. al., 1999) was 6.4 × 1e-8 beneficial mutations per
> > genome per generation. (1) The beneficial mutation rate obtained by
> > Imhof and Schlötterer was 4 x 1e-9 per genome per generation (see
> > reference links below). (2) Compare this with the detrimental mutation
> > rate for E. coli "in excess of 0.0002" per genome per generation. (3)
>
> Why quote beneficial mutation rates in viruses and bacteria, when I'm
> giving you the most recent estimates for humans?
I don't see that you are giving me beneficial mutation rates for humans
at all. The rates for bacteria, at least as far as ratios of
detrimental vs. neutral, are comparable to eukaryotic creatures - like
humans.
> Rates of both
> beneficial and deleterious mutations differ greatly between these
> organisms.
Yes, but the ratio is usually between 1 in 3,000 to 1 in 50,000.
> It's quite simple to compute these ratios in human coding
> regions using the estimates of Bustamante et. al., coupled with N&C's
> or Kimura's estimate of the proportion of deleterious non-synonymous
> substitutions.
Hmmmm - show me.
> > That produces a ratio of between 1 in ~3,000 to 1 in ~50,000. But what
> > about eukaryotes? "In sexual populations of higher eukaryotes, there is
> > extensive data showing that U >> K." (4)
>
> Certainly, taking K/U = 3/100, U >> K. But that K/U ratio is still
> much greater than 1/1000 that you were claiming, with no justification
> from human genomic data.
Note the Fay et al reference above.
> > "In general, organisms with
> > larger genomes appear to have a greater number of deleterious
> > mutations, although it does not appear that the deleterious mutation
> > rate is constant per base pair across these organisms." (5)
>
> Quoting rampantly again? This only means that the rates of deleterious
> mutations are higher in organisms with larger genomes, not that the
> rates of beneficial mutations don't increase in the same or even
> greater proportions.
I never meant to indicate otherwise. I fully expect the beneficial
rates to increase in proportion, but not at significantly greater rates
than detrimental mutations. I fully expect the ratios to be in the
range of at least 1 in 1000. But please, do show your own
calculations.
> > > Now, even if we were to accept
> > > Kimura's estimate that 86% of all functional mutations in primates are
> > > negative, we would still arrive at the conclusion that 0.09 * (1-0.86)
> > > = 1.3% percent of all functional mutations are beneficial, whereas .86
> > > + .135*(1-0.86)= 87.9% of all functional mutations are deleterious.
> > > That gives a ratio of 15 beneficial to 1000 negaitve, which is 15 times
> > > what you state. Of course, this ratio would be even bigger using N&C
> > > estimates of the fraction of deleterious mutations (about 30 to 1000).
> >
> > Perhaps you've misread Kimura?
>
> I haven't read that particular paper by Kimura - I took the figure from
> N&C.
>
> > It seems that the "86%" number is not the number of all "functional" mutations,
> > but of nonsynonymous substitutions in a functional protein - like the hemoglobin
> > protein used by Kimura in this particular study.
>
> My bad. I should have written "non-synonymous" instead of
> "functional".
It makes a very big difference. This was a complete misquote on your
part on a vital issue here. And you are chastising me for "quote
mining"?! Come on now . . .
> > N&C write:
> >
> > "What proportion of nonsynonymous changes are neutral and what
> > proportion are deleterious? The fraction that are neutral, fo, can be
> > calculated by comparing the total mutation rate, µt, with the
> > substitution rate, vo = foµt (KIMURA 1983A, KIMURA 1983B). The
> > proportion that are deleterious is 1 - fo. Using this approach, KIMURA
> > 1983B estimated that 86% of nonsynonymous substitutions are
> > deleterious. . .
>
> Yep, that's where I got the 86% percent figure.
Not that this doesn't translate into a 14% beneficial mutation rate
like you suggested. The 14% rate is the neutral mutation rate. Big
difference . . .
> > The genomic deleterious mutation rate is likely much
> > larger given our estimate that 80% of amino acid mutations are
> > deleterious and given that it does not include deleterious mutations in
> > noncoding regions, which may be quite common."
> >
> > http://www.genetics.org/cgi/content/full/156/1/297
>
> This phrase does not appear anywhere in N&C's 2000 Genetics paper you
> link to. Are you sure you have your quotes mined correctly?
Sorry, I must have hit the wrong button on my computer. The proper
reference for this passage is the Fay et al paper published in
Genetics:
In context, the quote reads:
"The number of deleterious mutations an individual carries and
their individual selection coefficients cannot be used in a
straightforward estimate of fitness reduction or genetic load. A
genomic deleterious mutation rate greater than one suggests epistatic
interactions between deleterious mutations in their translation to
fitness (CROW 1970). The genomic deleterious mutation rate in humans
was previously estimated to be at least 1.6 on the basis of an estimate
that 38% of amino acid mutations are deleterious. The genomic
deleterious mutation rate is likely much larger given our estimate that
80% of amino acid mutations are deleterious and given that it does not
include deleterious mutations in noncoding regions, which may be quite
common (SHABALINA and KONDRASHOV 1999). The combined NC*/S* ratio of
common SNPs from both surveys indicates 50% of the noncoding sites are
constrained and must serve some function. Because an equal number of
noncoding and amino acid-altering sites were surveyed, noncoding
mutation should contribute at least 60% (0.50/0.80) as much as coding
mutations to the total genomic deleterious mutation rate."
http://www.genetics.org/cgi/content/full/158/3/1227
Note that I not only list the passage for you guys, I also generally
try to give an easily accessible reference to a free online journal -
if possible. This is a very far cry from "quote mining". Even when
you completely misquote someone, like you did with your whole 14%
positive mutation rate, I at least give you the benefit of the doubt
and don't accuse you of blatant misrepresentation - "like an
evolutionist". Come on now? What are you trying to do here?
> Also, note that while we cannot estimate the rate of potential
> deleterious substitutions in the non-coding regions, we cannot estimate
> the rate of potential beneficial substitutions in those regions,
> either, but there's no reason to believe that beneficial mutations
> would be more uncommon, relatively speaking, in those regions, than
> deleterious mutations. Regardless, estimates of beneficial to
> deleterious ratios for the coding region of the human genome is all
> we've got at the moment.
I do believe that the ratio stays the same, but it is so outclassed
that the only important thing is the increase in the deleterious
mutation rate.
< snip >
Sean Pitman
www.DetectingDesign.com
Richard Forrest
Seanpit wrote:
> Richard Forrest wrote:
>
> > Sean, you haven't.
> > You have made assertions which have been shown to be false, refused to
> > read the references which demonstrate that, refused to make any changes
> > to the content of your essay on your web site even when it has been
> > shown to contain demonstrable falsehoods, and ignored or dismissed out
> > of hand any criticisms of that site.
> >
> > This is not "discussing" flood geology. It is simply a blanket denial
> > of any contrary evidence.
> >
> > RF
>
> Again, that is certainly your opinion. I find it most interesting then
> that you continue to want to talk to someone who you think incapable of
> true "discussion" ; )
I'm pointing out that you haven't discussed anything, not making the
assertion that you are incapable of discussion.
>
> Regardless, I'm not interested in talking to you about your notions
> about what you think I've gotten wrong on my website. So far I haven't
> seen anything from you that I find personally convincing.
So you stand by your statement "For fossilization to occur, burial
must be very rapid"?
This is in spite of the fact that I have posted several links to web
sites, published papers, and reference books which offer a detailed and
specific argument that this is not the case, and moreover indicate the
physical evidence in the form of fossils and formations to support this
argument which you can check for yourself.
Why do you not find this convicing?
You have in a subsequent post qualified that statement by saying that
it refers to the fossils you describe in your text. It's worth noting
that you have not changed your web site to reflect this. I can
understand why you don't because to do so would undermine your argument
for a global flood, but this hardly excuses it.
I have offered several references to taphonomic conditions in the
liassic deposits where some of the specimens to which you refer were
found. These references argue that the fossils are so well-preserved
because they were immersed in soupy substrate. This is not the rapid
burial on which your argument depends. Furthermore, those fossils
frequently show signs of colonisation by invertebrate scavengers, again
something inconstent with rapid burial.
What aspects of the evidence or argument do you not find convincing?
> That doesn't
> mean you're wrong or that many if not most others find you very
> convincing - but I don't.
What aspects of the evidence or argument do you not find convincing?
> Beyond that, I'm busy with other topics
> right now. I just don't have to time to discuss everything with
> everybody who wants to discuss his or her own favorite topic with me.
> Until I am, no one is forcing you to be here. You can always go
> elsewhere and join some other thread. You're certainly not contributing
> much of anything to the topic at hand in this thread.
>
> Sean Pitman
> www.DetectingDesign.com
If you are not interested in the subject, and show no signs of wanting
to learn about the subject, why do you post an essay on the internet
which deals with the subject?
It seems that you are quite content to post essays full of demonstrable
falsehoods on the internet.
Why is this?
RF
Seanpit
Let me make this as clear as I can. I'm not interested in talking to
you or anyone else about any other typics than the neutral gap problem
right now. I can't spend my limited time on all the topics that I'm
very much interested in. This doesn't mean I'm not interested in your
point of view. I am. I just don't have the time to discuss your point
of view right now. Currently, I'm more interested in other topics.
Sorry . . .
The reason why I haven't changed the stuff on my website dealing with
your favorite topics is because I haven't found your arugments
convincing. You may find that amazing and utterly mysterious since
your arguments are so overwhelmingly clear to you, but they aren't
clear to me. Perhaps someday they will be? For now though, I don't
have the time. Perhaps some other day? I'll let you know. Ok?
Until then, if you don't want to discuss the neutral gaps problem or
the problem of detrimental mutations building up faster than they can
be eliminated, then I'm really not interested in getting sidetracked at
the moment. Does that make sense? I'm sure there are others who would
be glad to talk to your about your ideas though. Just make a post and
I'm sure someone will want to talk to you about them.
Sean Pitman
www.DetectingDesign.com
RobinGoodfellow
> RobinGoodfellow wrote:
>
> < snip >
>
> > You seem to give a lot of weight to the fact that N&C seem to think
> > that the deleterious mutation rate is too high, and not at all to the
> > fact that they seem to think that humans and chimps diverged from a
> > common ancestor - even though they use the latter as a premise to
> > arrive at the former conclusion (the Ka/Ks ratios). Not too big on
> > consistency, are we?
>
> If U really is greater than 3, all slowly producing animals, to include
> both humans and apes, are headed for extinction.
That is your understanding, yes. Very few evolutionary biologists,
even those who accept the U>=3 rates, seem to share it. I wonder why?
> If U is significantly
> less than 3, then there is something wrong with the assumptions of
> human-ape common ancestry.
I have no idea how you came to that conclusion. Nothing about the
common ancestry of humans and apes implies requires a specific
*deleterious* mutation rate.
> Either way, something's not adding up. How
> is this consistency from your own side?
Well, given the fact that both points you bring up are wrong, I'd say
the consistency of my own side doesn't suffer very much.
> Beyond this, there is good real time evidence that humans do indeed
> sustain upwards of at least 175 mutations per person per generation.
> Oh, and by the way, my use of "29 mitotic divisions" before
> reproduction of the next generation is a male to female *average*. The
> average for woman is about 23, while the average for men is a good bit
> higher, creating an overall average of about "29". Was this really all
> that hard to understand from the way I wrote my essay?
You are the one who is not understanding me, Sean. I am talking about
estimates of *deleterious* mutation rates specifically, which relies on
the assumption that deleterious non-synonymous substitutions have been
eliminated by selection since humans and other primates split from one
another. Here's a paragraph you didn't address from the "Horse and
Donkey" thread:
"No matter. Assuming you have no problem with their computation of the
overall mutation rate, how do you reconcile their computation of the
deleterious mutation rate with your young life position? The
deleterious fraction is estimated as 1-Ka/Ks, where Ka/Ks is the
ratio of the number of non-synonymous to synonymous substitutions.
However, if you think life was created 10,000 years ago, you would need
to assume that the overwhelming majority of both synonymous and
non-synonymous substitutions were deliberately designed into chimp and
human genomes (unless you want to posit a past mutation rate
approximately 600 times higher than currently estimated). Thus, if
both synonymous and non-synonymous substitutions are designed, we
cannot use 1-Ka/Ks to estimate the deleterious fraction. (I suppose
you could use Ka/Ks estimates from human genes only, but I can't think
of any such studies off the top of my head.)"
And a follow-up from me, a little later in the thread:
"It occurred to me that there is plenty of human SNP data out there to
help make such a determination. Unfortunately, it doesn't help you.
Latest result seem to indicate that among humans, Ka/Ks rates are
pretty close to 1, meaning that they can't really be used to determine
deleterious mutation rates in a manner suggested by N&C. (See, e.g.
Genome Research, 14:2034-40). In order to compute these rates, you
have to start with idea that man and other primates diverged through
evolution. Tough luck, Sean."
> Now, if each human individual sustains 175 mutations per generation,
> how many of these will hit the "functional" area of the genome?
> Consider the following discussion by Fay et al published in the July
> 2001 issue of Genetics:
>
> "The genomic deleterious mutation rate in humans was previously
> estimated to be at least 1.6 on the basis of an estimate that 38% of
> amino acid mutations are deleterious. The genomic deleterious mutation
> rate is likely much larger given our estimate that 80% of amino acid
> mutations are deleterious and given that it does not include
> deleterious mutations in noncoding regions, which may be quite common
> (SHABALINA and KONDRASHOV 1999).
And how do you think the estimate of 80% amino acid mutations being
deleterious is arrived at? By computing the ratio of non-synonomous to
synonymous substitutions, and assuming that deleterious non-synonymous
substitutions have been eliminated from ancestral populations by
selection. The sort of assumption you can't be making, since you can
never know which substitutions were "designed" into the genomes and
which weren't at the time humans and chimps were separately created.
> The combined NC*/S* ratio of common
> SNPs from both surveys indicates 50% of the noncoding sites are
> constrained and must serve some function. Because an equal number of
> noncoding and amino acid-altering sites were surveyed, noncoding
> mutation should contribute at least 60% (0.50/0.80) as much as coding
> mutations to the total genomic deleterious mutation rate."(1)
>
> So, where is your notion that U=1.6 now? These figures put U well over
> 5. As far as the beneficial mutation rate, Fay et al go on to say:
>
> "The slightly deleterious fraction, f1, cannot exceed 0.80 since 20% of
> mutations were estimated to be neutral"
Meaning that, allowing that some of the substitutions were the result
of positive selection, the slightly deleterious fraction is somewhat
less than 80% percent. 80% is an upper bound.
> "... Extrapolating this
> proportion to the total amount of coding DNA in the genome (~5 x 10^7
> bp) yields an estimate of up to 1 advantageous substitution every ~200
> years since humans separated from old world monkeys 30 million years
> ago."
Quote mining again, are we? Let us restore the full paragraph:
"The large number of amino acid substitutions suggests a high rate of
adaptive evolution in primates. The expected number of amino acid
substitutions is 2382 (4151 x 70/122) based on the A/S ratio of common
polymorphism and the excess is 1278. Therefore, a large proportion,
***35%***, of amino acid substitutions between humans and old world
monkeys ***are estimated to have been driven bypositive selection***.
Extrapolating this proportion to the total amount of coding DNA in the
genome (~5 x 107 bp) yields an estimate of up to 1 advantageous
substitution every ~200 years since humans separated from old world
monkeys 30 million years ago (LI 1997 )." [Emphasis mine]
Now, you did provide the link to the article, and by now, you must have
assumed I would read it. What were you trying to accomplish by
omitting the highly relevant context above?
Incidentally, the quote, even without the context, does not mean what
you think it means. Substitution rates are measured *per population*
per generation, whereas mutation rates are measured *per individual*
per generation. Under the neutral model, the two quantities are equal:
within the context of slightly beneficial mutations, substitution rates
will be slighly higher than the mutation rate. So, a rate of 1
substitution every 200 years, is the rate of 0.125 substitutions per
population per generation, assuming 25 year generations. Which means
that the rate of slightly beneficial mutations will be slightly lower
than 0.125 per individual per generation. Taking U=3, the ratio of
beneficial to deleterious mutations is about 0.125/3, or slightly over
4 percent. Which is one percent higher than I originally estimated.
Thanks, Sean!
Here's a link on how substitution rates are computed from mutation
rates, within the context of the neutral model:
http://www.stat.berkeley.edu/users/terry/Classes/s260.1998/Week13a/week13a/node10.html
Now, how can you expect to discuss these topics and be taken seriously
when you don't even have rudimentary knowledge of basic population
genetics concepts?
> And you think that is an adequate enough fixation rate to compensate
> for a U of over 5?
OK. For U=5, the ratio of beneficial to deleterious is about 2.5%,
taking Fay et. al. figures. Yeah, I think that's high enough.
> 1. Justin C. Fay, Gerald J. Wyckoff, and Chung-I Wu, Positive and
> Negative Selection on the Human Genome, Genetics, Vol. 158, 1227-1234,
> July 2001 -http://www.genetics.org/cgi/content/full/158/3/1227
>
> > But, to answer your point, N&C had no way to assess the impact of
> > non-coding DNA on U. They could only compute U for coding DNA, and
> > those estimates vary considerably depending on their different
> > estimates of overall mutation rates.
>
> Ah, but N&C did comment that their estimate of U was likely biased
> downward because of this problem. You just choose to overlook that?
Of course not. But N&C don't say how far downward, do they? I choose
not to piss in the wind by trying to extrapolate to data that isn't
there.
> > N&C's lower bound estimate of U
> > is 1.5, and they site another estimate of 1.6 (from 1999, By
> > Eire-Walker and Keightley) that agrees with that lower bound. I'm not
> > saying that is necessarily the correct estimate either, but your desire
> > to accept U>=3 suggests that you've already made up your that we are
> > devolving towards extinction, and are cherry-picking data to support
> > this position.
>
> I'm not cherry picking to support my position. I'm just stating what
> the author's themselves believed to be the most likely estimate for U.
> They are the ones who said that U is most likely greater than 3, even
> greater than 5, not me. Rather, it seems like you are the one who is
> "cherry-picking" the extremes of the author's data set in an effort to
> support your views. The most recent data, according to the authors,
> not me, clearly supports U values of well over 5.
My point is, these rates are notoriously hard to estimate, and vary
widely between the loci being studied and the assumptions being made.
Until we have whole-genome studies assessing these rates, I'm unwilling
to believe any of these estimates: or rather, equally willing to
believe them all. And none of these numbers, by themselves, imply that
we or the chimps are headed anywhere toward extinction.
> > > In fact, the odds are most likely a
> > > whole lot higher than U=3 now that a lot more of what was thought to be
> > > "junk" DNA, because it doesn't code for proteins and whatnot, is no
> > > longer junk, but functional and constrained by natural selection after
> > > all. This has lead to the suggestion that U is actually "greater than
> > > 5" (see references below).
> >
> > The paper you reference by Rice you are referring to simply mentions
> > this figure in passing, citing three other studies, the latest of which
> > came of 1993. This pre-dates the U=1.6 estimate of Eire-Walker and
> > Keightley by 6 years, and N&C's estimate of U=3 by 7 years.
>
> And why do you think Rice chose to print the estimate of U = 5? Again,
> what do you think about the Fay paper will estimates of well over 5?
I think they are plausible, but far from as certain as you would like
to believe.
> > Of course,
> > the ironic part is that Rice's paper proposes specfically how epistasis
> > effects can compensate for such high deleterious mutation rates.
>
> Rice does point out several ways to slow down the effects of
> deleterious mutation rates, but he doesn't show how a high U, like a U
> above 3, can really be turned around. He does list off several kinds
> of epistasis as well as the positive effects of biased breeding, but
> none of these really compensate for a detrimental mutational load of U
> greater than 3. The net effect, even with buffering epistasis combined
> with biased mating is not enough to counteract a very high detrimental
> mutation rate given a low reproductive rate.
How nice of you to make this assertion, especially since Rice doesn't
actually explore the case of buffering epistasis *together* with biased
breeding, but rather studies their effects separately (see figure 3).
He finds that both can lower R_best (the required best class
reproductive-rate) by raising the product S*O* (standardized selection
gradient times opportunity for selection, both of which he defines as
the paper) above 1, as does reinforcing epsistasis when recombination
is present. The cumulative effects of the two types of epistasis,
combined with non-random mating, could be quite considerable. In fact,
according to Rice, non-random mating alone can have a very pronounced
effect. Here's the relevant part of the paper.
"Unlike the two forms of epistasis examined above, however, positive
assortative mating builds strong beneficial (i.e., variance enhancing)
linkage disequilibrium [ln(D) >> 0] and can thereby provide a
substantial reduction in the requisite mutational load. Even weak
positive assortative mating can produce a large reduction in load (data
not shown). Obviously, negative assortative mating would increase the
requisite load of a recombining population." (i.e. The last sentence
goes to show that some forms of non-random mating can actually lead to
an increase in mutational load.)
> For example, take a steady state population of 1 billion, half women,
> half men, with a reproductive rate of 10 children per woman and U of 3.
> All of these people in this initial population start out with equally
> good genomes. In other words, all of them are at "R-best". Now, each
> woman has 10 pregnancies to populate the next generation. Out of these
> 5 billion 1 in 20 are still at R-best. The rest have at least 1
> detrimental mutation with an average of 3 detrimental mutations. That
> means the next generation will start out with at most 250 million
> R-best individuals. The rest will be R-not as good. Even if these 250
> million R-best mate exclusively with each other, regardless of the
> presence of buffering epistasis or not, the number of R-best in the
> next generation will be a mere 62 million in a steady state population
> of 1 billion. How is this decline reversed without dramatically increasing the
> reproductive rate?
Now consider the next-best class. After one generation of matings
(starting from the optimally fit population), the number of inviduals
in the next-best class (carrying one deleterious mutation each) will be
around 750 million. (See Kimura and Moruyama, Genetics 54:1337-1351,
Equation 3.1). Now, assume that the best class, like the next-best
class, also mates with one another. Assuming infinite loci (which are
assumed in calculating genetic loads - see the equation for w-bar right
under equation 3.1 in the K&M paper), the deleterious mutations carried
by a pair of individuals in each class are all at different loci.
Thus, thanks to genetic recombination, one in four offspring of parents
in this class will have no deleterious mutations whatsoever. If each
of 375 (750/2) million couples produce 10 offspring, and 1 in 20 of
those has no deleterious mutations, that's 3.75e8*10/4*20 ~ 50 million
individuals entering the best class from the next-best in generation 2.
Plus, another 15 million from the second-best class, e.t.c So, in two
generations later, the number best-fit individuals is over 125 million,
rather than just 62 mil. While this still shows a decline in the
number of best fit individuals, this decline is twice as small in
proportion to the first generation. Furthermore, as the less-fit
classes grow, so will the proportion of all recombinant offspring
returning to the more-fit classes, effectively lowering mutational load
requirements. Again, I strongly encourage you to read Rice's paper.
The important thing here is that, once recombination is present, the
lower fitness classes are no longer members of the "living dead", since
invidiuals from lower-fitness classes can produce offspring in
higher-fitness classes. Given non-random mating, these rates of
re-entry into higher-fitness can be amplified considerably.
> > > Your notion of U is in reality as low as 1.5 is highly unlikely.
> >
> > It's not my notion at all. I am arguing that in order for your idea to
> > have some merit, you need to be able to shown not only that U>=3 (Which
> > is by no means clear),
>
> It is quite clear that most on your own side feel that U is well over 3
> - even over 5.
Most people that you cite, yes. But the evidence is not quite there
yet, and likely won't be until whole-genome studies are conducted.
When they are, I suspect the result will be that it's meaningless to
talk about one single U, since the estimates will likely vary quite a
bit from locus to locus.
> > but that epistasis effects aren't real or
> > sufficient, beneficial mutations can't maintain a an overall level of
> > fitness, deleterious effects are not, on average, so small as to have
> > extremely little effect on the overall fitness of the population in
> > practical terms, and some other things besides. I am certainly willing
> > to beleive that U>=3, but a) it is far from certain, and b) it hardly
> > consitutes evidenc e that we are devolving toward extinction.
>
> Sure it does. If U is over 3 and the beneficial mutation rate isn't
> anywhere near what you say it is, what's left to reverse the trend and
> make more R-bests in each generation than are lost as a ratio of the
> current vs. the previous number of R-bests?
Why don't you actually read Rice's paper, instead of looking for quotes
supporting your position? He details a number of explanations.
> Not only is the beneficial rate nowhere near as high as you say, even
> if it were 3 in 100 mutations, how would this help? Even if every one
> of these 3 beneficial mutations became fixed, you'd have to get rid of
> 100 detrimental ones for each of these beneficial mutations to tilt the
> balance.
Because beneficial mutations tend to spread faster through a population
than deleterious ones. Besides, when talking about the overall effects
on fitness, it is the individual effect of each mutation, not merely
their, number that matters. A strongly beneficial mutation will likely
quickly fix in a population, raising the overall level of fitness. A
strongly deleterious mutation will likely be quickly eliminated, thus
having no effect on the overall fitness in the long run. The
assumption that the only thing that matters for an organism's fitness
is the number of deleterious mutations it carries may be nice for
performing genetic load computations, but it is a considerable
over-simplification of reality. (Or didn't you know that this
assumption is being made? Again, read Kimura and Maruyama.)
[snip]
> > > I have read several of Kimura's papers as well as several papers from
> > > others dealing with this problem by suggesting some form of epistasis
> > > (multiplicative increases in the effects of detrimental mutations). The
> > > problem is that with detrimental mutation rates as high as 3 per person
> > > per generation, positive epistasis only increases the death rate, but
> > > does not clear detrimental mutations from the gene pool faster than the
> > > fitness of the most fit individual in the population decreases.
> >
> > Apparently, you haven't read enough papers. What you describe is
> > reinforcing epistasis, which is only one several possible forms of
> > positive epistasis. Rice, for instance also discusses the possible
> > effects of buffering and pathway epistasis, both of which can have a
> > significant effect on reducing genetic load in sexually reproducing
> > populations according to his simulations.
>
> Yes, but his simulations make some very interesting assumptions, such
> as biased mating and biased translocations. How does this happen in
> the real world?
First, Rice's simulations make do not assume any translocation biases -
this is pure invention on your part. But, talking about interesting
assumptions ... Does it happen in the real world that all mutations
carry the exact same deleterious effect on fitness, and all that
matters is their number? Does it happen in the real world that the
number of potential loci for mutations is infinite, and in steady
state, the population may contain individuals carrying arbitrarily
large (essentially infinite) numbers of deleterious mutations? Does it
happen in the real world that mutations arrive into a population in
numbers precisely following a Poisson distribution? Does 1-Ka/Ks
really give a good estimate for the fraction of deleterious amino acid
substitutions? All those assumptions are made in calculating genetic
loads, Sean. (again, I strongly urge you to read Kimura and Maruyama:
at least the section on asexual reporduction), and all of them are only
meant to apporoximate reality. Just as the assumption that all
organisms mate approximately in their fitness class: i.e. positive
assortative mating.
> > > Of course, sexual reproduction is supposed to help out with this problem,
> > > but sexual reproduction doesn't even help unless mating is done in a
> > > non-random way.
> >
> > And of course, in real life, mating happens in a purely random fashion.
>
> Do you know how many detrimental mutations your wife has? Does your
> wife know how many you have?
Since I am not married, the answer is a clear and resounding no. :)
> > I don't know if you're married, but if you are, I hope you didn't just
> > ask the first girl you saw on the street to be your wife - and if you
> > did, I hope for both your sakes she didn't say "yes".
>
> My wife is downright gorgeous, and looks quite "fit" to me.
Kudos to you, then! However, you are a doctor, quite intelligent, a
pretty nice guy (extrapolating from your behavior on this forum), and
from the picture on your on your website, reasonably good-looking. I
am sure that your wife, in addition to being gorgeous, is also smart,
educated, nice, and has a host of other traits that are found to be
desirable in human society. Do you still think still she would have
married you if you were an abusive jerk? Would you have given her any
consideration had she been grotesque in appearance and dumb as a post?
Sure, such matches do happen, but they are the exception rather than
the norm. Rather, part the reason that you married to one another is
that you are both in each other's "league" - though, I hope, it is only
a very small part of the whole picture.
Note the above is primarily meant as an analogy, and not a very good
one, since human notions of what is desirable are, to a large degree,
artificial and culturally induced, and don't readily correlate with
biological fitness. Still, the idea that organisms reproduce primarily
with other organisms in their "fitness class" isn't really that radical
or improbable.
[snip]
> > More seriously,
> > don't you think that an organism suffering many deleterious mutations
> > will have much lower chances of mating successfully, especially with
> > organisms of the "most fit" genotype?
>
> Not necessarily. Studies have shown that even with a marked increase
> in detrimental mutations a population can continue to grow for some
> time. The odds are very good that many in the R-fittest group will
> actually mate with those of less fit groups.
But, the point is that's it's not just the R-fittest group that's
important here. With recombination, there is a flow of individuals
from lower-fit classes to higher-fit classes in each generation, so the
dynamics the entire population become important.
> There just aren't very
> many R-fittest to go around and some of those less fit babes look fit
> enough - you know?
LOL. Sure, I dig it.
> > There is fierce competition for
> > mates in both the human and the larger animal kingdom: as the result,
> > genetically fit individuals have a higher propensity of mating with one
> > another, and individuals of lower fitness are forced to mate with one
> > another, assuming they can mate successfully at all. This phenomenon
> > is called positive assortive mating, and Rice shows that it can
> > drastically reduce mutational load by eliminating whole swaths of
> > deleterious mutations at once, even in the absence of all other
> > epistaisis effects.
>
> I'm not saying that biased mating does not occur. I'm just saying that
> it isn't nearly complete and that even if it was, it really wouldn't
> solve the problem. What you need is biased mating and biased
> translocations as well.
It doesn't need to solve the problem completely. It needs, in
combination with various epistasis effects, to slow down the problem
sufficiently for beneficial mutations to compensate for the difference.
> > > Of course, there are several problems with this non-random mating
> > > notion. With a detrimental mutation rate of 3 per individual per
> > > generation, only 2 of 40 individuals will have a neutral fitness
> > > balance relative to the original parent population
> > > What are the odds that these 2 individuals will actually mate "preferentially" so that
> > > the overall number of individuals with equivalent parent-level fitness
> > > does not decrease in each generation?
> >
> > Non-random mating does not mean "only best fit individuals mate with
> > each other". In most fit individuals do not derive any benefits from
> > recombination or epistatic effects, since they do not suffer any
> > deleterious mutations in the first place, relative to the rest of the
> > population. Non-random mating could mean any number of things,
> > including the positive assortative mating scenario described above:
> > that is, individuals within the same fitness classes tend to mate with
> > one another.
>
> There's a lot of mixing and matching you know. The fact is that we are
> only concerned about the R-fittest class here.
No, we are not. Read my example above, and then read Rice and
understand what he means by "assortative mating". In fact, the
individuals in the R-fittest class cannot by themselves benefit from
epistasis effects or non-random mating, since those can eliminate or
mitigate the effects of deleterious mutations, which the R-fittest
class does not even have. You really don't understand the subject you
are arguing here, Sean.
Now, you are correct that there is a lot of "mixing and matching", and
that assortative matching is an approximation. But so is pretty every
other concept used to compute mutational loads. Why accept some
unrealistic theoretical models, but not others?
> How do we get more
> R-fittest in each generation? You don't do that by mating your
> R-fittest with those that are less than R-fittest.
>
> > > In every generation every individual receives around 3 deleterious
> > > mutations. The "most fit" phenotype in each generation comes in at a
> > > ratio of about 1 in 20. With that ratio, unless very fortuitous mating
> > > and high reproductive rates take place, the direction is still
> > > extinction. There is no equilibrium here. The only way to avoid this
> > > problem is by having the most fit individuals mate pretty much
> > > exclusively with each other and still make a whole lot of children in
> > > each generation.
> >
> > I will concede this point, more or less.
>
> This is the whole point! This is where the rubber meets the road!
Actually, my mistake. "I will concede this point" should have been
placed after "there is no equilibrium here" in your paragraph. I was
talking about the asexual case here, so mating doesn't even enter into
the picture.
> > My argument was that unlike
> > in the infinite allele/infinite pool of deleterious mutations models
> > that are used to derive genetic loads, the potential for deleterious
> > mutations in a population will go down once a particular slightly
> > deleterious allele becomes fixed - and in fact, the potential for a
> > reverse beneficial mutation opens up (more on that below). In other
> > words, the deleterious mutation rate is not a constant thing.
>
> U is not a constant thing? Where do you get that idea? Have a
> reference for this one? Reverse mutations do happen, but not nearly
> fast enough to keep up in slowly reproducing populations.
Think about it. Every time an individual acquires a deleterious
mutation at a given locus, it means that locus can no longer suffer a
deleterious mutation (well, not quite, but I hope you see where this is
going), and in fact the potential for a beneficial reverse mutation is
created at that locus. Given a finite amount of loci, once enough
deleterious mutations have accumulated, the decreased potential for
deleterious mutaton combined with the increased potential for reverse
mutations can have a significant impact on the effective deleterious
mutation rate (i.e. the rate of deleterious minus compensatory
mutations). Unless, of course, you assume infinite loci, which is
exactly what Kimura did in deriving his equilibrium load equation.
> > However,
> > if the number of loci that can undergo such mutations is high enough
> > and U decreases sufficiently slowly, a population will likely undergo a
> > very significant reduction in fitness - though not necessarily
> > extinction - the genetic load approximations start to break down.
>
> Why not extinction?
It's certainly possible. But if U happens to drop sufficiently after a
number of deleterious fixations, but the population survives, it simply
may reach an equilibrium of lower fitness. I am not sure how likely
this is, but there's nothing inherently impossible about it.
> > > Deleterious mutations happen on an individual basis - more than 3 per
> > > individual per generation. These are then passed on to the next
> > > generation from parents to children and do not leave the gene pool
> > > until they die out of it with the premature death of an individual who
> > > didn't get a chance to replicate. They don't need to "propagate" any
> > > other way to be a problem. Their rate of propagation is not
> > > significantly related to the size of the population, but to the
> > > reproductive rate and death rate.
> >
> > As the population size grows, you will see a broad spectrum of mutants
> > with accumulation slightly deleterious mutations, but any one specific
> > deleterious mutation at a specific locus will have a much harder time
> > achieving fixation in the population. For sexual populations, this is
> > where recombination can act to eliminate multiple deleterious mutations
> > in fell swoops, especially in the context of non-random mating and
> > epistasis.
>
> You don't need to have fixation for deleterious mutations to be a
> problem. Each individual sustains 3 detrimental mutations. The next
> generation will have 6, the next 9, etc. None of these needs to become
> fixed for the overall fitness level of the population to decline. How
> is recombination going to help? How does random recombination know how
> to cluster the bad together preferentially for batch elimination?
When two individuals with N deleterious random mutations (interspersed
with equal probability throughout the genome) recombine their genes,
the recombinant will have either more than N or fewer than N
deleterious mutations with high probability. If it is fewer than N,
the higher-fitness class is propped up, improving the overall fitness
of the population. If it is more than N, the new individual has lower
fitness than either parent, and is less likely to propagate its genes.
In either case, the overall fitness of the population will be improved
in the long run.
> It is quite clear that if such clustering could actually be realized that
> non-random mating and epistasis would most certainly take care of the
> rest, but I don't understand how recombination is going to help here.
Well, I hope the above clears things up for you somewhat.
> > > > whereas beneficial mutaitons will arise and
> > > > spread much more rapidly (this is one of the basic theorems of
> > > > population genetics).
> > >
> > > Not quite true.
> >
> > Ah, Quote-Mining, the discerning creationist's way of doing science
> > (TM)! None of those quotes actually say what you want them to say when
> > placed in proper context, Sean. Let's examine closer, shall we?
>
> Lets . . .
>
> > The first paper, "Beneficial Mutations, Hitchhiking and the Evolution
> > of Mutation Rates in Sexual Populations" by Toby Johnson (Genetics,
> > Vol. 151, 1621-1631, April 1999), says nothing about the fixation rates
> > of beneficial mutations, but rather their effect on modifiers of
> > mutation rate. They concude that, according to their model,
> > deleterious mutations have more of effect on the overall mutation rate
> > due to stronger indirect selection of hitckhiking rate modifiers. It
> > has nothing to do with whether or not beneficial mutations fix faster
> > in a population or have more of an effect on the overall fitness of the
> > population compared to their deleterious counterparts.
>
> It seems to me that what the authors do say does play into how
> beneficial mutations affect fitness vs. detrimental mutations. The
> authors write, "Relative to an asexual population, increased levels of
> recombination reduce the effects of beneficial mutations more rapidly
> than those of deleterious mutations."
Yes, and they say this *specifically* regarding the effects of such
mutations on the indirect selection for a hitchiking modifier of the
overall mutation rate. In other words, one very specific special case,
in the context of one very specific theoretical model constructed by
Johnson. This has nothing to do with whether detrimental mutations
have more of a direct effect on the overall fitness of the population
than beneficial ones in the general case.
> Now, I'm not saying that beneficial mutations do not fix faster than
> detrimental mutations. Of course they fix faster.
So, when you wrote "not quite true" in response to my "whereas
beneficial mutaitons will arise and spread much more rapidly [in larger
populations]", you didn't actually mean it? Incidentally, in case you
misunderstood, I meant that in larger populations beneficial mutations
will spread faster, and deleterious mutation will spread slower, *than
in smaller populations* - and thus, the larger the population, the
better beneficial mutations will be at offsetting the negative fitness
effects of deleterious mutations. Is that what you were objecting to
when you replied "not quite true"?
> I'm just saying that
> beneficial mutations are so outnumbered by detrimental mutations that
> they just can't keep up. And, this is made even worse by the fact that
> recombination reduces the effects of beneficial mutations over and
> above what they are able to do in asexual populations.
Which is exactly what the article you cited is *not* about.
> There's really no quote-mining here. I'm not intending to indicate
> anything other than what the authors true intent really is in their
> paper. You just misinterpret my intent - or so it seems.
OK, I'll give you the benefit of the doubt and assume that in your
fervor to prove to me that we are rapidly devolving toward extinction,
you misunderstood this paper, and three others, and then only cut and
pasted the relevant sentences that seem to support your arguments.
That's much better, Sean.
[snip]
> > > "The probability of fixation of a given beneficial mutation decreases
> > > with both population size and mutation rate" (2).
> >
> > This paper, "Adaptive dynamics during experimental evolution of RNA
> > viruses" by Elena et. al., talks about the adaptive stage in the
> > evolution of a viruses, when multiple beneficial mutations are in
> > competition with one another, a phenomenon known as clonal
> > interference. In this stage, due to the presence of a large variety of
> > beneficial mutations, "the probability of fixation of a given
> > beneficial mutation decreases with both population size and mutation
> > rate", as you so aptly mis-quote.
>
> It's not a mis-quote at all. That's exactly what I quoted - word for
> word. There is no misquote here.
http://en.wikipedia.org/wiki/Misquote
You most certainly omitted important context, not mention entire parts
of the sentence containing that quote.
> > Of course, the "winning" beneficial
> > mutations are the ones that contribute to the greatest increase in
> > overall population fitness. Again, this paper says nothing about the
> > effects on fitness from beneficial mutations versus the effects from
> > deleterious ones, or the differences in their fixation rates.
>
> The point is that even given multiple beneficial mutations, they will
> not all become fixed because of the interference problem. Obviously,
> the most beneficial mutation will most likely make it, but not all
> beneficial mutations will.
How does it contradict a basic premise of population genetics that
beneficial mutations will arise and spread more rapidly in larger
populations, which you claimed was "not quite true", using this quote
to back up your position? To quote the very next sentence in the
paper, that you conviniently omitted "(2) As population size or
mutation rate increases, adaptive substitutions result in larger
fitness increases."
> This reduces the effective beneficial
> mutation rate even more. The most beneficial mutations and actually
> become fixed certainly do not counteract the hoard of detrimental
> mutations with which the R-fittest must contend.
What counts is not the rate of beneficial or deleterious mutations, but
their effect on the overall on fitness. If a particular beneficial
mutation wins because it contributes the most to the fitness of a
population, so much the better. In addition, clonal interference
effects are strongest only in those cases where opportunity for
positive selection is already high and thus beneficial mutation rates
are higher than usual.
Non-sequitur. The overall *per nucleotide* mutation rates don't need
to be increased for both deleterious and beneficial rates *per genome*
to increase. Human mutation rates per nucleotide per generation are
much lower than bacterial, but our deleterious and beneficial mutation
rates per genome per generation are much higher, due in no small part
to our larger genomes. At any rate, how does this article contradict
the notion that beneficial mutaitons will arise and spread more rapidly
in larger populations, which you claimed was "not quite true"?
> Even an normal rates,
> beneficial mutations are far outnumbered by detrimental mutations and
> these detrimental mutations do play a significant role in limiting the
> effectiveness of beneficial mutations as well as the fixation of
> beneficial mutations in the gene pool.
Now this is a novel claim. How exactly do deleterious mutations
prevent beneficial ones from achieving fixation (unless, of course, in
a stroke of misfortune, a beneficial mutation gets obliterated by a
series of reverse mutations in every organism in its lineage)?
Besides, I am not claiming that beneficial mutations strictly lead to a
steady increase in the fitness of populations. I'm under no delusion
that we are currently evolving toward a state of perfection. I'm just
saying the relatively high rates of beneficial mutations in humans,
combined with other effects, are probably quite enough to keep our
species from devolving into extinction.
> > > "Rare reverse and compensatory
> > > mutations can move deleterious mutations, via genetic hitchhiking,
> > > against the flow of genetic polarization. But this is a minor
> > > influence, analogous to water turbulence that occasionally transports a
> > > pebble a short distance upstream. . .
> >
> > We are now back to Rice's article. First, notice here here that he is
> > talking about reverse and compensating mutations - i.e. mutations that
> > specifically undo the effects of a preceding deleterious mutation - and
> > not beneficial mutation in general.
>
> Oh, come now. Compensatory mutations are about the most common type of
> beneficial mutation there is. You just can't get much better than this
> as far as rate of beneficial mutations is concerned.
Even if this were true - and this is just the sort of claim you may
want to provide a citation for - it doesn't mean compensatory mutations
are those that contribute most to the overall fitness increases in the
population. In fact, reverse and compensatory mutations by themselves
can never achieve an overall increase in fitness, since all they
accomplish is undoing the effects of previous deleterious mutations.
> > Secondly, at this point, he is
> > talking about asexually reproducing populations specifically. Later
> > on, Rice goes on to say:
> >
> > "Overall, epistasis and nonrandom mating can cause recombination to
> > build the best class faster than its own net reproductive rate.
> > Clearly, when R_best(realized) << R_best(req) [the best realized
> > reproductive rate is much smaller than the required rate], then
> > deterministic mutation accumulation will lead to extinction. But when
> > the increment (R_best(req) - R_best(realized)) is smaller, a mutational
> > Red-Queen may ensue with mutation accumulation being accommodated by
> > perpetual compensating adaptation. Recombination, by reducing
> > R_best(req) via epistasis and nonrandom mating, extends the permissible
> > range of phenotypic complexity (large Ud [rate of deleterios mutation])
> > that can potentially evolve."
>
> Exactly. Asexual populations cannot benefit by positive epistasis and
> sexual populations can only benefit if, somehow, recombinations happen
> in a biased manner. But how does a random process achieve this bias?
Perhaps because mating isn't nearly a random process? And aside from
non-random mating, and epistatic effects, Rice does not invoke any
other mechanisms in order to achieve reduction in mutational load. The
need for biased translocation is something you invented simply because
you don't understand the actual processes involved.
> > > Whenever demographic, ecological,
> > > and/or physiological constraints cause, R-best t to be less than eUd,
> > > then the progenitor class will decline in size each generation and
> > > deterministic mutation accumulation will ensue. Such mutation
> > > accumulation will be opposed by reverse and compensatory mutations, but
> > > if R-best is much less than eUd, then net mutation accumulation will
> > > ensue" (4).
> >
> > Notice the constant use of qualifiers such as "whenever" and "if" in
> > this parpagraph. And again, he is talking about reverse and
> > compensatory mutations specifically: not about adaptations in general.
>
> Not true. In this case R-best will decline regardless of the type of
> "beneficial" mutation in question given R-best t to be less than eUd.
This is your assertion, and is in no way reflected in the paper in
question.
> Reverse and compensatory mutations are just about the best you've got.
> I wouldn't just throw them out if I were you. With a compensatory
> mutation, you almost there anyway. These are by far the most likely
> beneficial mutations a genome can come across.
I'd love to see some evidence for your assertion that reverse and
compensatory mutations are "by far the most common", or more
importantly, that they have the largest effect on improving fitness, as
compared to, say, novel adaptations. But I'm not throwing them out -
Rice is, primarily because modeling them would be inconvenient for his
analysis. I don't think the effects of compensatory mutations can be
quite so callously disregarded as slighly deleterious mutations
accumulate in various lineages, but I wasn't the one who wrote the
paper.
> < snip >
>
> > [snip]
> >
> > > > > > 5) The accumulation of positive selective effects in the population.
> > > > >
> > > > > These do not accumulate nearly as fast as the negative mutations - Even
> > > > > in literature the suggested ratio is less than 1 beneficial to 1000
> > > > > negative.
> > > >
> > > > Do you mean "classic English literature"? Because, in the scientific
> > > > literature, the ratios are considerably higher. Looking at the article
> > > > by Bustamante et. al., they have found evidence that, of the genetic
> > > > loci they examined, 9.0% show signs of positive selection, whereas
> > > > 13.5% show signs of negative selection.
> > >
> > > That is not the "rate" of beneficial vs. detrimental mutations. The
> > > rate of detrimental mutations far outpaces the rate of beneficial
> > > mutations.
> >
> > I acknowledge that, and below, I have shown how to compute the ratio of
> > beneficial to determinetal mutations from Bustamante et. al. Using
> > Nachmann and Crowell's estimate of 73% of non-synonymous subsitituions
> > being deleterious, we can compute the rate as 30 beneficial mutations
> > to 1000 deleterious.
>
> I don't think so. Refer to the Fay et al reference above an not that
> they suggest an 80% detrimental mutation rate for non-synonymous
> substitutions being deleterious - with 20% neutral with respect to
> function.
As Fay et. al., 20% of non-synonymous substitutions being neutral with
respect to function means that *at most* 80% of non-synonymous
substitutions are deleterious - i.e. a fraction of those could have
been beneficial.
> They go on to propose a very low beneficial mutation rate
> that is far lower than 3%.
Actually, as I've shown above, from their figures we can compute that
the ratio of beneficial to deleterious mutations is actually above 4%.
It's just that you don't understand the difference between mutation
rate and substitution rate.
> How in the world did you make this
> calculation anyway? Don't you have to know two of the three (both the
> detrimental and neutral rates) to calculate the third (the beneficial
> rate)? I don't see where you made this calculation?
The calculation is actually in the quoted text below. But here it is
again:
"Looking at the article by Bustamante et. al., they have found evidence
that, of the genetic loci they examined, 9.0% show signs of positive
selection, whereas 13.5% show signs of negative selection. Now, even
if we were to accept Kimura's estimate that 86% of all [non-synonymous]
mutations in primates are negative, we would still arrive at the
conclusion that 0.09 * (1-0.86) = 1.3% percent of all [non-synonymous]
mutations are beneficial, whereas .86 + .135*(1-0.86)= 87.9% of all
[non-synonymous] mutations are deleterious. That gives a ratio of 15
beneficial to 1000 negaitve, which is 15 times what you state. Of
course, this ratio would be even bigger using N&C estimates of the
fraction of deleterious mutations (about 30 to 1000)."
Edit: I replaced every instance of "functional" - to be honest, I have
no idea why I wrote it in the first place - with non-synonymous, which
is what I originally should have written. If you read the rest of the
paragraph, you will actually see that the calculations actually bear
this out (i.e. I never claim the 14% figure that you attribute to me
later on.) To see how I arrived at the 3% percent figure using N&C's
data, substitute their estimate of 73% for the fraction of deleterious
non-synonymous substitutions instead of Kimura's 86%.
> > > For E. coli, the estimated value for the beneficial mutation rate
> > > (Miralles et. al., 1999) was 6.4 × 1e-8 beneficial mutations per
> > > genome per generation. (1) The beneficial mutation rate obtained by
> > > Imhof and Schlötterer was 4 x 1e-9 per genome per generation (see
> > > reference links below). (2) Compare this with the detrimental mutation
> > > rate for E. coli "in excess of 0.0002" per genome per generation. (3)
> >
> > Why quote beneficial mutation rates in viruses and bacteria, when I'm
> > giving you the most recent estimates for humans?
>
> I don't see that you are giving me beneficial mutation rates for humans
> at all. The rates for bacteria, at least as far as ratios of
> detrimental vs. neutral, are comparable to eukaryotic creatures - like
> humans.
Again, that is an assertion of your own making. Neither Bustamante et.
al., nor Fay et. al. agree with you on this.
> > Rates of both
> > beneficial and deleterious mutations differ greatly between these
> > organisms.
>
> Yes, but the ratio is usually between 1 in 3,000 to 1 in 50,000.
Or, as shown, above in humans it's more like in 1 in 25 to 1 in 50.
Not a big difference, really. After all, how could anyone expect that
complex multi-cellular organisms would be any different from bacteria?
> > It's quite simple to compute these ratios in human coding
> > regions using the estimates of Bustamante et. al., coupled with N&C's
> > or Kimura's estimate of the proportion of deleterious non-synonymous
> > substitutions.
>
> Hmmmm - show me.
Open your eyes and have a look.
> > > That produces a ratio of between 1 in ~3,000 to 1 in ~50,000. But what
> > > about eukaryotes? "In sexual populations of higher eukaryotes, there is
> > > extensive data showing that U >> K." (4)
> >
> > Certainly, taking K/U = 3/100, U >> K. But that K/U ratio is still
> > much greater than 1/1000 that you were claiming, with no justification
> > from human genomic data.
>
> Note the Fay et al reference above.
Yes, they bear out my figures quite nicely, thanks.
> > > "In general, organisms with
> > > larger genomes appear to have a greater number of deleterious
> > > mutations, although it does not appear that the deleterious mutation
> > > rate is constant per base pair across these organisms." (5)
> >
> > Quoting rampantly again? This only means that the rates of deleterious
> > mutations are higher in organisms with larger genomes, not that the
> > rates of beneficial mutations don't increase in the same or even
> > greater proportions.
>
> I never meant to indicate otherwise. I fully expect the beneficial
> rates to increase in proportion, but not at significantly greater rates
> than detrimental mutations. I fully expect the ratios to be in the
> range of at least 1 in 1000. But please, do show your own
> calculations.
They are actually coming up right in the quoted text below. If you
didn't fixate on my admittedly incorrect usage of the term
"functional", you might have even noticed them.
> > > > Now, even if we were to accept
> > > > Kimura's estimate that 86% of all functional mutations in primates are
> > > > negative, we would still arrive at the conclusion that 0.09 * (1-0.86)
> > > > = 1.3% percent of all functional mutations are beneficial, whereas .86
> > > > + .135*(1-0.86)= 87.9% of all functional mutations are deleterious.
> > > > That gives a ratio of 15 beneficial to 1000 negaitve, which is 15 times
> > > > what you state. Of course, this ratio would be even bigger using N&C
> > > > estimates of the fraction of deleterious mutations (about 30 to 1000).
> > >
> > > Perhaps you've misread Kimura?
> >
> > I haven't read that particular paper by Kimura - I took the figure from
> > N&C.
> >
> > > It seems that the "86%" number is not the number of all "functional" mutations,
> > > but of nonsynonymous substitutions in a functional protein - like the hemoglobin
> > > protein used by Kimura in this particular study.
> >
> > My bad. I should have written "non-synonymous" instead of
> > "functional".
>
> It makes a very big difference. This was a complete misquote on your
> part on a vital issue here. And you are chastising me for "quote
> mining"?! Come on now . . .
Yes, but it was an honest mistake on my part, albeit a pretty stupid
one. But if you actually took time to read and - *gasp* - understand
my calculation, you would have seen that I was talking exactly about
non-synonymous rates.
> > > N&C write:
> > >
> > > "What proportion of nonsynonymous changes are neutral and what
> > > proportion are deleterious? The fraction that are neutral, fo, can be
> > > calculated by comparing the total mutation rate, µt, with the
> > > substitution rate, vo = foµt (KIMURA 1983A, KIMURA 1983B). The
> > > proportion that are deleterious is 1 - fo. Using this approach, KIMURA
> > > 1983B estimated that 86% of nonsynonymous substitutions are
> > > deleterious. . .
> >
> > Yep, that's where I got the 86% percent figure.
>
> Not that this doesn't translate into a 14% beneficial mutation rate
> like you suggested. The 14% rate is the neutral mutation rate. Big
> difference . . .
Where on earth did I suggest a 14% rate of beneficial mutations? Who
is misquoting whom now?
Rather, if you still don't understand the above calculation - I am
assuming that 86% of non-synonymous substitutions are deleterious, and
the rest are either neutral, beneficial, or deleterious but not yet
eliminated by selection. Then, relying on Bustamante's data, I am
assuming that of that 14%, 9% of the substitutions are beneficial.
That gives an overall .14 * .09 ~ 0.013. That is, of all
non-synonymous substitutions, I compute that 1.3% are beneficial (using
Kimura's estimate of 86% deleterious, which is probably too high).
That is very kind of you, and I appreciate the links. However, while
in this particular instance you've given the entire paragraph,
elsewhere you've only pasted one or two sentences that you deem
relevant - without regard to broader context. If you are not doing
this with the intention to deceive - and I don't believe you are - then
why? The only explanation I can think of is that you are hunting
through the literature for any bit of text that might support your
position, and than cut and paste that text without giving regard to the
broader context within the article, even when the next or previous
sentence already happens to contradict what you are saying. I don't
consider that type of behavior to be very intellectually honest.
> Even when
> you completely misquote someone, like you did with your whole 14%
> positive mutation rate,
Who is misquoting whom now? Look through my post and tell me where I
suggested a 14% positive percent mutation rate. I don't blame you for
thinking that is what I meant after my unfortuntate usage of
"functional", but please don't put words in my mouth.
> I at least give you the benefit of the doubt
> and don't accuse you of blatant misrepresentation - "like an
> evolutionist". Come on now? What are you trying to do here?
I am trying to point out that cherry-picking sentences from the
literature that seem to support your position without making the
slightest endeavor to understand them in their proper context will not
do wonders for your scientific credibility. Assuming, of course, you
care about such things.
> > Also, note that while we cannot estimate the rate of potential
> > deleterious substitutions in the non-coding regions, we cannot estimate
> > the rate of potential beneficial substitutions in those regions,
> > either, but there's no reason to believe that beneficial mutations
> > would be more uncommon, relatively speaking, in those regions, than
> > deleterious mutations. Regardless, estimates of beneficial to
> > deleterious ratios for the coding region of the human genome is all
> > we've got at the moment.
>
> I do believe that the ratio stays the same, but it is so outclassed
> that the only important thing is the increase in the deleterious
> mutation rate.
You are fully entitled to your beliefs, of course. But as usual, they
are not borne out by actual evidence. Life is often like that, I am
afraid.
Incidentally, I must apologetically inform you that this is likely my
last reply of any length to one of your posts. Next Tuesday, I'll be
leaving for a two-week long vacation where I'll have blissfully little
access to Internet. After I come back, I'll have to get my act
together and make one last concerted push so that, after many years, I
can finally graduate from the ineptly named graduate school. Meaning
that, for at least the next eight to ten months, I'll need to leave
behind newsgroup postings and other such pleasant distractions. So
good luck to you, and try not to fall into any neutral gaps while I'm
gone!
Cheers,
Leonid.
Richard Forrest
Seanpit wrote:
> Richard Forrest,
>
> Let me make this as clear as I can. I'm not interested in talking to
> you or anyone else about any other typics than the neutral gap problem
> right now. I can't spend my limited time on all the topics that I'm
> very much interested in. This doesn't mean I'm not interested in your
> point of view. I am. I just don't have the time to discuss your point
> of view right now. Currently, I'm more interested in other topics.
> Sorry . . .
>
> The reason why I haven't changed the stuff on my website dealing with
> your favorite topics is because I haven't found your arugments
> convincing. You may find that amazing and utterly mysterious since
> your arguments are so overwhelmingly clear to you, but they aren't
> clear to me. Perhaps someday they will be? For now though, I don't
> have the time. Perhaps some other day? I'll let you know. Ok?
It's not my arguments, Sean.
It's the evidence. It's clear and unambiguous.
You base your argument for a global flood on the assertion that fossils
are created by rapid burial.
You have conceeded here that not all fossils show signs of rapid
burial. Therefore the fossil record is not evidence for a global flood.
You have no argument.
RF
Seanpit
Richard Forrest wrote:
> It's not my arguments, Sean.
> It's the evidence. It's clear and unambiguous.
>
> You base your argument for a global flood on the assertion that fossils
> are created by rapid burial.
>
> You have conceeded here that not all fossils show signs of rapid
> burial. Therefore the fossil record is not evidence for a global flood.
>
> You have no argument.
>
> RF
Obviously, I disagree. Your notion of how a global flood/catastrophe
should act is rather odd. During a very complex global catastrophe, to
include many centuries of catastrophic aftermath, there most certainly
would be regions of calm interspersed with catastrophe. The geologic
column and the fossil record, taken as a whole, are far more consistent
with massive closely spaced catastrophes than with relatively slow
accumulation over millions and even billions of years.
I find none of your arguments to be inconsistent with this statement.
Sean Pitman
www.DetectingDesign.com
Seanpit
John Harshman wrote:
> >>This is not "discussing" flood geology. It is simply a blanket denial
> >>of any contrary evidence.
> >
> > Again, that is certainly your opinion. I find it most interesting then
> > that you continue to want to talk to someone who you think incapable of
> > true "discussion" ; )
>
> Nobody says you are incapable, just that you haven't done it so far. We
> retain hopes that some day you will.
LOL - you are just too hilarious John! ; ) It seems to me that only
those who agree with you are capable of even a basic "discussion"?
Really now. At least I extend you this courtesy from my own
perspective. I do not say that you have yet to really "discuss"
anything. Isn't that a rather backhanded thing to say? But hey, if
that's how you feel, that's how you feel.
> > Regardless, I'm not interested in talking to you about your notions
> > about what you think I've gotten wrong on my website.
>
> You aren't concerned with getting your claims right? That seems odd.
I think my claims are right and that yours are wrong. Now, I might
still be wrong in this and you might also be wrong. It is possible you
know. Just because I don't have the time to talk to everyone about
everything doesn't mean I'm not interested in improving. I'm just more
interested in other things right now. If you want to talk about those
topics, the topics I'm currently interested in "discussing", great. If
not, I suggest you go elsewhere for now. I do thank you for your
concern and your efforts to try to improve my understanding though.
That's very thoughtful of you.
Sean Pitman
www.DetectingDesign.com
Richard Forrest
Seanpit wrote:
> Richard Forrest wrote:
>
> > It's not my arguments, Sean.
> > It's the evidence. It's clear and unambiguous.
> >
> > You base your argument for a global flood on the assertion that fossils
> > are created by rapid burial.
> >
> > You have conceeded here that not all fossils show signs of rapid
> > burial. Therefore the fossil record is not evidence for a global flood.
> >
> > You have no argument.
> >
> > RF
>
> Obviously, I disagree. Your notion of how a global flood/catastrophe
> should act is rather odd.
I have no notion at all of how a global flood would act. The whole
concept is ridiculous at so many levels that I cannot imagine how one
could occur except by miraculous means. If you are invoking miracles to
explain the global flood, fine - you are not presenting a scientific
argument, and you can believe it if you chose.
> During a very complex global catastrophe,
...quite unlike the one described in the Bible, incidentally: that was
a pretty simple flood. It rained a lot, the water level rose, and after
a few months receeded again....
> to
> include many centuries of catastrophic aftermath,
...which curiously was not reported in the Bible, the source on which
you base your belief in a global flood. The biblical account tells us
that as soon as the flood was over, Moses was commanded to go forth and
multiply. Not very nice of God if there were centuries of "catastophic
aftermath" to the flood.
But perhaps you haven't read the Biblical account.
> there most certainly
> would be regions of calm interspersed with catastrophe.
IF there was a global flood, who knows what the effects would be? Our
knowledge of basic physics would be so violated by such an occurence
that we would have to rethink most of what we know about the physical
sciences.
> The geologic
> column and the fossil record, taken as a whole, are far more consistent
> with massive closely spaced catastrophes than with relatively slow
> accumulation over millions and even billions of years.
The chalk of southern England, and which also covers wide swathes of
northern Europe is up to 1000 feet thick, and composed almost entirely
of the fossils of tiny living organisms.
Please explain how this could have been laid down by a massive
catastophe?
The beds of the Westbury formation of the Lower Lias show a record of
climatic changes driven by astronomical rhythms which show that they
were laid down over a period of hundreds of thousands of years. The
same patterns are found in the Kimmeridge Clay, where they represent a
time-span of at least two million years. The same patterns are found in
the Oxford Clay, and several other formations.
These cycles are based on pertubations of the axis of the earth as it
orbits the sun. If these cycles, which match closely the measurements
of the orbit of the earth made today were created over the period of a
few centuries you assert, we may as well abandon Newtonian graviation
along with the rest of physics.
>
> I find none of your arguments to be inconsistent with this statement.
>
So you you deny that astronomical cycles exist, or do you have some
reason to think that Newton and all the physicists who followed him go
it completely and utterly wrong?
You may not find my arguments "inconsistent with" your statement. I
doubt that anyone else agrees with you.
> Sean Pitman
> www.DetectingDesign.com
Ernest Major
Seanpit <seanpi...@naturalselection.0catch.com> writes
>
>Richard Forrest wrote:
>
>> It's not my arguments, Sean.
>> It's the evidence. It's clear and unambiguous.
>>
>> You base your argument for a global flood on the assertion that fossils
>> are created by rapid burial.
>>
>> You have conceeded here that not all fossils show signs of rapid
>> burial. Therefore the fossil record is not evidence for a global flood.
>>
>> You have no argument.
>>
>> RF
>
>Obviously, I disagree. Your notion of how a global flood/catastrophe
>should act is rather odd. During a very complex global catastrophe, to
>include many centuries of catastrophic aftermath, there most certainly
>would be regions of calm interspersed with catastrophe. The geologic
>column and the fossil record, taken as a whole, are far more consistent
>with massive closely spaced catastrophes than with relatively slow
>accumulation over millions and even billions of years.
Fossil coral reefs? Fossil coal swamps stacked one on top of the other?
Fossil evaporite deposits? Fossil sand dunes? Thousands of feet of rocks
composed of the skeletal parts of plankton? Interbedded lavas and ash
falls, with fossil soils on the upper surfaces? Salt domes and other
sedimentary diapirs? Intrusive igneous rocks? Intruding other igneous
rocks? Green River Shales?
>
>I find none of your arguments to be inconsistent with this statement.
>
>Sean Pitman
>www.DetectingDesign.com
>
--
alias Ernest Major
--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.371 / Virus Database: 267.14.0/203 - Release Date: 15/12/2005
john.1...@gmail.com
Basically, Sean Pitman is a crank when it comes to geology. He has been
shown
time and again how increadibly untenable the notion of the flood is,
and yet, he
persists in his denial. I stomped him rather thorougly about a year ago
with
regard to his rediculous notions of geology. He switched context, and
finally
ran away, and has not gone back to the topic.
Now he is pounding his neutral gap hypothesis. If he has such a
convincing discussion,
I say that he should write it up, and submit it to a journal. And, also
I would suggest that
he be a man about it, and post the reviewer replies here so we can see
what actual
experts in the field have to say.
>
> Sean Pitman
John Stockwell | jo...@dix.Mines.EDU
Center for Wave Phenomena (The Home of Seismic Un*x)
Colorado School of Mines
Golden, CO 80401 | http://www.cwp.mines.edu/cwpcodes
voice: (303) 273-3049
R. Baldwin
news:1134719085.3...@f14g2000cwb.googlegroups.com...
> Seanpit wrote:
>> RobinGoodfellow wrote:
>>
[megasnip]
Since Sean is modeling his sequence space in the space defined by amino acid
codons, and in the space defined by nucleotides, how do the substution rate
numbers you two are discussing translate into his model?
Ron O
It is pretty bad for him. The more mutations happening the more his
neutral sequence space doesn't matter. He knows at some level that a
population at equilibrium for mutation selection balance has such a
great diversity of genetic background that 2 and probably 3 neutral
steps are no problem to cross. He made this calculation himself. He
has retreated to what he considers to be large enough neutral gaps, but
he doesn't have any evidence that they exist. He hasn't even
documented where something had to cross a neutral gap of three
substitutions before selection could come into play. Sad but true.
Just ask him for his verifiable example of needing to cross a gap of
three mutations. He used to claim that beta galactosidase was an
example, but it backfired on him because the original work crossed the
gap in one mutation, with two giving higher activity. The gap was even
crossed in multiple species and not just in E. coli. He tried to make
a big deal about the fact that if you took out the protein that could
be mutated that single or double mutations were more difficult to come
by, but evolution doesn't care about every species being able to do
these things, and he had to admit that single clonal colonies are far
from mutation selection equilibrium. He never did explain how you
could get beta gal activity evolving in other species of bacteria if it
was really as difficult as he claimed. I think his own references
claimed that you could jump his improbable gap in at least three
species, and I bet that they didn't try more than 10 or 20.
Ron Okimoto
Glenn
That's not even convincing with more familiar subjects. Andolfatto
seemed rather clear, that they are more difficult to find if they are
buried in junk, not because they are rare. If he had thought that the
results of his study only seemed to him to be relevant to the subjects
of his research and not humans or organisms in general, it occurs to
me that he would have said so, or at least made that more clear,
instead of seemingly implying that they would be difficult to find if
buried in junk.
>
> > The argument that
> > removing parts of DNA and seeing no ill effects provides some
evidence
> > that those parts were not under selective pressure is not
convincing
> > to me. Are there others?
>
> Yes. Andolfatto's findings are based on testing rates of divergence
> between species. If the rate is much slower than the neutral rate,
he
> supposes that stabilizing selection accounts for this. If it's much
> higher, he supposes that positive selection accounts for this. So if
you
> accept his conclusions you need to accept his methods. If you try
these
> methods on the human genome you find no such departures from
neutrality.
> (The news article implies that his tests are more sensitive than
> previous ones, but on the surface it looks like a form of the HKA
test,
> and Andolfatto was a student of the K in that test, so I wonder. HKA
and
> similar tests have certainly been performed on human sequences.)
So you don't know if what he employed was similar or even based on
previous methods, yet you make a definitive statement about what you
would find on humans/apes.
>
> Let's not confuse "junk DNA" with "non-coding DNA", though. It's
been
> known for a long time that many non-coding sequences are functional.
> They just do not, in the human genome at least, compose a large
fraction
> of the whole. That's based on the same sorts of test that Andolfatto
used.
I prefer the definition that seems the common sense usage:
http://www.sciencedaily.com/upi/index.php?feed=Science&article=UPI-1-20051019-14451400-bc-us-junkdna.xml
"Junk DNA is so-called because it doesn't contain instructions for
protein-coding genes and appears to have little or no function."
John, if you do not know a sequence has a corresponding function, how
do you know what to look for? Simple sequence matches?
>
> I can think of a number of potential objections, though I don't know
if
> he already thought of them. If recombination rates are low enough,
> selection on coding or other regions under obvious selection can
> influence frequencies at sites far from the directly affected site.
It's
> also possible that the sites he used to estimate the neutral rate
are
> themselves under positive selection, though I have trouble seeing
how
> that could be. However, it's more likely that the sites used to
estimate
> that rate are under weak negative selection, and that might explain
some
> of the positive selection results; i.e. those are really the neutral
> sites, and his clock is slow. There are previous papers that show
silent
> sites in exons to be under weak negative selection.
>
> Interesting result, though, and I wonder what it means. How are the
> conserved sequences distributed, and what do they do? Ditto for the
> positively selected ones.
>
From the link above:
"Sequencing of the complete genome in humans, fruit flies, nematodes
and plants has revealed the number of protein-coding genes is much
more similar among these species than expected," he said. "Curiously,
the largest differences between major species groups appear to be the
amount of 'junk' DNA, rather than the number of genes."
Do you disagree?
RobinGoodfellow
I am not sure if they do at all: that is, I don't think the two topics
are related. First, in his calculations, Sean skips the codon part
completely, and defines sequence space directly in terms of amino acid
sequences - which, in itself, is a reasonable initial simplification of
the problem. (Pretty much every study out there seeking to explore the
relationship between protein sequence and structure, for example, takes
a similar appoach.) More importantly, since humans and chimps diverged
from one another fairly recently in our evolutionary history - or, if
we were take the ID view, since the Designer has made us closely in the
chimps' image - the observed amino acid polymorphisms between the
coding sequences of humans and chimps represent only a very minor
subset of all possibilities for encoding the proteins in question. And
that overall set of possibilities for encoding a protein is what Sean
is ultimately after, because he believes that knowing the size of this
set is all that is needed to obtain an estimate of the time that
evolution needs to "find" one sequence in that set, and hence to find
any sequence at that overall "level of complexity". At least that's
how I understand him.
Of course, substitution rates (not to be confused with mutation rates)
are measured based on the premise that humans and other primates
diverged from a common ancestor several million years ago. I can't
quite put my finger on it, but it seems to me that this premise is
somehow inconsistent with the rest of Sean's views.
Cheers,
Leonid.
Glenn
That's not even convincing with more familiar subjects. Andolfatto
seemed rather clear, that they are more difficult to find if they are
buried in junk, not because they are rare. If he had thought that the
results of his study only seemed to him to be relevant to the subjects
of his research and not humans or organisms in general, it occurs to
me that he would have said so, or at least made that more clear,
instead of seemingly implying that they would be difficult to find if
buried in junk.
>
> > The argument that
> > removing parts of DNA and seeing no ill effects provides some
evidence
> > that those parts were not under selective pressure is not
convincing
> > to me. Are there others?
>
> Yes. Andolfatto's findings are based on testing rates of divergence
> between species. If the rate is much slower than the neutral rate,
he
> supposes that stabilizing selection accounts for this. If it's much
> higher, he supposes that positive selection accounts for this. So if
you
> accept his conclusions you need to accept his methods. If you try
these
> methods on the human genome you find no such departures from
neutrality.
> (The news article implies that his tests are more sensitive than
> previous ones, but on the surface it looks like a form of the HKA
test,
> and Andolfatto was a student of the K in that test, so I wonder. HKA
and
> similar tests have certainly been performed on human sequences.)
So you don't know if what he employed was similar or even based on
previous methods, yet you make a definitive statement about what you
would find on humans/apes.
>
> Let's not confuse "junk DNA" with "non-coding DNA", though. It's
been
> known for a long time that many non-coding sequences are functional.
> They just do not, in the human genome at least, compose a large
fraction
> of the whole. That's based on the same sorts of test that Andolfatto
used.
I prefer the definition that seems the common sense usage:
http://www.sciencedaily.com/upi/index.php?feed=Science&article=UPI-1-20051019-14451400-bc-us-junkdna.xml
"Junk DNA is so-called because it doesn't contain instructions for
protein-coding genes and appears to have little or no function."
John, if you do not know a sequence has a corresponding function, how
do you know what to look for? Simple sequence matches?
>
> I can think of a number of potential objections, though I don't know
if
> he already thought of them. If recombination rates are low enough,
> selection on coding or other regions under obvious selection can
> influence frequencies at sites far from the directly affected site.
It's
> also possible that the sites he used to estimate the neutral rate
are
> themselves under positive selection, though I have trouble seeing
how
> that could be. However, it's more likely that the sites used to
estimate
> that rate are under weak negative selection, and that might explain
some
> of the positive selection results; i.e. those are really the neutral
> sites, and his clock is slow. There are previous papers that show
silent
> sites in exons to be under weak negative selection.
>
> Interesting result, though, and I wonder what it means. How are the
> conserved sequences distributed, and what do they do? Ditto for the
> positively selected ones.
>
John Harshman
Feel free to believe anything you like.
Yes. You understand that in closely related species, like human and
chimp, this is a no-brainer. Corresponding sequences differ only by a
few bases in a thousand.
No. This is common knowledge.
R. Baldwin
>
> R. Baldwin wrote:
>> "RobinGoodfellow" <lmey...@gmail.com> wrote in message
>> news:1134719085.3...@f14g2000cwb.googlegroups.com...
>> > Seanpit wrote:
>> >> RobinGoodfellow wrote:
>> >>
>> [megasnip]
>>
>> Since Sean is modeling his sequence space in the space defined by amino
>> acid
>> codons, and in the space defined by nucleotides, how do the substution
>> rate
>> numbers you two are discussing translate into his model?
(I left out a word above- it should have read "and not in the space defined
by nucleotides").
Thanks. I had leaped to the conclusion that, because Sean had started a new
thread from the middle of an old one, it must have been related to his
"neutral gaps" model. Sean had several times recently stated that he was
only interested in discussing his "neutral gaps" model, which helped confirm
my conclusion. I was curious why he would discuss a mutation rate at all
when his model ignores silent mutations.
I see now that this actually came from the "Horse and Donkey" thread, which
I wasn't following. I'm not going to read it, because it is too old and too
long, so I'll give up my notion that Sean might be confusing mutation rate
with substitution rate.
RobinGoodfellow
Don't be so hasty! If you read upthread - specifically, the post that
you originally responded to - you'll see that the good doctor appears
to have done exactly that.
Seanpit
> RobinGoodfellow wrote:
> > Seanpit wrote:
> > > RobinGoodfellow wrote:
> >
> > < snip >
> >
> > > You seem to give a lot of weight to the fact that N&C seem to think
> > > that the deleterious mutation rate is too high, and not at all to the
> > > fact that they seem to think that humans and chimps diverged from a
> > > common ancestor - even though they use the latter as a premise to
> > > arrive at the former conclusion (the Ka/Ks ratios). Not too big on
> > > consistency, are we?
> >
> > If U really is greater than 3, all slowly producing animals, to include
> > both humans and apes, are headed for extinction.
>
> That is your understanding, yes. Very few evolutionary biologists,
> even those who accept the U>=3 rates, seem to share it. I wonder why?
This doesn't mean that evolutionary biologists can explain how
detrimental mutations that enter the gene pool at this rate (U > 3) can
actually be cleared in a way that does not lead to extinction. And,
many of them, to include Crow himself, actually admit this. They
really don't know how to explain it other than to say that some sort of
epistasis must be involved. But, when you really get down to it, none
of the proposed forms of positive or buffering epistasis or even sexual
recombination, or both, really solve the problem when U is so high.
> > If U is significantly
> > less than 3, then there is something wrong with the assumptions of
> > human-ape common ancestry.
>
> I have no idea how you came to that conclusion. Nothing about the
> common ancestry of humans and apes implies/requires a specific
> *deleterious* mutation rate.
That's how U was calculated; right? - By assuming human-ape common
ancestry? If the real value of 3 is in fact significantly less than 3,
then wouldn't there be something wrong with the human-ape comparisons
that return results greater than 3?
> > Either way, something's not adding up. How
> > is this consistency from your own side?
>
> Well, given the fact that both points you bring up are wrong, I'd say
> the consistency of my own side doesn't suffer very much.
If U really is greater than 3, then there seems to be a big mystery
that your side has yet to explain. If U really is significantly less
than 3, the whole notion of being able to calculate U from human-ape
comparisons seems just a bit screwing since human-ape comparisons
really do indicate that the value for U is probably greater than 3 -
like even greater than 5.
> > Beyond this, there is good real time evidence that humans do indeed
> > sustain upwards of at least 175 mutations per person per generation.
> > Oh, and by the way, my use of "29 mitotic divisions" before
> > reproduction of the next generation is a male to female *average*. The
> > average for woman is about 23, while the average for men is a good bit
> > higher, creating an overall average of about "29". Was this really all
> > that hard to understand from the way I wrote my essay?
>
> You are the one who is not understanding me, Sean.
First off, Leonid, you did question why I would put "23" as my estimate
for the number of mitotic divisions for women and then use "29" as my
overall estimate even though it was obvious from my essay that the "29"
estimate was a combined average of both woman and men together.
Beyond this question of yours, I do understand what you are talking
about in the following:
> I am talking about
> estimates of *deleterious* mutation rates specifically, which relies on
> the assumption that deleterious non-synonymous substitutions have been
> eliminated by selection since humans and other primates split from one
> another. Here's a paragraph you didn't address from the "Horse and
> Donkey" thread:
>
> "No matter. Assuming you have no problem with their computation of the
> overall mutation rate, how do you reconcile their computation of the
> deleterious mutation rate with your young life position? The
> deleterious fraction is estimated as 1-Ka/Ks, where Ka/Ks is the
> ratio of the number of non-synonymous to synonymous substitutions.
> However, if you think life was created 10,000 years ago, you would need
> to assume that the overwhelming majority of both synonymous and
> non-synonymous substitutions were deliberately designed into chimp and
> human genomes (unless you want to posit a past mutation rate
> approximately 600 times higher than currently estimated). Thus, if
> both synonymous and non-synonymous substitutions are designed, we
> cannot use 1-Ka/Ks to estimate the deleterious fraction. (I suppose
> you could use Ka/Ks estimates from human genes only, but I can't think
> of any such studies off the top of my head.)"
I do believe that many of the synonymous and non-synonymous differences
between different "kinds" of creatures, like humans and apes, are based
on original design - tailored according to the different needs of
different kinds of creatures. That is why many of these functional
differences are maintained over time and cannot therefore be used as
any sort of molecular clock to estimate times of divergence.
Functional differences are maintained by natural selection over time as
"different" because of different functional needs.
> And a follow-up from me, a little later in the thread:
>
> "It occurred to me that there is plenty of human SNP data out there to
> help make such a determination. Unfortunately, it doesn't help you.
> Latest result seem to indicate that among humans, Ka/Ks rates are
> pretty close to 1, meaning that they can't really be used to determine
> deleterious mutation rates in a manner suggested by N&C. (See, e.g.
> Genome Research, 14:2034-40). In order to compute these rates, you
> have to start with idea that man and other primates diverged through
> evolution. Tough luck, Sean."
I don't believe in the basis all of the N&C estimates. I'm just
arguing that the results of these estimates are inconsistent with the
ToE itself.
> > Now, if each human individual sustains 175 mutations per generation,
> > how many of these will hit the "functional" area of the genome?
> > Consider the following discussion by Fay et al published in the July
> > 2001 issue of Genetics:
> >
> > "The genomic deleterious mutation rate in humans was previously
> > estimated to be at least 1.6 on the basis of an estimate that 38% of
> > amino acid mutations are deleterious. The genomic deleterious mutation
> > rate is likely much larger given our estimate that 80% of amino acid
> > mutations are deleterious and given that it does not include
> > deleterious mutations in noncoding regions, which may be quite common
> > (SHABALINA and KONDRASHOV 1999).
>
> And how do you think the estimate of 80% amino acid mutations being
> deleterious is arrived at? By computing the ratio of non-synonomous to
> synonymous substitutions, and assuming that deleterious non-synonymous
> substitutions have been eliminated from ancestral populations by
> selection. The sort of assumption you can't be making, since you can
> never know which substitutions were "designed" into the genomes and
> which weren't at the time humans and chimps were separately created.
Again, I'm just showing that these calculations are inconsistent with
the ToE. These are not my calculations. These calculations come from
your side of this issue.
> > The combined NC*/S* ratio of common
> > SNPs from both surveys indicates 50% of the noncoding sites are
> > constrained and must serve some function. Because an equal number of
> > noncoding and amino acid-altering sites were surveyed, noncoding
> > mutation should contribute at least 60% (0.50/0.80) as much as coding
> > mutations to the total genomic deleterious mutation rate."(1)
> >
> > So, where is your notion that U=1.6 now? These figures put U well over
> > 5. As far as the beneficial mutation rate, Fay et al go on to say:
> >
> > "The slightly deleterious fraction, f1, cannot exceed 0.80 since 20% of
> > mutations were estimated to be neutral"
>
> Meaning that, allowing that some of the substitutions were the result
> of positive selection, the slightly deleterious fraction is somewhat
> less than 80% percent. 80% is an upper bound.
You do realize that the authors of these papers you quote actually
assume a ratio of Ud/Ub of 1000 to 1? (see below) . . .
> > "... Extrapolating this
> > proportion to the total amount of coding DNA in the genome (~5 x 10^7
> > bp) yields an estimate of up to 1 advantageous substitution every ~200
> > years since humans separated from old world monkeys 30 million years
> > ago."
>
> Quote mining again, are we?
Oh, come on now . . . The paragraph you quote below says exactly the
same thing as the smaller portion says. It just explains how the
authors calculated this positive substitution rate. The rate itself,
however, means exactly what I said it means. There's no "quote mining"
going on here Leonid. The intent of the authors is preserved in what I
quoted.
Perhaps it is you who doesn't quite understand the implications of what
the authors said? Given this potential misunderstanding of yours, you
accuse me of quote mining - thinking that either I don't understand
what the authors are saying or that I'm deliberately quote mining to
support something I already believe without considering the true
meaning of what the authors are trying to say? What about the
possibility that you don't understand what the authors are saying in
this passage?
> Let us restore the full paragraph:
Lets . . .
> "The large number of amino acid substitutions suggests a high rate of
> adaptive evolution in primates. The expected number of amino acid
> substitutions is 2382 (4151 x 70/122) based on the A/S ratio of common
> polymorphism and the excess is 1278. Therefore, a large proportion,
> ***35%***, of amino acid substitutions between humans and old world
> monkeys ***are estimated to have been driven bypositive selection***.
> Extrapolating this proportion to the total amount of coding DNA in the
> genome (~5 x 107 bp) yields an estimate of up to 1 advantageous
> substitution every ~200 years since humans separated from old world
> monkeys 30 million years ago (LI 1997 )." [Emphasis mine]
Quoting the 35% number is not the "rate" of positive mutations - not
even close. It is the ratio of substitutions that are currently
thought to be the result of positive selection. That is not the same
thing as the rate at which positive mutations occur. The rate of 1
advantageous substitution every 200 years is also *not* even close to
the rate at which positive mutations occur per individual per
generation - despite your suggestion to the contrary.
> Now, you did provide the link to the article, and by now, you must have
> assumed I would read it. What were you trying to accomplish by
> omitting the highly relevant context above?
I fail to see the significant difference between the paragraph you
quoted and the smaller portion I quoted. The final outcome is the same.
The true position of the authors is preserved even in the smaller
quotation. Nothing was changed. It seems that the only problem is your
interpretation of what the authors were trying to say in this passage.
> Incidentally, the quote, even without the context, does not mean what
> you think it means.
Oh really? Lets see . . .
> Substitution rates are measured *per population*
> per generation, whereas mutation rates are measured *per individual*
> per generation.
Exactly . . . but keep going.
> Under the neutral model, the two quantities are equal:
Right! But, they are not equal if the mutation is not neutral. If the
mutation is functional, the substitution rate will not equal the actual
mutation rate. You base your calculations on the assumption that the
functionally beneficial substitution rate does in fact closely equal
the beneficial mutation rate. Perhaps this assumption of yours isn't
quite right?
> within the context of slightly beneficial mutations, substitution rates
> will be slighly higher than the mutation rate.
It seems to me that this is where you goof up. In the context of high U
values, the beneficial mutations that actually make it to fixation will
not be those that are just slightly beneficial. Therefore, the positive
substitution rate is not going to be "slightly" higher than the actual
positive mutation rate. The substitution rate would be a whole lot
higher. Beyond this, consider that population geneticists, like
Whitlock, assume a detrimental vs. beneficial mutation estimate of at
least 1000 to 1:
"Given a reasonable value of λd of about −0.02 (Lynch and Walsh
1998; but see Keightley 1994; García-Dorado 1997) and guesses of the
values of λb and Ud/Ub of 0.02 and 1000, respectively, the critical
value of the effective size would be about 125."
Michael C. Whitlock, Fixation of New Alleles and the Extinction of
Small Populations: Drift Load, Beneficial Alleles, and Sexual
Selection, Evolution, Vol. 54, No. 6, pp. 1855–1861. -
http://evol.allenpress.com/evolonline/?request=get-document&issn=0014-3820&volume=054&issue=06&page=1855
Beyond this, real time Ud/Ub rates, obtained from studies of bacteria
and other microorganisms, have returned ratios of between 3,000 and
50,000 to 1. You counter by arguing that humans and microorganisms
might be a whole lot different with respect to such ratios.
Personally, I don't see why this should be. I do understand how the
overall mutation rates might be a whole lot different, but I don't
really understand why the *ratio* of Ud/Ub would be significantly
different.
Now, before you go off and argue that Whitlock demonstrated that only a
relatively small population of just a few thousand is needed to
overcome a ratio of Ud/Ub of 1000 to 1, consider that Whitlock did not
take very low reproductive rates into consideration in his calculation
of R-best (Ne). He also did not consider the fact that higher Ud rates
significantly reduce the Ub rate (see below).
Note also that the Y-gene in human males does not undergo recombination
during sexual reproduction. According to your own position, then, the
information within the Y-chromosome is indeed caught in the "Living
Dead". Positive epistasis doesn't help when genetic information is
reproduced asexually. So, even by your own theory, the male line of
the human race is headed for extinction. Women will end up taking over
the world! - - which I'm sure isn't a problem for most women. ; )
And, by that time, I'm sure they'll have found a way to reproduce
without us men. The technology is already here anyway - just a bit
more polishing is all.
> So, a rate of 1
> substitution every 200 years, is the rate of 0.125 substitutions per
> population per generation, assuming 25 year generations. Which means
> that the rate of slightly beneficial mutations will be slightly lower
> than 0.125 per individual per generation.
Actually, the rate of slightly beneficial mutations will be quite a bit
lower. The substitution rate of 0.125 per population per generation is
based on a concentration of more than slightly beneficial mutations in
the larger population into a smaller subpopulation over time with many
of the more slightly beneficial mutations being lost before fixation.
Out of all the millions of beneficial mutations that might have been
realized in a given generation, only 0.125 make it to fixation. This
means that the actual beneficial mutation rate is much much lower than
the beneficial substitution rate.
> Taking U=3, the ratio of
> beneficial to deleterious mutations is about 0.125/3, or slightly over
> 4 percent. Which is one percent higher than I originally estimated.
> Thanks, Sean!
Again, this calculation of yours assumes that the beneficial
substitution rate is only "slightly" higher than the actual beneficial
mutation rate. This is mistaken since the beneficial substitution rate
is far higher than the actual beneficial mutation rate.
> Here's a link on how substitution rates are computed from mutation
> rates, within the context of the neutral model:
http://www.stat.berkeley.edu/users/terry/Classes/s260.1998/Week13a/week13a/node10.html
This linked discussion assumes neutral mutations. It doesn't calculate
the difference between the actual beneficial substitution rate and the
actual beneficial mutation rate - which are very different things.
The beneficial substitution rate is affected by many factors to include
the population size, the rate of detrimental mutations, and the form of
positive epistasis actually in play.
"Deleterious mutations affect drastically the dynamics of fixation
of the beneficial mutations. As previously demonstrated (Campos 2003;
Peck 1994), the probability of fixation (Pfix) of advantageous mutants
is a decrease function of the mutation rate U. This is a consequence of
the possibility of occurrence of such beneficial mutation in a genome
with a large amount of segregated deleterious mutations, i.e., in a
genome with very low fitness value . . . Figure 1: Probability of
ultimate fixation as a function of the mutation rate U. The parameters
are N = 1000, = 10 and sd = 0.1. The data points correspond to the
simulation results over 100,000 runs, and the thick line is the
theoretical prediction according to Eq. 11." (1)
In this figure, you will notice that as U approaches 1, the rate of
fixation of beneficial mutations drops off dramatically. When U =
0.001, Pfix = ~0.16. But, when U = 1, Pfix is less than 0.02. What do
you think happens to Pfix when U is over 3? - or over 5? The authors
continue:
"For high mutation rate U the agreement between the simulations and
the theoretical prediction is less satisfactory than those seen for
small and intermediate values of U. This problem occurs due to the
occurrence of the Muller’s ratchet phenomenon in finite populations,
at which the continuous accumulation of deleterious mutations leads to
loss of the best adapted classes of individuals as the population
evolves. Thus, for very high U the population never reaches the
equilibrium regime as supposed in the theoretical formulation (Campos
2003)."
… Only those mutations awarding a large beneficial effect to the
individuals have a non-negligible chance to reach fixation." (1)
Oh, but what about sexual vs. asexual populations?
"Relative to an asexual population, increased levels of
recombination reduce the effects of beneficial mutations more rapidly
than those of deleterious mutations. However, the role of beneficial
mutations in determining the evolutionarily stable mutation rate may
still be significant if the function describing the cost of
high-fidelity replication has a shallow gradient." (2)
In other words, if the cost function is very shallow, some beneficial
mutations might just have a chance of making it to fixation in a
sexually reproductive population.
Consider the following discussion of this problem:
"When recombination is absent, mutations are trapped in their
original genetic background [background trapping; see (7–14)]. In
this case, the effect of the genetic background on organismal fitness
persists and interferes with direct selection on a mutation. The only
mutations that will ultimately be fixed in a population are those that
originated by chance in genotypes whose fitness, which results from the
combined effects of the new mutation and the background, is high
(7–14); i.e., they must originate in the “Progenitor tail” shown
in Fig. 1B. All other mutations are eliminated deterministically
because they are trapped in lower-fitness lineages (genetic
backgrounds) that are destined to eventual extinction (the “Living
Dead,” Fig. 1B) (10). . . . Overall, theory predicts that
recombination increases the realized strength of selection because it
prevents background-trapping and thereby reduces the dilution of direct
selection by background selection. This prediction can be tested
equally well in the context of the accumulation of beneficial mutations
or the removal of deleterious mutations because both are a consequence
of the same phenomenon—reduced interference between direct versus
background selection. . . . . In the absence of recombination, most new
beneficial mutations are trapped in the living dead (Fig. 1B) when the
genetic standard deviation in fitness, sW(gen), is large relative to
the selection coefficient s. As a selective sweep of lineages derived
from the progenitor tail progresses, the net strength of selection (42)
declines to zero for beneficial mutations trapped in the living dead,
and ultimately becomes negative. In contrast, the net strength of
selection for a segregating beneficial mutation is always positive in a
recombining population. . . . "(6)
So, it seems that beneficial mutations in non-R-best (Ne) individuals
are only capable of contributing to the R-best population in sexually
reproducing individuals. So, does the reproductive rate play any part
in maintaining the number of R-best in a population? And, if so, how
much of a part does the reproductive rate play?
1. http://arxiv.org/PS_cache/q-bio/pdf/0310/0310006.pdf
2. http://www.genetics.org/cgi/content/full/151/4/1621
3. http://homepage.univie.ac.at/Reinhard.Buerger/04WhitlockBuerger.pdf
5.
http://www.springerlink.com/media/9f52hyxwvk0xpmkged4t/contributions/r/4/k/3/r4k343448275070w.pdf
6.
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/50.pdf
7. http://www.christianforums.com/t1155768-the-quiet-thread.html&page=4
> Now, how can you expect to discuss these topics and be taken seriously
> when you don't even have rudimentary knowledge of basic population
> genetics concepts?
I fail to see where you have shown that high detrimental mutation rates
in slowly reproducing populations can be overcome by positive epistasis
or beneficial mutation rates that are most likely lower than 1 per 1000
detrimental mutations at best. Given your vast knowledge of how
population genetics works, where have you or anyone else solved the
problem?
> > And you think that is an adequate enough fixation rate to compensate
> > for a U of over 5?
>
> OK. For U=5, the ratio of beneficial to deleterious is about 2.5%,
> taking Fay et. al. figures. Yeah, I think that's high enough.
Well, it still seems to me that your calculations here are mistaken.
The ratio of Ud/Ub is still most likely less than 1 in 1,000 - at best.
> > 1. Justin C. Fay, Gerald J. Wyckoff, and Chung-I Wu, Positive and
> > Negative Selection on the Human Genome, Genetics, Vol. 158, 1227-1234,
> > July 2001 -http://www.genetics.org/cgi/content/full/158/3/1227
> >
> > > But, to answer your point, N&C had no way to assess the impact of
> > > non-coding DNA on U. They could only compute U for coding DNA, and
> > > those estimates vary considerably depending on their different
> > > estimates of overall mutation rates.
> >
> > Ah, but N&C did comment that their estimate of U was likely biased
> > downward because of this problem. You just choose to overlook that?
>
> Of course not. But N&C don't say how far downward, do they? I choose
> not to piss in the wind by trying to extrapolate to data that isn't
> there.
Actually Crow did suggest that U could easily be over 5.
> > > N&C's lower bound estimate of U
> > > is 1.5, and they site another estimate of 1.6 (from 1999, By
> > > Eire-Walker and Keightley) that agrees with that lower bound. I'm not
> > > saying that is necessarily the correct estimate either, but your desire
> > > to accept U>=3 suggests that you've already made up your that we are
> > > devolving towards extinction, and are cherry-picking data to support
> > > this position.
> >
> > I'm not cherry picking to support my position. I'm just stating what
> > the author's themselves believed to be the most likely estimate for U.
> > They are the ones who said that U is most likely greater than 3, even
> > greater than 5, not me. Rather, it seems like you are the one who is
> > "cherry-picking" the extremes of the author's data set in an effort to
> > support your views. The most recent data, according to the authors,
> > not me, clearly supports U values of well over 5.
>
> My point is, these rates are notoriously hard to estimate, and vary
> widely between the loci being studied and the assumptions being made.
> Until we have whole-genome studies assessing these rates, I'm unwilling
> to believe any of these estimates: or rather, equally willing to
> believe them all. And none of these numbers, by themselves, imply that
> we or the chimps are headed anywhere toward extinction.
If you do accept that the most likely value for U is over 3, then this
does certainly indicate that all slowly reproducing creatures, like
humans and apes, are indeed headed for extinction.
> > > > In fact, the odds are most likely a
> > > > whole lot higher than U=3 now that a lot more of what was thought to be
> > > > "junk" DNA, because it doesn't code for proteins and whatnot, is no
> > > > longer junk, but functional and constrained by natural selection after
> > > > all. This has lead to the suggestion that U is actually "greater than
> > > > 5" (see references below).
> > >
> > > The paper you reference by Rice you are referring to simply mentions
> > > this figure in passing, citing three other studies, the latest of which
> > > came of 1993. This pre-dates the U=1.6 estimate of Eire-Walker and
> > > Keightley by 6 years, and N&C's estimate of U=3 by 7 years.
> >
> > And why do you think Rice chose to print the estimate of U = 5? Again,
> > what do you think about the Fay paper will estimates of well over 5?
>
> I think they are plausible, but far from as certain as you would like
> to believe.
If you accept these estimates as plausible, then it seems equally
plausible that extinction is indeed inevitable.
< snip >
It seems to me that the contribution to R-best from the second best
would be less than 5 million - but this doesn't take away from your
main point here.
> So, in two
> generations later, the number best-fit individuals is over 125 million,
> rather than just 62 mil.
In 2 generations, the contribution to R-best from the lesser fit would
be around 55 million. However, in subsequent generations, there become
fewer and fewer lesser fit in each category. The most of the sand
continues to go downhill for each category with only a relatively few
grains rolling uphill in each generation. The overall current,
however, continues downhill. The basis for R-best just drains away and
R-best continues to decline over time.
> While this still shows a decline in the
> number of best fit individuals, this decline is twice as small in
> proportion to the first generation. Furthermore, as the less-fit
> classes grow, so will the proportion of all recombinant offspring
> returning to the more-fit classes, effectively lowering mutational load
> requirements. Again, I strongly encourage you to read Rice's paper.
The problem is that the fitness sand rolls downhill for each class of
offspring. In each generation each class will contribute far more to
the lower than to the upper classes. While recombination does indeed
slow this process down a bit, it doesn't stop the steady decline and
the relatively low rate of beneficial mutations is not enough to
balance out the problem.
Now, it is true that less fit individuals do in fact increase their
odds of sustaining a beneficial mutation. But, this is based on the
marked relative increase in detrimental mutations that can sustain
either reverse or compensatory mutations. The problem is that the rate
of beneficial fixes to these already detrimental changes is not great
enough to cope with the onslaught. Add to this the problem that some
types of detrimental mutations cannot be easily fixed. The gap created
by some detrimental mutations is just too great for reversion or
compensatory mutations cross - and so they remain.
"As the mean fitness of a population declines and as more and more
deleterious alleles become fixed in the population, the proportion and
mean magnitude of beneficial mutations is expected to increase. This is
because for most deleterious alleles, there exists not only a reverse
mutation but also the potential for mutations at other loci to
compensate for the loss in function."
> The important thing here is that, once recombination is present, the
> lower fitness classes are no longer members of the "living dead", since
> invidiuals from lower-fitness classes can produce offspring in
> higher-fitness classes. Given non-random mating, these rates of
> re-entry into higher-fitness can be amplified considerably.
This is true. But, at high rates of detrimental mutations, combined
with low reproductive rates, this amplification is just not enough to
keep up. Rice's own paper, which you keep telling me to read even
though I've read it several times by now, backs this concept up. He
does demonstrate a slowing of the decline with the use of various types
of epistasis as well as genetic recombination, but none of these are
able to completely reverse the trend given high Ud rates and low
reproductive rates.
< snip >
> > Exactly! The problem is that deleterious mutation rates are far more
> > likely to be increased than beneficial mutation rates. Try exposing an
> > organism to radiation and see what happens.
>
> Non-sequitur. The overall *per nucleotide* mutation rates don't need
> to be increased for both deleterious and beneficial rates *per genome*
> to increase. Human mutation rates per nucleotide per generation are
> much lower than bacterial, but our deleterious and beneficial mutation
> rates per genome per generation are much higher, due in no small part
> to our larger genomes. At any rate, how does this article contradict
> the notion that beneficial mutaitons will arise and spread more rapidly
> in larger populations, which you claimed was "not quite true"?
Actually, the mutation rate for bacteria, although highly variable,
averages about 1e-8 per bp per generation. The human per generation
rate is about the same at about 1e-8 per bp per generation. It is just
that the generation times between humans and bacteria are very much
different.
The problem is that many calculations in population genetics do not
consider the fact that high mutation rates are lethal regardless of the
population size. Calculations cannot simply consider a single ratio of
Ud/Ub because increasing overall mutation rates markedly increase Ud
relative to Ub in all classes of mutational load within a population.
Very high mutation rates in slowly reproducing individuals simply
cannot be tolerated because of this problem. Beneficial mutations
simply aren't increased nearly enough to compensate for the marked
relative increase in the detrimental mutation rate.
< snip >
> > Oh, come now. Compensatory mutations are about the most common type of
> > beneficial mutation there is. You just can't get much better than this
> > as far as rate of beneficial mutations is concerned.
>
> Even if this were true - and this is just the sort of claim you may
> want to provide a citation for - it doesn't mean compensatory mutations
> are those that contribute most to the overall fitness increases in the
> population. In fact, reverse and compensatory mutations by themselves
> can never achieve an overall increase in fitness, since all they
> accomplish is undoing the effects of previous deleterious mutations.
Exactly . . . Yet, these are the types of beneficial mutations that
are increased with decreasing fitness of a population. Other types of
beneficial mutations are relatively rare and are not increased with
decreasing population fitness.
< snip >
> > > > For E. coli, the estimated value for the beneficial mutation rate
> > > > (Miralles et. al., 1999) was 6.4 × 1e-8 beneficial mutations per
> > > > genome per generation. (1) The beneficial mutation rate obtained by
> > > > Imhof and Schlötterer was 4 x 1e-9 per genome per generation (see
> > > > reference links below). (2) Compare this with the detrimental mutation
> > > > rate for E. coli "in excess of 0.0002" per genome per generation. (3)
> > >
> > > Why quote beneficial mutation rates in viruses and bacteria, when I'm
> > > giving you the most recent estimates for humans?
> >
> > I don't see that you are giving me beneficial mutation rates for humans
> > at all. The rates for bacteria, at least as far as ratios of
> > detrimental vs. neutral, are comparable to eukaryotic creatures - like
> > humans.
>
> Again, that is an assertion of your own making. Neither Bustamante et.
> al., nor Fay et. al. agree with you on this.
Actually, they do agree with me (see above). It is just that your
assumptions, that the beneficial mutation rate is at all close to the
beneficial substitution rate, may not be correct here.
>
> > > Rates of both
> > > beneficial and deleterious mutations differ greatly between these
> > > organisms.
> >
> > Yes, but the *ratio* is usually between 1 in 3,000 to 1 in 50,000.
>
> Or, as shown, above in humans it's more like in 1 in 25 to 1 in 50.
> Not a big difference, really. After all, how could anyone expect that
> complex multi-cellular organisms would be any different from bacteria?
As far as the ratio of Ud/Ub? - I really don't see how the difference
would be all that significant. Do you? The authors of the papers you
reference simply don't support your calculations like you think they
do. Beyond this, what explanation do you have to support your assertion
of such a markedly different ratio of Ud/Ub between eukaryotes and
prokaryotes?
< snip >
> > It makes a very big difference. This was a complete misquote on your
> > part on a vital issue here. And you are chastising me for "quote
> > mining"?! Come on now . . .
>
> Yes, but it was an honest mistake on my part, albeit a pretty stupid
> one. But if you actually took time to read and - *gasp* - understand
> my calculation, you would have seen that I was talking exactly about
> non-synonymous rates.
Which is also a problem since your calculations are based on mistaken
assumptions. That is why I didn't understand your honest mistake as
being at all reasonable. It just made no sense to me so I didn't
"figure it out" that you actually made an honest mistake.
< snip >
> Incidentally, I must apologetically inform you that this is likely my
> last reply of any length to one of your posts. Next Tuesday, I'll be
> leaving for a two-week long vacation where I'll have blissfully little
> access to Internet. After I come back, I'll have to get my act
> together and make one last concerted push so that, after many years, I
> can finally graduate from the ineptly named graduate school. Meaning
> that, for at least the next eight to ten months, I'll need to leave
> behind newsgroup postings and other such pleasant distractions. So
> good luck to you, and try not to fall into any neutral gaps while I'm
> gone!
Have fun and best of luck to you as well. I'm sure you'll do fine at
staying out of the neutral gaps yourself! ; )
> Cheers,
> Leonid.
Sean Pitman
www.DetectingDesign.com
Seanpit
> > I see now that this actually came from the "Horse and Donkey" thread, which
> > I wasn't following. I'm not going to read it, because it is too old and too
> > long, so I'll give up my notion that Sean might be confusing mutation rate
> > with substitution rate.
>
> Don't be so hasty! If you read upthread - specifically, the post that
> you originally responded to - you'll see that the good doctor appears
> to have done exactly that.
Actually, it seems to me that you have confused the concepts of
mutation rates with substitution rates. You're calculations depend
upon the notion that the rate of beneficial substitutions is very
similar to the rate of beneficial mutations. This simply isn't true.
In fact, it is interesting to note that the authors of the papers you
reference assume a much lower ratio for Ub/Ud than you do.
Sean Pitman
www.DetectingDesign.com
R. Baldwin
Would you mind pointing out precisely where you think Leonid has made
incorrect calculations? He has indicated, a few posts back, where he
believes you did.
>
Seanpit
> Would you mind pointing out precisely where you think Leonid has made
> incorrect calculations? He has indicated, a few posts back, where he
> believes you did.
Sean Pitman
www.DetectingDesign.com
Seanpit
Seanpit wrote:
> R. Baldwin wrote:
>
> > Would you mind pointing out precisely where you think Leonid has made
> > incorrect calculations? He has indicated, a few posts back, where he
> > believes you did.
http://groups.google.com/group/talk.origins/msg/76e6c60147647348?dmode=source
> Sean Pitman
> www.DetectingDesign.com
RobinGoodfellow
Seanpit wrote:
> RobinGoodfellow wrote:
[snip]
> > So, a rate of 1
> > substitution every 200 years, is the rate of 0.125 substitutions per
> > population per generation, assuming 25 year generations. Which means
> > that the rate of slightly beneficial mutations will be slightly lower
> > than 0.125 per individual per generation.
>
> Actually, the rate of slightly beneficial mutations will be quite a bit
> lower. The substitution rate of 0.125 per population per generation is
> based on a concentration of more than slightly beneficial mutations in
> the larger population into a smaller subpopulation over time with many
> of the more slightly beneficial mutations being lost before fixation.
> Out of all the millions of beneficial mutations that might have been
> realized in a given generation, only 0.125 make it to fixation. This
> means that the actual beneficial mutation rate is much much lower than
> the beneficial substitution rate.
Wait a minute! You're claiming here that out of millions of beneficial
mutations, only 0.125 make it to fixation, and that the vast majority
of beneficial mutations are being lost before fixation. But this means
that beneficial mutation rate is actually *much higher* than the
substitution rate: i.e. for every X beneficial mutations per individual
per generation, only 1 per 8 generations makes it to substitution.
Could you be any more confused, Sean?
[snip]
RobinGoodfellow
Sean never actually gets around to doing so. He *asserts* my estimate
of beneficial mutation rate is too high, then proceeds to make an
argument which, if correct, would demonstrate that my estimate of
beneficial mutation rate is actually way too low (see my brief reply to
him), then quotes from several unrelated small-scale theoretical
studies that already assume their beneficial and deleterious rates a
priori and have nothing to do with real genomic data, and finally skips
over my entire calculation of beneficial rates from Bustamante's data,
pausing briefly only to reassure me that I am wrong. All in all,
vintage Sean Pitman.
Unfortunately, I won't have the time to answer him in detail for at
least few weeks. Though, given that these discussions are quickly
proving an exercise in futility, I'm not sure if that is such a bad
thing.
Cheers,
Leonid.
R. Baldwin
Could you be more specific, Sean? What calculation did Leonid post that you
believe is incorrect, and why?
Steve Schaffner
> Seanpit wrote:
> > RobinGoodfellow wrote:
> > > Seanpit wrote:
> > > > RobinGoodfellow wrote:
> > >
> > > < snip >
> > >
> > > > You seem to give a lot of weight to the fact that N&C seem to think
> > > > that the deleterious mutation rate is too high, and not at all to the
> > > > fact that they seem to think that humans and chimps diverged from a
> > > > common ancestor - even though they use the latter as a premise to
> > > > arrive at the former conclusion (the Ka/Ks ratios). Not too big on
> > > > consistency, are we?
> > >
> > > If U really is greater than 3, all slowly producing animals, to include
> > > both humans and apes, are headed for extinction.
> >
> > That is your understanding, yes. Very few evolutionary biologists,
> > even those who accept the U>=3 rates, seem to share it. I wonder why?
>
> This doesn't mean that evolutionary biologists can explain how
> detrimental mutations that enter the gene pool at this rate (U > 3) can
> actually be cleared in a way that does not lead to extinction.
This has already been pointed out to you, but I doubt you understood
it. A deleterious mutation rate > 3 only threatens extinction if you
make certain assumptions about how selection works. If all selection
were soft, for example, then any rate of deleterious mutations could
be accomodated. What happens is that deleterious mutations accumulate
in the population, with more and more entering every generation. The
number of deleterious mutations in each individual varies, however,
and the ones with the highest load are least likely to
reproduce. Eventually, the number of deleterious alleles being removed
by the failure of the most unfit individuals equals the input from
new mutations, and the population is at equilibrium.
This scenario isn't realistic, because some selection is hard, but no
one knows enough about the details of selection to say what a
realistic maximum genetic load would be. Certainly you don't.
--
Steve Schaffner s...@broad.mit.edu
Immediate assurance is an excellent sign of probable lack of
insight into the topic. Josiah Royce
Seanpit
I'm quite surprised that you came back with this comment Leonid. Don't
you see? - The number of beneficial mutations in a population is a
function of the size of the population as well as the rate of
beneficial mutations in that population. For argument's sake, lets
assume that we have a Ud/Ub ratio of 1000 to 1 in a population of 10
billion. If the Ud rate were 3 per individual per generation, the Ub
rate would be 0.003 per individual per generation. Out of the 10
billion, how many Ub mutations would there be in one generation? The
answer is 30 million - right? Out of this 30 million, how many will
make it to fixation? Less than 1 of them - right?
So, you see, the rate of beneficial mutations can indeed be very much
lower than the rate of beneficial substitutions. In fact, the authors
of the papers you reference do use the 1000 to 1 ratio in their
estimates of Ud vs Ub (Ref).
Who's confused here? ; )
Sean Pitman
www.DetectingDesign.com
Reference:
Michael C. Whitlock, Fixation of New Alleles and the Extinction of
Small Populations: Drift Load, Beneficial Alleles, and Sexual
Selection, Evolution, Vol. 54, No. 6, pp. 1855-1861. -
http://evol.allenpress.com/evolonline/?request=get-document&issn=0014...
Seanpit
Steve Schaffner wrote:
> This has already been pointed out to you, but I doubt you understood
> it. A deleterious mutation rate > 3 only threatens extinction if you
> make certain assumptions about how selection works. If all selection
> were soft, for example, then any rate of deleterious mutations could
> be accomodated.
That's completely wrong. In fact, near-neutral deleterious mutations
are the most destructive type of mutations to a gene pool. Soft
selection only puts off the inevitable and contributes to the decline
of a population toward extinction. The only way to reverse this trend
is to send more individuals up the chain than are descending down the
chain of R-best toward R-worst. With high mutation rates in slowly
reproducing individuals, the flow of the current is steadily downward.
A few individuals do make it a short way up the current, but are
eventually overcome by the vast multitude that is drifting down-current
with each generation.
> What happens is that deleterious mutations accumulate
> in the population, with more and more entering every generation. The
> number of deleterious mutations in each individual varies, however,
> and the ones with the highest load are least likely to
> reproduce.
That's correct. The problem is that the number of individuals that are
entering lower and lower levels of relative fitness, heading toward
this threshold cutoff, is growing with each generation in slowing
reproducing populations that are sustaining high mutation rates.
> Eventually, the number of deleterious alleles being removed
> by the failure of the most unfit individuals equals the input from
> new mutations, and the population is at equilibrium.
This is simply not true for slowly reproducing populations that are
sustaining high mutation rates. You need to read the papers I've
listed that are trying to deal with this problem. You certainly don't
understand it.
> This scenario isn't realistic, because some selection is hard, but no
> one knows enough about the details of selection to say what a
> realistic maximum genetic load would be. Certainly you don't.
Well, you clearly don't even understand the basic issues involved and
aren't contributing to this discussion at all. Sorry, but even Leonid
is doing a far better job here than you are.
> Steve Schaffner s...@broad.mit.edu
> Immediate assurance is an excellent sign of probable lack of
> insight into the topic. Josiah Royce
Exactly . . . How are you so sure you're right on this one? Where are
your references to back up your most unusual notions?
Sean Pitman
www.DetectingDesign.com
John Harshman
> Steve Schaffner wrote:
>
>
>>This has already been pointed out to you, but I doubt you understood
>>it. A deleterious mutation rate > 3 only threatens extinction if you
>>make certain assumptions about how selection works. If all selection
>>were soft, for example, then any rate of deleterious mutations could
>>be accomodated.
>
>
> That's completely wrong. In fact, near-neutral deleterious mutations
> are the most destructive type of mutations to a gene pool. Soft
> selection only puts off the inevitable and contributes to the decline
> of a population toward extinction. The only way to reverse this trend
> is to send more individuals up the chain than are descending down the
> chain of R-best toward R-worst. With high mutation rates in slowly
> reproducing individuals, the flow of the current is steadily downward.
> A few individuals do make it a short way up the current, but are
> eventually overcome by the vast multitude that is drifting down-current
> with each generation.
There is excellent evidence againt this position that you are not
thinking of, i.e. the evidence that lineages with long generations are
millions of years old, and yet haven't suffered any such extinction. Of
course in order to consider this evidence you have to directly confront
both deep time and common descent, which you refuse to do.
All this quite aside from the population genetics of it.
Seanpit
John Harshman wrote:
> There is excellent evidence againt this position that you are not
> thinking of, i.e. the evidence that lineages with long generations are
> millions of years old, and yet haven't suffered any such extinction. Of
> course in order to consider this evidence you have to directly confront
> both deep time and common descent, which you refuse to do.
>
> All this quite aside from the population genetics of it.
You don't understand what's going on here John. Consider that if
humans and apes really are descended from some common ancestor who
lived many millions of years ago that genetic comparisons suggest a
very high mutation rate. This is inconsistent because such a high
mutation rate, to include a very high detrimental mutation rate, if a
reality, would lead to extinction of both humans and apes - not toward
their independent evolution. Something is obviously wrong here one way
or another. If you are right, then the conclusions of scientists who
do genetic comparisons between humans and apes is wrong. If they are
right, then your notion of a common ancestor millions of years ago is
wrong. You both can't be right.
Sean Pitman
www.DetectingDesign.com
Seanpit
R. Baldwin wrote:
> Could you be more specific, Sean? What calculation did Leonid post that you
> believe is incorrect, and why?
I believe Leonid's calculation of the beneficial mutation rate in
humans is wrong. He suggests a rate of about 0.125 beneficial
mutations per person per generation for an overall ratio of detrimental
vs beneficial of about 30 to 1 instead of my suggestion of at least
1000 to 1.
For a more detailed explanation of why I think Leonid's estimates here
are based on mistaken assumptions, see the link to my detailed response
to Leonid.
http://groups.google.com/group/talk.origins/msg/76e6c60147647348?dmod...
Sean Pitman
www.DetectingDesign.com
John Harshman
> John Harshman wrote:
>
>
>>There is excellent evidence againt this position that you are not
>>thinking of, i.e. the evidence that lineages with long generations are
>>millions of years old, and yet haven't suffered any such extinction. Of
>>course in order to consider this evidence you have to directly confront
>>both deep time and common descent, which you refuse to do.
>>
>>All this quite aside from the population genetics of it.
>
>
> You don't understand what's going on here John.
I'll concede that one of us doesn't.
> Consider that if
> humans and apes really are descended from some common ancestor who
> lived many millions of years ago that genetic comparisons suggest a
> very high mutation rate. This is inconsistent because such a high
> mutation rate, to include a very high detrimental mutation rate, if a
> reality, would lead to extinction of both humans and apes - not toward
> their independent evolution. Something is obviously wrong here one way
> or another. If you are right, then the conclusions of scientists who
> do genetic comparisons between humans and apes is wrong. If they are
> right, then your notion of a common ancestor millions of years ago is
> wrong. You both can't be right.
There is a third possibility, that there is indeed an evolutionary
mechanism you don't know about that removes deleterious mutations at
higher rates than you imagine. There are other possibilities too. Why
are creationists so fond of false dichotomies?
But OK, let's go with it and pretend there is such a dichotomy. Let's
put the evidence favoring deep time and common descent against the
evidence favoring some kind of catastrophic spiral toward extinction.
Rejecting the first would require us to reject much of what we know of
physics and geology as well as biology. Rejecting the second would
require us to suppose that a few estimates of hard-to-determine
parameters, with lots of unknowns, might be a bit off. I know which one
you want to pick -- but why?
Steve Schaffner
> Steve Schaffner wrote:
>
> > This has already been pointed out to you, but I doubt you understood
> > it. A deleterious mutation rate > 3 only threatens extinction if you
> > make certain assumptions about how selection works. If all selection
> > were soft, for example, then any rate of deleterious mutations could
> > be accomodated.
>
> That's completely wrong. In fact, near-neutral deleterious mutations
> are the most destructive type of mutations to a gene pool. Soft
> selection only puts off the inevitable and contributes to the decline
> of a population toward extinction.
That depends entirely on what effect these deleterious mutations have
on the absolute, rather than the relative, fitness. "Deleterious"
here means that they place the individual at a disadvantage relative
to others in the population. Their effect on the survival of the
population as a whole can be large, small, zero or even positive.
Simply saying that deleterious mutations are accumulating does not
tell you what is happening to the population.
> The only way to reverse this trend
> is to send more individuals up the chain than are descending down the
> chain of R-best toward R-worst. With high mutation rates in slowly
> reproducing individuals, the flow of the current is steadily downward.
> A few individuals do make it a short way up the current, but are
> eventually overcome by the vast multitude that is drifting down-current
> with each generation.
No, this is wrong. The number does not increase without limit. When
enough deleterious alleles have accumulated, the fluctations in the
number that get passed to the next generation become large compared to
the number entering from new mutation, and the differential loss of
the least fit balances new mutations. For a simple model with
multiplicative fitness, a population size of 10,000 and three new
deleterious mutations per generation, eadh with an effect of -1% on
fitness, the equilibrium genetic load is about 150 deleterious
mutations carried by each individual. Is that number sustainable by a
population? That's an empirical question. Mathematical modelling
won't tell you without empirical input (and simply shouting "We're all
gonna die" doesn't address the question).
> > What happens is that deleterious mutations accumulate
> > in the population, with more and more entering every generation. The
> > number of deleterious mutations in each individual varies, however,
> > and the ones with the highest load are least likely to
> > reproduce.
>
> That's correct. The problem is that the number of individuals that are
> entering lower and lower levels of relative fitness, heading toward
> this threshold cutoff, is growing with each generation in slowing
> reproducing populations that are sustaining high mutation rates.
>
> > Eventually, the number of deleterious alleles being removed
> > by the failure of the most unfit individuals equals the input from
> > new mutations, and the population is at equilibrium.
>
> This is simply not true for slowly reproducing populations that are
> sustaining high mutation rates. You need to read the papers I've
> listed that are trying to deal with this problem. You certainly don't
> understand it.
I've read the papers, or at least some of them. What I'm saying is
simple math.
> > This scenario isn't realistic, because some selection is hard, but no
> > one knows enough about the details of selection to say what a
> > realistic maximum genetic load would be. Certainly you don't.
>
> Well, you clearly don't even understand the basic issues involved and
> aren't contributing to this discussion at all. Sorry, but even Leonid
> is doing a far better job here than you are.
I'm sorry, but you don't even understand what the issues are. And I
dare say I've published more papers about selection than you have.
> Exactly . . . How are you so sure you're right on this one? Where are
> your references to back up your most unusual notions?
Start by reading _Fifty years of genetic laod_, by Bruce Wallace.
--
Steve Schaffner
> Seanpit wrote:
>
> > Consider that if
> > humans and apes really are descended from some common ancestor who
> > lived many millions of years ago that genetic comparisons suggest a
> > very high mutation rate. This is inconsistent because such a high
> > mutation rate, to include a very high detrimental mutation rate, if a
> > reality, would lead to extinction of both humans and apes - not toward
> > their independent evolution. Something is obviously wrong here one way
> > or another. If you are right, then the conclusions of scientists who
> > do genetic comparisons between humans and apes is wrong. If they are
> > right, then your notion of a common ancestor millions of years ago is
> > wrong. You both can't be right.
>
> There is a third possibility, that there is indeed an evolutionary
> mechanism you don't know about that removes deleterious mutations at
> higher rates than you imagine. There are other possibilities too. Why
> are creationists so fond of false dichotomies?
There are probably quite a few things going on. One is that we can
probably tolerate quite a high genetic laad. It seems clear that many
diseases have multigenic risk factors, and that many risk alleles are
common, so we all have quite a few of them, and yet most of manage to
function pretty well. Another factor is that some fraction of
deleterious alleles prevent the production of viable gametes or
zygotes, and so waste minimal reproductive potential. Another is
negative interactions between deleterious alleles, leading to greater
than multiplicative disadvantage to those with more of them.
--
Seanpit
Steve Schaffner wrote:
< snip >
> > The only way to reverse this trend
> > is to send more individuals up the chain than are descending down the
> > chain of R-best toward R-worst. With high mutation rates in slowly
> > reproducing individuals, the flow of the current is steadily downward.
> > A few individuals do make it a short way up the current, but are
> > eventually overcome by the vast multitude that is drifting down-current
> > with each generation.
>
> No, this is wrong. The number does not increase without limit. When
> enough deleterious alleles have accumulated, the fluctations in the
> number that get passed to the next generation become large compared to
> the number entering from new mutation, and the differential loss of
> the least fit balances new mutations. For a simple model with
> multiplicative fitness, a population size of 10,000 and three new
> deleterious mutations per generation, eadh with an effect of -1% on
> fitness, the equilibrium genetic load is about 150 deleterious
> mutations carried by each individual. Is that number sustainable by a
> population? That's an empirical question. Mathematical modelling
> won't tell you without empirical input (and simply shouting "We're all
> gonna die" doesn't address the question).
If you have a steady-state population of 10,000 individuals and, in
each generation, each one of the offspring sustain an average of 3
additional detrimental mutations (-1% fitness effect) you can only
reach an equilibrium *if* the reproductive rate high enough. The whole
problem is at least partially dependent upon the reproductive rate. If
the reproductive rate is too low, the population will indeed head for
extinction. If you think not, then please do explain why not.
For example, if the reproductive rate is very low, in each generation
the various levels of fitness from R-best to R-threshold will end up
giving more individuals to the lower levels of fitness than to the next
higher level. The numbers of R-best, as well as every other lower
category, will steadily decline until the original level of R-best will
no longer have any members. The same thing will happen to the new
lower-level R-best . . . and so on, until R-best-current is so low that
it is just above the level of extinction itself. At this point, the
population as a whole will indeed quickly go into extinction.
As far as I understand the problem, the only way to overcome this
steady decline is to increase the reproductive rate dramatically or
increase the rate of beneficial mutations dramatically. Genetic
recombination does slow things down a bit, but not enough to compensate
for reproductive rates as low as humans or apes or elephants - given a
detrimental mutation rate > 3 per individual per generation.
If you have some sort of insight into why reproductive rates are not
really an issue here, I'd love to hear your explanation.
> > > Eventually, the number of deleterious alleles being removed
> > > by the failure of the most unfit individuals equals the input from
> > > new mutations, and the population is at equilibrium.
> >
> > This is simply not true for slowly reproducing populations that are
> > sustaining high mutation rates. You need to read the papers I've
> > listed that are trying to deal with this problem. You certainly don't
> > understand it.
>
> I've read the papers, or at least some of them. What I'm saying is
> simple math.
It seems to me that your math assumes a very large reproductive rate.
For example, lets say that every individual with any detrimental
mutation fails to reproduce. What happens to the population sustaining
high levels of detrimental mutations? It goes extinct unless it can
dramatically increase the reproductive rate or the beneficial mutation
rate. Sure, if you make some threshold number like 10,000 detrimental
mutations before an individual will cross over and not be able to
reproduce, you put off things a bit and allow for recombination to slow
things down even more. However, unless the population is able to pick
up its reproductive rate or beneficial mutation rate, how can it reach
equilibrium? How do you prevent the R-best category from continued
decline over time?
> I'm sorry, but you don't even understand what the issues are. And I
> dare say I've published more papers about selection than you have.
Why not explain, then, how a population can actually reach equilibrium,
in light of very high detrimental mutation rates, regardless of its
reproductive rate. How is this done?
> Start by reading _Fifty years of genetic laod_, by Bruce Wallace.
You know, I get a lot of people telling me to read this or that book or
journal article and this will solve my problems. Usually, though, when
I do actually get a chance to read some of these books or articles,
they don't solve the problem I'm considering at all. So, I'm much more
interested in hearing your own explanation of how you think the problem
is solved. I don't think it is nearly as simple as you make it out to
be. Could be wrong though. Use whatever books or other resources you
want, but I'd like to hear your own explanation.
> Steve Schaffner s...@broad.mit.edu
Sean Pitman
www.DetectingDesign.com
Seanpit
Steve Schaffner wrote:
> There are probably quite a few things going on. One is that we can
> probably tolerate quite a high genetic laad. It seems clear that many
> diseases have multigenic risk factors, and that many risk alleles are
> common, so we all have quite a few of them, and yet most of manage to
> function pretty well. Another factor is that some fraction of
> deleterious alleles prevent the production of viable gametes or
> zygotes, and so waste minimal reproductive potential. Another is
> negative interactions between deleterious alleles, leading to greater
> than multiplicative disadvantage to those with more of them.
Doesn't this form of "reinforcing epistasis" actually hasten the
decline of a slowly reproducing population undergoing high detrimental
mutation rates? Look at Figure 2 in Rice's article and let me know how
you believe this form of epistasis actually solves the problem.
http://www.lifesci.ucsb.edu/eemb/faculty/rice/publications/pdf/40.pdf
Steve Schaffner
> If you have a steady-state population of 10,000 individuals and, in
> each generation, each one of the offspring sustain an average of 3
> additional detrimental mutations (-1% fitness effect) you can only
> reach an equilibrium *if* the reproductive rate high enough. The whole
> problem is at least partially dependent upon the reproductive rate. If
> the reproductive rate is too low, the population will indeed head for
> extinction. If you think not, then please do explain why not.
You need some reproductive excess, sure, but you don't need much. In
the simple model I quoted in a previous post, the reproductive rate
was four per breeding pair.
> For example, if the reproductive rate is very low, in each generation
> the various levels of fitness from R-best to R-threshold will end up
> giving more individuals to the lower levels of fitness than to the next
> higher level. The numbers of R-best, as well as every other lower
> category, will steadily decline until the original level of R-best will
> no longer have any members. The same thing will happen to the new
> lower-level R-best . . . and so on, until R-best-current is so low that
> it is just above the level of extinction itself. At this point, the
> population as a whole will indeed quickly go into extinction.
>
> As far as I understand the problem, the only way to overcome this
> steady decline is to increase the reproductive rate dramatically or
> increase the rate of beneficial mutations dramatically. Genetic
> recombination does slow things down a bit, but not enough to compensate
> for reproductive rates as low as humans or apes or elephants - given a
> detrimental mutation rate > 3 per individual per generation.
>
> If you have some sort of insight into why reproductive rates are not
> really an issue here, I'd love to hear your explanation.
I did already give the explanation, but it obviously didn't come
across, so I'll try again. The inexorable increase in deleterious
mutations you describe is true only in the absence of recombination.
Introducing sexual reproduction and recombination doesn't just slow
things down a bit; it fundamentally changes the dynamics.
Consider a hypothetical population of 10,000 diploids, each having
exactly 97 mildly deleterious alleles, and a mutation rate of three
new deleterious alleles per generation. They mate and produce 20,000
offspring, with an average of 100 deleterious per individual. They
don't all have 100, however. Because the offspring have a random
assortment of alleles from the two parents, it's possible to have more
_or fewer_ deleterious alleles than either of their parents. That last
point is the key to understanding how diploids avoid mutational
meltdown.
In a simple model (random mating, arbitrarily large amounts of
recombination), each offspring inherits a Poisson random number of
deleterious alleles, with mean given by the average of the two
parent's numbers. In our simple example, that means the offspring
have 100 +/- 10 deleterious alleles apiece. So quite a large fraction
of the 2nd generation population has fewer deleterious alleles than
the 1st generation. Assuming constant population size, only half of
this generation will go on to reproduce, and that half will
disproportionately come from the least loaded end of the population,
since they have higher relative fitness.
How heavy that weighting is depends on the selection coefficient. For
small coefficients, there is only a small bias, and the number of
deleterious mutations will increase. But as it increases, the spread
in the mutational load distribution in each generation also
increases. If we had started with an initial load of 397 instead of
97, then the spread in the 2nd generation would be +/- 20 instead of
+/- 10. Since the input from new mutations is always fixed (at 3, in
this case), eventually the load will always increase to the point that
the spread (combined with biased reproductive success) always brings
the population into equilibrium. That's why Rice's paper talks about
the required load to reach equilibrium.
Whether the required load is more than the population can handle
without going extinct is, as I said, an empirical question. Simply
quoting the deleterious mutation rate (even if it were well measured)
cannot tell you that.
>
> It seems to me that your math assumes a very large reproductive rate.
> For example, lets say that every individual with any detrimental
> mutation fails to reproduce. What happens to the population sustaining
> high levels of detrimental mutations?
That's hard selection, not soft selection. Hard selection imposes
constraints on reproductive rate that soft selection (in itself) does not.
> > Start by reading _Fifty years of genetic laod_, by Bruce Wallace.
>
> You know, I get a lot of people telling me to read this or that book or
> journal article and this will solve my problems. Usually, though, when
> I do actually get a chance to read some of these books or articles,
> they don't solve the problem I'm considering at all. So, I'm much more
> interested in hearing your own explanation of how you think the problem
> is solved.
Uh, Sean, you asked me for references. Complaining because I gave you
a reference in response is a little silly.
I'm going to see if I can come up with a lower bound on the number of
deleterious alleles currently residing in the average human. How many
do you think we've got? 1 per person? 10, 100, 1000?
Steve Schaffner
> Steve Schaffner wrote:
>
> > There are probably quite a few things going on. One is that we can
> > probably tolerate quite a high genetic laad. It seems clear that many
> > diseases have multigenic risk factors, and that many risk alleles are
> > common, so we all have quite a few of them, and yet most of manage to
> > function pretty well. Another factor is that some fraction of
> > deleterious alleles prevent the production of viable gametes or
> > zygotes, and so waste minimal reproductive potential. Another is
> > negative interactions between deleterious alleles, leading to greater
> > than multiplicative disadvantage to those with more of them.
>
> Doesn't this form of "reinforcing epistasis" actually hasten the
> decline of a slowly reproducing population undergoing high detrimental
> mutation rates? Look at Figure 2 in Rice's article and let me know how
> you believe this form of epistasis actually solves the problem.
No. Read the paper. Reinforcing epistasis increases the efficiency
of selection and reduces the required genetic load. What it does is
increase the bias against those in the more loaded part of the load
spectrum (see my last post).
Seanpit
> > So, you admit that both you and Zach realize my requirement for
> > selection based on improvement in beneficial meaning/function. Yet,
> > Zach decided to go ahead and write his programs without this little
> > detail. It isn't my job to tell Zach how to program a program to
> > recognize improvement in meaning/function. That's his job. I just
> > know that his programs don't do it.
>
>
> Concerning "my job":
>
> Just for the record, I am not a professional software engineer, which should
> have been apparent. It's just that computers are such a valuable tool that
> some facility is a great aid in analyzing all sorts of data. Rather, I am
> more than happy to use my meager skills to confirm your predictions! You
> just have to tell us what you mean specifically by "improvement in
> meaning/function" with regards to the English language.
One way you could do it is to set up a command-type program where
various words and phrases in the English language translate into
certain actions. Higher-level actions would require longer command
sequences to make the resulting function "work". Another way would be
to use computer code to code for a functional system like a chess
program that you could see and interact with on the computer screen.
Now, starting with a short portion of this larger program, evolve the
larger program using random mutation to the underlying code that
provides options for selection where the selector cannot see the
changes in the underlying code, only the resulting changes on the
screen. See how far you can go toward evolution of a
Shakespearean-sized code sequence this way.
You see Zach, meaning/function is context dependent. The phrase, "to
be or not" may or may not make sense depending upon the
situation/environment in which it is read or spoken. You have to
create this context into your programs somehow. At this point, it just
isn't there.
> Zachriel, angel that rules over memory, presides over the planet Jupiter.
> Member AMF, Angelic Motive Force: Pushing planets on celestial spheres - one
> epoch at a time.
> http://zachriel.blogspot.com/
Sean Pitman
www.DetectingDesign.com
Seanpit
Steve Schaffner wrote:
Give me some time on this one. I'm still thinking about it. I think
you might actually be right, but I've gotta consider it a bit more.
> Steve Schaffner s...@broad.mit.edu
> Immediate assurance is an excellent sign of probable lack of
> insight into the topic. Josiah Royce
Sean Pitman
www.DetectingDesign.com
John Harshman
This is something I've never previously seen a creationist do: admit the
possibility that he might be wrong, even on the most trivial point. I am
impressed. No implied smileys.
John Harshman
Sean's refusal to consider this argument, even though it relates
directly to his neutral gap theory, is, I think, telling:
Seanpit
Steve Schaffner wrote:
> > If you have some sort of insight into why reproductive rates are not
> > really an issue here, I'd love to hear your explanation.
>
> I did already give the explanation, but it obviously didn't come
> across, so I'll try again. The inexorable increase in deleterious
> mutations you describe is true only in the absence of recombination.
> Introducing sexual reproduction and recombination doesn't just slow
> things down a bit; it fundamentally changes the dynamics.
Let's see here. You started with a steady state population of 10,000
diploids each having 97 mildly detrimental mutations. They mate to
produce 20,000 offspring. Given an average Ud of 3, the average number
of deleterious mutations, per individual, is 100. Of course, given the
Poisson distribution, some will have more and some will have less. You
argue that, "quite a large fraction of the 2nd generation population
has fewer deleterious alleles than the 1st generation." Therefore, it
is clear to see that the sexually reproducing population can go uphill,
so to speak. Sound good so far?
Without any additional mutations of any kind affecting this population,
genetic recombination will divide the population about half and half,
in a normalized binomial distribution, where one side has equal or more
and the other has equal or less than 97 mutations. Out of the 20,000
offspring, about 47% will have a beneficial mutational load. That's
9400 offspring that have moved at least 1 level uphill.
Now, what happens when you bring in a detrimental mutation rate of Ud =
3? How many of these 9,400 stay on the plus side of the equation? The
majority of these 9,400 have only a few beneficial mutations. The
detrimental mutations will also be distributed in a Poisson
distribution pattern where those with the least beneficial mutations
will be hit with the least number of detrimental mutations and those
with the most beneficial mutations will be hit with the most
detrimental mutations.
For example, of those with a neutral or positive mutational balance,
About 1140 of them will have 0 beneficial mutations. Of these, about 56
will not be hit by any detrimental mutations and will maintain a
neutral mutational balance.
About 1,100 of them will have 1 beneficial mutation. Of these, 54 will
not be hit by any detrimental mutations and stay with 1 beneficial
mutation. 164 will be hit by 1 detrimental mutation and maintain a
neutral mutational balance and the rest will have a detrimental
mutational balance.
About 1040 of them will have 2 beneficial mutations. Of these, about 52
will not be hit by any detrimental mutations and stay with 2 beneficial
mutations. 162 will be hit by 1 detrimental mutation and go to the
level of 1 beneficial mutation. 233 will be hit by 2 detrimental
mutations and maintain a neutral mutational balance, and the rest will
have a detrimental mutational balance.
Etc . . .
In short, out of the 20,000 offspring, less than 3,000 will have either
a neutral balance or better. The rest will have a detrimental
mutational load relative to the parent generation.
It seems then that a reproductive rate of just 4 is not enough to
maintain equilibrium with a detrimental mutation rate of Ud = 3. In
other words, the sand will roll downhill much faster than it rolls
uphill. Overall, each level of fitness will loose much more to the
lower levels than it contributes to the higher levels of fitness. How
is this compensated for without significantly increasing the
reproductive rate? And, for rates of Ud=5, it gets a whole lot worse.
Beyond this little problem, consider that the Y-chromosome in males
does not undergo significant recombination events. It basically
undergoes asexual replication. So, how is the male population helped
by recombination and/or epistasis at all? Doesn't it follow that males
are headed for extinction in slowly reproducing populations even faster
than females?
Sean Pitman
www.DetectingDesign.com
Seanpit
Steve Schaffner wrote:
> > Doesn't this form of "reinforcing epistasis" actually hasten the
> > decline of a slowly reproducing population undergoing high detrimental
> > mutation rates? Look at Figure 2 in Rice's article and let me know how
> > you believe this form of epistasis actually solves the problem.
>
> No. Read the paper. Reinforcing epistasis increases the efficiency
> of selection and reduces the required genetic load. What it does is
> increase the bias against those in the more loaded part of the load
> spectrum (see my last post).
Reinforcing epistasis removes those with higher detrimenal loads. How
does it help in keeping the overall ballance in equilibrium given a
slowly reproducing population?
Let's think about this a bit. You started with a steady state
Etc . . .
How does reinforcing epistasis solve the problem? Removing those with
higher numbers of detrimental mutations doesn't seem to help replace
the loss of those with low numbers of detrimental mutations.
Beyond this little problem, consider that the Y-chromosome in males
does not undergo significant recombination events. It basically
undergoes asexual replication. So, how is the male population helped
by recombination and/or epistasis at all? Doesn't it follow that males
are headed for extinction in slowly reproducing populations even faster
than females?
> Steve Schaffner s...@broad.mit.edu
> Immediate assurance is an excellent sign of probable lack of
> insight into the topic. Josiah Royce
Sean Pitman
www.DetectingDesign.com
Seanpit
John Harshman wrote:
> > Give me some time on this one. I'm still thinking about it. I think
> > you might actually be right, but I've gotta consider it a bit more.
>
> This is something I've never previously seen a creationist do: admit the
> possibility that he might be wrong, even on the most trivial point. I am
> impressed. No implied smileys.
I've admitted that I was wrong several times in this forum. The fact
is though, many of your evolutionist buddies in this forum have just as
much trouble admitting even trivial errors as creationists do. I think
its a human thing, especially a male thing, to have difficulty
admitting error or even the possibility of error.
Sean Pitman
www.DetectingDesign.com
John Harshman
We don't have a good sample, but it's been my impression that
creationists on TO have been less willing to admit being wrong than
non-creationists have, granted that there is still a strong overall bias
against anyone admitting they were wrong that might swamp any other effect.
Steve Schaffner
Your numbers are a little off, but they're close enough. You
shouldn't have stopped there, however. Here are the number of
offspring, classified by the number of inherited deleterious alleles,
and the number of them that have 97 or fewer after new mutations are
added:
inherited N(offspring) N(del <= 97)
97 1144 57
96 1133 226
95 1098 465
94 1043 675
93 971 792
92 885 811
91 791 764
90 692 684
89 593 591
If you keep going, you will find a total of 7285 offspring with 97
or fewer deleterious alleles, compared to 12715 with > 97 deleterious
alleles. (If you'd just used the Poisson approximation (100 +/- 10
deleterious alleles), you'd have gotten 8148 offspring with <= 97,
and a normal approximation would give 7264.)
That's the raw number of mutations. Now apply selection. How much
more likely are the 7285 offspring (or the 6265 offspring with < 97
deleterious alleles) to survive and reproduce than the ones with more
deleterious mutations (some of them with many more)? That depends on
the selection coefficient. If the coefficient is high, so that, say,
the average lightly loaded offspring is twice as likely to reproduce
as the average heavily loaded offspring, then the lightly loaded
offspring will actually contribute more to the next generation
than the heavily loaded ones, and the mean number of deleterious
alleles will decrease.
Note that this is well within the reproductive capacity that I
specified. Take the extreme case: all 7285 lightly loaded offspring
form part of the next breeding generation, along with 2715 heavily
loaded ones. Has the mean number of deleterious alleles in the
population increased or decreased?
> It seems then that a reproductive rate of just 4 is not enough to
> maintain equilibrium with a detrimental mutation rate of Ud = 3. In
> other words, the sand will roll downhill much faster than it rolls
> uphill. Overall, each level of fitness will loose much more to the
> lower levels than it contributes to the higher levels of fitness. How
> is this compensated for without significantly increasing the
> reproductive rate?
Since you haven't done the calculation correctly, and haven't
considered the effect of selection (rather an important omission when
you're studying selection itself), your conclusion is premature. More
accurately, it's wrong.
> And, for rates of Ud=5, it gets a whole lot worse.
> How does reinforcing epistasis solve the problem? Removing those with
> higher numbers of detrimental mutations doesn't seem to help replace
> the loss of those with low numbers of detrimental mutations.
You have a basic misunderstanding here. At equilibrium, half the
population will have more than the mean number of deleterious
alleles. If you have 5000 with < 97 and 500 with > 97, that doesn't
mean you're losing half of the lightly loaded ones each generation,
it means you're in equilibrium.
> Beyond this little problem, consider that the Y-chromosome in males
> does not undergo significant recombination events. It basically
> undergoes asexual replication. So, how is the male population helped
> by recombination and/or epistasis at all? Doesn't it follow that males
> are headed for extinction in slowly reproducing populations even faster
> than females?
This was exactly why it was suggested for a while that the Y
chromosome was headed for extinction. Note a couple of points.
First, there are not many genes (twenty-some, I believe) in the
non-recombining portion of the Y, so the number of deleterious
mutations there is likely to be quite small (even accounting for the
higher mutation rate in males). Second, the large palindromic repeats
on the Y enable gene conversion within the chromosome to play the
same role as recombination between autosomes, permitting the genes in
those regions to escape mutational decay.
--
Steve Schaffner
Program in Medical and Population Genetics
The Broad Institute of MIT and Harvard
Seanpit
Steve Schaffner wrote:
Hmmmm . . . This does make sense.
< snip >
> > Beyond this little problem, consider that the Y-chromosome in males
> > does not undergo significant recombination events. It basically
> > undergoes asexual replication. So, how is the male population helped
> > by recombination and/or epistasis at all? Doesn't it follow that males
> > are headed for extinction in slowly reproducing populations even faster
> > than females?
>
> This was exactly why it was suggested for a while that the Y
> chromosome was headed for extinction. Note a couple of points.
> First, there are not many genes (twenty-some, I believe) in the
> non-recombining portion of the Y, so the number of deleterious
> mutations there is likely to be quite small (even accounting for the
> higher mutation rate in males). Second, the large palindromic repeats
> on the Y enable gene conversion within the chromosome to play the
> same role as recombination between autosomes, permitting the genes in
> those regions to escape mutational decay.
Can you elaborate on this a bit more? Or, are there any fairly easily
accessible papers that describe this process? Thanks . . .
> Steve Schaffner
> Program in Medical and Population Genetics
> The Broad Institute of MIT and Harvard
Sean Pitman
www.DetectingDesign.com
Steve Schaffner
[On palindromic repeats on the Y chromosome:]
> Can you elaborate on this a bit more? Or, are there any fairly easily
> accessible papers that describe this process? Thanks . . .
Here's a brief description (including an appropriately skeptical note
-- this is still pretty speculative stuff), and has references to the
original papers:
http://www.wellcome.ac.uk/en/genome/thegenome/hg01f015.html
This is more of a press release, but provides more info:
http://www.brightsurf.com/news/june_03/HHMI_news_061903.html
--
Seanpit
Steve Schaffner wrote:
> "Seanpit" <seanpi...@naturalselection.0catch.com> writes:
>
> [On palindromic repeats on the Y chromosome:]
>
> > Can you elaborate on this a bit more? Or, are there any fairly easily
> > accessible papers that describe this process? Thanks . . .
>
> Here's a brief description (including an appropriately skeptical note
> -- this is still pretty speculative stuff), and has references to the
> original papers:
>
> http://www.wellcome.ac.uk/en/genome/thegenome/hg01f015.html
>
> This is more of a press release, but provides more info:
>
> http://www.brightsurf.com/news/june_03/HHMI_news_061903.html
Interesting. I'm still a bit skeptical myself about how beneficial
internal Y-chromosome shuffling would actually be. I'll have to think
about it a bit more.
I've also been thinking a bit more about the numbers in your last
example. In the first generation you start off with 10,000 individuals
with 97 detrimental mutations. In the second generation you end up
with ~7,300 individuals with < = 97 detrimental mutations and ~2,700
with more than 97 detrimental mutations. Every level contributed far
more individuals to the lower than to the higher levels. I'm just
wondering what the 3rd and 4th and 5th . . . etc. generations would
look like? Is there a simulation available where these parameters can
be plugged in and tested? Sure, some sand does indeed roll uphill, but
the majority does indeed seem to roll downhill. How is equilibrium
reached like this?
You explain:
> Now apply selection. How much
> more likely are the 7285 offspring (or the 6265 offspring with < 97
> deleterious alleles) to survive and reproduce than the ones with more
> deleterious mutations (some of them with many more)? That depends on
> the selection coefficient. If the coefficient is high, so that, say,
> the average lightly loaded offspring is twice as likely to reproduce
> as the average heavily loaded offspring, then the lightly loaded
> offspring will actually contribute more to the next generation
> than the heavily loaded ones, and the mean number of deleterious
> alleles will decrease.
For argument's sake, lets say that no offspring with more than 97
mutations are able to reproduce at all. Now, instead of 5,000 couples
mating to populate the next generation, we only have about 3,600.
True, these are more fit, but there are fewer of them. If they still
have the same maximum reproductive rate of 4 per couple, then only
about 14,500 offspring will be produced instead of the 20,000 produced
by the original parent population. Of these, won't most go downhill
instead of uphill? Won't most of these cross the boundary and end up
below the 97 detrimental mutation threshold?
> Note that this is well within the reproductive capacity that I
> specified. Take the extreme case: all 7285 lightly loaded offspring
> form part of the next breeding generation, along with 2715 heavily
> loaded ones. Has the mean number of deleterious alleles in the
> population increased or decreased?
The mean number has indeed decreased, but so has the absolute number of
offspring at or above the 97 starting point. What happens in the
subsequent generations?
> Steve Schaffner s...@broad.mit.edu
> Immediate assurance is an excellent sign of probable lack of
> insight into the topic. Josiah Royce
Sean Pitman
www.DetectingDesign.com
Steve Schaffner
Well, you can try to write down equations, which is what that Rice
paper did. If you look at his equation for the necessary load in
diploids, you'll find a factor for the new mutation rate, one for the
spread in the number of deleterious mutations carried (which is just
the statistical scatter from sampling), and one for the dependence of
fitness on the number of mutations.
Or you can, as you suggest, use a simulation. I've got one I wrote a
year or two ago, if you have the ability to compile C code.
> You explain:
>
> > Now apply selection. How much
> > more likely are the 7285 offspring (or the 6265 offspring with < 97
> > deleterious alleles) to survive and reproduce than the ones with more
> > deleterious mutations (some of them with many more)? That depends on
> > the selection coefficient. If the coefficient is high, so that, say,
> > the average lightly loaded offspring is twice as likely to reproduce
> > as the average heavily loaded offspring, then the lightly loaded
> > offspring will actually contribute more to the next generation
> > than the heavily loaded ones, and the mean number of deleterious
> > alleles will decrease.
>
> For argument's sake, lets say that no offspring with more than 97
> mutations are able to reproduce at all. Now, instead of 5,000 couples
> mating to populate the next generation, we only have about 3,600.
> True, these are more fit, but there are fewer of them. If they still
> have the same maximum reproductive rate of 4 per couple, then only
> about 14,500 offspring will be produced instead of the 20,000 produced
> by the original parent population. Of these, won't most go downhill
> instead of uphill? Won't most of these cross the boundary and end up
> below the 97 detrimental mutation threshold?
If there is a hard threshold at 97, then considering a population that
starts with exactly 97 deleterious mutations each is not a good model
for thinking about the situation. Think about a hard limit at, say,
110 instead. You will have more than enough offspring surviving to
produce the next generation, and the really highly loaded tail will be
completely removed. The effect is to increase the overall fitness of
the population at equilibrium, not decrease it.
What you describe is known as truncation selection, and is one extreme
solution proposed to explain how humans could support a high
deleterious mutation rate. (See the discussion section of the Nachman
and Crowell paper, where they mention it.) An ideal poplulation with
3.0 deleterious mutations per gen (2x reproductive capacity, Npop =
10,000, selection coeff = 0.01) comes to equilibrium at around 300
deleterious mutation per individual. (I think I reported that number
incorrectly previously.) With truncation selection operating, with a
threshold of 97 mutations, the population reaches equilibrium with 87
mutations per individual.
> > Note that this is well within the reproductive capacity that I
> > specified. Take the extreme case: all 7285 lightly loaded offspring
> > form part of the next breeding generation, along with 2715 heavily
> > loaded ones. Has the mean number of deleterious alleles in the
> > population increased or decreased?
>
> The mean number has indeed decreased, but so has the absolute number of
> offspring at or above the 97 starting point. What happens in the
> subsequent generations?
Mating in each generation tightens the distribution (combining two
randomly selected individuals from the population reduces the variance
on their average by 1/sqrt(2)), as does selection (which also biases
the distribution towards the low end). Production of new offspring
broadens the distribution, as random numbers of new mutations and
random sampling of existing mutations introduce scatter. Since we
started with zero variance in our thought experiment, initially the
variance will increase, and then stabilize. Selection keeps the high
tail from expanding indefinitely, and new mutations keep the low tail
from expanding.
--
Seanpit
Steve Schaffner wrote:
> > For argument's sake, lets say that no offspring with more than 97
> > mutations are able to reproduce at all. Now, instead of 5,000 couples
> > mating to populate the next generation, we only have about 3,600.
> > True, these are more fit, but there are fewer of them. If they still
> > have the same maximum reproductive rate of 4 per couple, then only
> > about 14,500 offspring will be produced instead of the 20,000 produced
> > by the original parent population. Of these, won't most go downhill
> > instead of uphill? Won't most of these cross the boundary and end up
> > below the 97 detrimental mutation threshold?
>
> If there is a hard threshold at 97, then considering a population that
> starts with exactly 97 deleterious mutations each is not a good model
> for thinking about the situation. Think about a hard limit at, say,
> 110 instead. You will have more than enough offspring surviving to
> produce the next generation, and the really highly loaded tail will be
> completely removed. The effect is to increase the overall fitness of
> the population at equilibrium, not decrease it.
You will have enough offspring to produce the next generation, but will
the number of those with fewer than 97 mutations increase or decrease
or stay the same given your reproductive rate of 4 per couple? It
seems to me that in each generation, the number of those with fewer
than 97 deleterious mutations will indeed decrease at this low rate of
reproduction regardless of where you draw the line, 110 or otherwise.
> What you describe is known as truncation selection, and is one extreme
> solution proposed to explain how humans could support a high
> deleterious mutation rate. (See the discussion section of the Nachman
> and Crowell paper, where they mention it.)
Yep . . .
> An ideal poplulation with
> 3.0 deleterious mutations per gen (2x reproductive capacity, Npop =
> 10,000, selection coeff = 0.01) comes to equilibrium at around 300
> deleterious mutation per individual. (I think I reported that number
> incorrectly previously.) With truncation selection operating, with a
> threshold of 97 mutations, the population reaches equilibrium with 87
> mutations per individual.
I'm not sure about this given a reproductive rate of only 4 per couple.
> > > Note that this is well within the reproductive capacity that I
> > > specified. Take the extreme case: all 7285 lightly loaded offspring
> > > form part of the next breeding generation, along with 2715 heavily
> > > loaded ones. Has the mean number of deleterious alleles in the
> > > population increased or decreased?
> >
> > The mean number has indeed decreased, but so has the absolute number of
> > offspring at or above the 97 starting point. What happens in the
> > subsequent generations?
>
> Mating in each generation tightens the distribution (combining two
> randomly selected individuals from the population reduces the variance
> on their average by 1/sqrt(2)), as does selection (which also biases
> the distribution towards the low end). Production of new offspring
> broadens the distribution, as random numbers of new mutations and
> random sampling of existing mutations introduce scatter. Since we
> started with zero variance in our thought experiment, initially the
> variance will increase, and then stabilize. Selection keeps the high
> tail from expanding indefinitely, and new mutations keep the low tail
> from expanding.
Yes, given a high enough reproductive rate. I just don't see how a
reproductive rate of 4 is going to result in equilibrium in this
particular case.
Consider another example of a steady state population of 5,000
individuals each starting out with 7 detrimental mutations and an
average detrimental mutation rate of 3 per individual per generation
(easier on my calculator). Given a reproductive rate of 4 offspring
per each one of the 2,500 couples (10,000 offspring), in one
generation, how many offspring will have the same or fewer detrimental
mutations than the parent generation?
The Poisson approximation shows that out of 10,000 offspring, only
~2,202 of them would have the same or fewer than the original number of
detrimental mutations of the parent population. This leaves ~7,798
with more detrimental mutations than the parent population. Of course,
in order to maintain a steady state population of 5,000, natural
selection must cull out 5,000 of these 10,000 offspring before they are
able to reproduce. Given a preference, those with more detrimental
mutations will be less fit by a certain degree and will be removed from
the population before those that are more fit (less detrimental
mutations). Given strong selection pressure, the second generation
might be made up of ~2,200 more fit individuals and only ~2,800 less
fit individuals with the overall average showing a decline as compared
with the original parent generation. If selection pressure is strong,
so that the majority of those with more than 7 detrimental mutations
are removed from the population, the next generation will only have
about 1,100 mating couples as compared to 2,500 in the original
generation. With a reproductive rate of 4 per couple, only 4,400
offspring will be produced as compared to 10,000 originally. Even if
the limit were drawn at more than 7 detrimental mutations, wouldn't the
final outcome still be the same?
It seems then that in order to keep up with this loss, the reproductive
rate must be increased or the population will head toward extinction.
Given a detrimental mutation rate of Ud = 3 in a sexually reproducing
population, the average number of offspring needed to keep up would be
around 20 per breeding couple (2e^Ud/2). While this is about half that
required for an asexual population (2e^Ud), it is still quite
significant.
How is my thinking off base here?
> Steve Schaffner s...@broad.mit.edu
> Immediate assurance is an excellent sign of probable lack of
> insight into the topic. Josiah Royce
Sean Pitman
www.DetectingDesign.com
Steve Schaffner
> Steve Schaffner wrote:
> >
> > If there is a hard threshold at 97, then considering a population that
> > starts with exactly 97 deleterious mutations each is not a good model
> > for thinking about the situation. Think about a hard limit at, say,
> > 110 instead. You will have more than enough offspring surviving to
> > produce the next generation, and the really highly loaded tail will be
> > completely removed. The effect is to increase the overall fitness of
> > the population at equilibrium, not decrease it.
>
> You will have enough offspring to produce the next generation, but will
> the number of those with fewer than 97 mutations increase or decrease
> or stay the same given your reproductive rate of 4 per couple? It
> seems to me that in each generation, the number of those with fewer
> than 97 deleterious mutations will indeed decrease at this low rate of
> reproduction regardless of where you draw the line, 110 or otherwise.
I don't have much intuition myself about whether the population would
survive or not. Running a simulation, I see that the population does
just fine with a reproductive rate of 4/couple, but goes extinct with
3/couple. (with a truncation threshold of 110 deleterious mutations).
Keep in mind that this is an unrealistic model, in which there is no
global loss of fitness for having deleterious mutations, just a cost
relative to the most fit living individual. (This can lead to some
weird situations, where adding a single, lightly loaded individual to
a population dramatically decreases the global fitness of the
population as a whole, causing a population crash. Not exactly real
biology.)
With a hard threshold at 7 and 5000 in the population, I find
extinction for reproductive rates of 7/pair or less. 8 seems to be
ok. (Again, that's in this toy model.)
> It seems then that in order to keep up with this loss, the reproductive
> rate must be increased or the population will head toward extinction.
> Given a detrimental mutation rate of Ud = 3 in a sexually reproducing
> population, the average number of offspring needed to keep up would be
> around 20 per breeding couple (2e^Ud/2). While this is about half that
> required for an asexual population (2e^Ud), it is still quite
> significant.
>
> How is my thinking off base here?
I'd say your thinking is pretty close to on target here. You do need
a higher reproduction rate with a low, hard limit on the number of
tolerable deleterious mutations -- not surprising, since the selection
is much more hard than soft in this scenario. I don't think it has to
be as high as you claim, but a realistic assessment of how high the
reproduction rate would have to be is not currently possible. It
requires more information about selection -- frequency-dependent,
lethel, globally deleterious, interacting, etc -- than we have.
David Wilson
talk.origins Steve Schaffner <s...@phosphorus.broad.mit.edu> wrote:
> "Seanpit" <seanpi...@naturalselection.0catch.com> writes:
>
... [snip] ...
> > around 20 per breeding couple (2e^Ud/2). ....
If Ud = 3 then 2e^(Ud/2) is nowhere near 20. It's about 9 (8.963378... to
be more precise).
> > While this is about half that
> > required for an asexual population (2e^Ud), ....
No it's not. It's precisely the _square root_ of _twice_ this latter figure,
which, for your value of Ud = 3, turns out to be about _one quarter_ of it.
> > ... it is still quite
> > significant.
> >
> > How is my thinking off base here?
>
> I'd say your thinking is pretty close to on target here. ...
>
I'd beg to differ. I don't dispute Dr Sean's or your conclusions about his
truncation selection model, but _for that model_ the argument he gives in
his final paragraph seems to me to be nonsensical.
The problem is that the requirement of 2*e^(Ud/2) offspring per breeding
couple which he cites as necessary to maintain a constant population size
is based on a calculation of the mean fitness, e^(-Ud/2), of a population
(relative to its theoretically "optimal" genotypes) _under completely
different assumptions_. Here are some of them:
i) all the relevant fitnesses interact multiplicatively;
ii) the magnitude of the selection coefficient for each mutation is
"fairly small" (i.e. "small" relative to 1; I'd guess that 10% should
be nearly small enough, even though this is quite "large" as selection
coefficients go);
iii) effectively all the mutations are recessive. (If effectively all
the mutations were dominant, the mean fitness relative to the optimal
genotype would be e^(-Ud), rather than e^(-Ud/2). With incomplete
dominance I presume the mean fitness would be somewhere between these two
figures, but I haven't actually carried out the calculation for that case).
iv) population size is effectively infinite;
("Effectively infinite" here really does mean unrealistically large.
Unless there actually did exist individuals in the population with
genotypes very close to the theoretically optimal, it would _not_ be
necessary for every breeding pair to produce as many as 2*e^(Ud/2)
offspring to maintain mutation-selection equilibrium and a constant
population size. But for Ud=3 and typical selection coefficients of
about -0.01, the frequencies of genotypes close to the optimal are
_minuscule_ at mutation-selection equilibrium, and it would require a
total population size in excess of 10^65 for the expected number of
such genotyptes to exceed 1).
Now, if you work out the mean number of deleterious mutations carried per
individual at mutation-selection equilibrium under these assumptions, it
turns out to be huge. If the total number of genes (i.e. loci) subject
to deleterious mutation is L, and all the deleterious mutations have
the same selection coefficient, -s, the mean is Sqrt( 2*L*Ud/s) (If
effectively all the mutations were dominant this figure would be Ud/s---
much smaller, but still quite large for typical values of the selection
coefficient). The behaviour of a model like this, with fitnesses
_interacting multiplicatively_ over many more than seven loci, and an
extremely large population size, seems hardly likely to me to bear much
resemblance to that of a truncation selection model with a threshold of
seven deleterious mutations and a population size of 5000.
Thus I can't see how the figure of 2*e^(Ud/2) offspring per pair,
required to maintain mutation-selection balance and constant population
size for _this_ model, can have any relevance whatever to the analysis of
the truncation selection model Dr Sean had proposed in his preceding
paragraph.
--------------------------------------------------------------------------
David Wilson
SPAMMERS_fingers@WILL_BE_fwi_PROSECUTED_.net.au
(Remove underlines and upper case letters to obtain my email address.
Steve Schaffner
> In article <ydlu0ck...@phosphorus.broad.mit.edu> on January 4th in
> talk.origins Steve Schaffner <s...@phosphorus.broad.mit.edu> wrote:
>
> > "Seanpit" <seanpi...@naturalselection.0catch.com> writes:
> > >
> > > How is my thinking off base here?
> >
> > I'd say your thinking is pretty close to on target here. ...
> >
>
> I'd beg to differ. I don't dispute Dr Sean's or your conclusions about his
> truncation selection model, but _for that model_ the argument he gives in
> his final paragraph seems to me to be nonsensical.
Allow me to rephrase my response: "I'd say you're pretty close to on
target in thinking that a truncation model with a low threshold will
require a larger reproductive capacity, and I'm going to ignore the
rest of your post because I'm tired of discussing the matter."
[...]
> Now, if you work out the mean number of deleterious mutations carried per
> individual at mutation-selection equilibrium under these assumptions, it
> turns out to be huge. If the total number of genes (i.e. loci) subject
> to deleterious mutation is L, and all the deleterious mutations have
> the same selection coefficient, -s, the mean is Sqrt( 2*L*Ud/s)
Are you sure about the factor of 2 in there? For a recessive
deleterious allele, the equilibrium frequency is sqrt(u/s), where u
is the per locus deleterious mutation rate. The mean number of
deleterious alleles is then L*sqrt(u/s) = sqrt(L*Ud/s), since
Ud = L*u. What am I missing?
> (If
> effectively all the mutations were dominant this figure would be Ud/s---
> much smaller, but still quite large for typical values of the selection
> coefficient).
More generally, N = Ud/hs for partial dominance, with fitnesses 1,
1-hs and 1-s. The case I simulated (without truncation) corresponds to
h = 0.5, with my selection coefficient defined as s/2. So my
simulation with Ud = 3 and s = .02 should have returned an equilibrium
value of 300. Which it did, which is nice since I've never bothered
to check the simulation.
--
David Wilson
In article <ydlr77m...@phosphorus.broad.mit.edu> on January 5th in
talk.origins Steve Schaffner <s...@phosphorus.broad.mit.edu> wrote:
> David Wilson <see_sig@for_my.address> writes:
>
> > In article <ydlu0ck...@phosphorus.broad.mit.edu> on January 4th in
> > talk.origins Steve Schaffner <s...@phosphorus.broad.mit.edu> wrote:
> >
> > > "Seanpit" <seanpi...@naturalselection.0catch.com> writes:
> > > >
> > > > How is my thinking off base here?
> > >
> > > I'd say your thinking is pretty close to on target here. ...
> > >
> >
> > I'd beg to differ. I don't dispute Dr Sean's or your conclusions about his
> > truncation selection model, but _for that model_ the argument he gives in
> > his final paragraph seems to me to be nonsensical.
>
> Allow me to rephrase my response: "I'd say you're pretty close to on
> target in thinking that a truncation model with a low threshold will
> require a larger reproductive capacity, and I'm going to ignore the
> rest of your post because I'm tired of discussing the matter."
>
That's pretty much what I thought you probably meant.
> [...]
> > Now, if you work out the mean number of deleterious mutations carried per
> > individual at mutation-selection equilibrium under these assumptions, it
> > turns out to be huge. If the total number of genes (i.e. loci) subject
> > to deleterious mutation is L, and all the deleterious mutations have
> > the same selection coefficient, -s, the mean is Sqrt( 2*L*Ud/s)
>
> Are you sure about the factor of 2 in there? For a recessive
> deleterious allele, the equilibrium frequency is sqrt(u/s), where u
> is the per locus deleterious mutation rate. The mean number of
> deleterious alleles is then L*sqrt(u/s) = sqrt(L*Ud/s), since
> Ud = L*u. What am I missing?
>
Possibly my understanding of the term "locus" is wrong. Your L appears
to be my 2*L. By "number of genes (i.e. loci)" I meant the number of
alleles in a haploid genome, whereas you appear to be taking it to mean
the total number in the diploid genome. Is the term "locus" really just
as ambiguous as the term "gene", or is my usage simply wrong? And does
this actually clear up the discrepancy, or is there still something
amiss?
> > (If
> > effectively all the mutations were dominant this figure would be Ud/s---
> > much smaller, but still quite large for typical values of the selection
> > coefficient).
>
> More generally, N = Ud/hs for partial dominance, with fitnesses 1,
> 1-hs and 1-s. ...
Thanks for the info. (This approximation must surely break down for h
sufficiently close to zero, though).
Steve Schaffner
> In article <ydlr77m...@phosphorus.broad.mit.edu> on January 5th in
> talk.origins Steve Schaffner <s...@phosphorus.broad.mit.edu> wrote:
>
> > David Wilson <see_sig@for_my.address> writes:
> > > Now, if you work out the mean number of deleterious mutations carried per
> > > individual at mutation-selection equilibrium under these assumptions, it
> > > turns out to be huge. If the total number of genes (i.e. loci) subject
> > > to deleterious mutation is L, and all the deleterious mutations have
> > > the same selection coefficient, -s, the mean is Sqrt( 2*L*Ud/s)
> >
> > Are you sure about the factor of 2 in there? For a recessive
> > deleterious allele, the equilibrium frequency is sqrt(u/s), where u
> > is the per locus deleterious mutation rate. The mean number of
> > deleterious alleles is then L*sqrt(u/s) = sqrt(L*Ud/s), since
> > Ud = L*u. What am I missing?
> >
>
> Possibly my understanding of the term "locus" is wrong. Your L appears
> to be my 2*L. By "number of genes (i.e. loci)" I meant the number of
> alleles in a haploid genome, whereas you appear to be taking it to mean
> the total number in the diploid genome. Is the term "locus" really just
> as ambiguous as the term "gene", or is my usage simply wrong? And does
> this actually clear up the discrepancy, or is there still something
> amiss?
No, your usage isn't wrong. I simply forgot (again) that I'm a
diploid. (It's possible there's some connection between this fact and
the kids I see running around my house every now and then.)
N_deleterious = 2Lq, and Ud = 2Lu. Sigh.
> > More generally, N = Ud/hs for partial dominance, with fitnesses 1,
> > 1-hs and 1-s. ...
>
> Thanks for the info. (This approximation must surely break down for h
> sufficiently close to zero, though).
Yes, the approximation assumes that terms like hq^2 (where q is the
allele frequency) can be neglected compared to hsq. See (for example)
http://www.zi.biologie.uni-muenchen.de/institute/zi/abtlgn/evolutionsbiologie/Lectures/Evobio/Evo8-Summary.pdf
RobinGoodfellow
Seanpit wrote:
> RobinGoodfellow wrote:
> > Seanpit wrote:
> > > RobinGoodfellow wrote:
> >
> > [snip]
> >
> > > > So, a rate of 1
> > > > substitution every 200 years, is the rate of 0.125 substitutions per
> > > > population per generation, assuming 25 year generations. Which means
> > > > that the rate of slightly beneficial mutations will be slightly lower
> > > > than 0.125 per individual per generation.
> > >
> > > Actually, the rate of slightly beneficial mutations will be quite a bit
> > > lower. The substitution rate of 0.125 per population per generation is
> > > based on a concentration of more than slightly beneficial mutations in
> > > the larger population into a smaller subpopulation over time with many
> > > of the more slightly beneficial mutations being lost before fixation.
> > > Out of all the millions of beneficial mutations that might have been
> > > realized in a given generation, only 0.125 make it to fixation. This
> > > means that the actual beneficial mutation rate is much much lower than
> > > the beneficial substitution rate.
> >
> > Wait a minute! You're claiming here that out of millions of beneficial
> > mutations, only 0.125 make it to fixation, and that the vast majority
> > of beneficial mutations are being lost before fixation. But this means
> > that beneficial mutation rate is actually *much higher* than the
> > substitution rate: i.e. for every X beneficial mutations per individual
> > per generation, only 1 per 8 generations makes it to substitution.
> > Could you be any more confused, Sean?
>
> I'm quite surprised that you came back with this comment Leonid. Don't
> you see? - The number of beneficial mutations in a population is a
> function of the size of the population as well as the rate of
> beneficial mutations in that population. For argument's sake, lets
> assume that we have a Ud/Ub ratio of 1000 to 1 in a population of 10
> billion. If the Ud rate were 3 per individual per generation, the Ub
> rate would be 0.003 per individual per generation. Out of the 10
> billion, how many Ub mutations would there be in one generation? The
> answer is 30 million - right? Out of this 30 million, how many will
> make it to fixation? Less than 1 of them - right?
My bad. It appears I misunderstood what you were saying. I thought
you were talking about the number of beneficial mutations per
individual, not per population.
Notice, however, that your paragraph above actually makes no argument
as to why the beneficial mutation rate *should* be much lower than the
beneficial substitution rate, merely that it can be.
> So, you see, the rate of beneficial mutations can indeed be very much
> lower than the rate of beneficial substitutions.
Indeed, though it would actually depend on the selective advantage
conferred by the beneficial mutation, and the population size.
Regardless, I would agree that I jumped the gun by estimating the
beneficial mutation rate from the substitution rate alone. Just as you
jumped the gun by quoting the figure of 1 beneficial *substitution*
every 200 years, and concluding that the beneficial mutation rate must
be extremely low compared to the deleterious rate.
> In fact, the authors of the papers you reference do use the 1000 to 1 ratio in their
> estimates of Ud vs Ub (Ref).
Actually, I never referenced that paper - you did. And the authors
simply assume this ratio for their analysis: they don't actually
attempt to derive it, especially not from human genomic data.
> Who's confused here? ; )
We both are, it would appear. :)
RobinGoodfellow
My bad. It appears I misunderstood what you were saying. I sought you
RobinGoodfellow
My bad. It appears I misunderstood what you were saying. I thought
Ron O
Steve Schaffner wrote:
> "Seanpit" <seanpi...@naturalselection.0catch.com> writes:
>
> > Steve Schaffner wrote:
> >
SNIP:
> >
> > Well, you clearly don't even understand the basic issues involved and
> > aren't contributing to this discussion at all. Sorry, but even Leonid
> > is doing a far better job here than you are.
I can't believe Sean wrote this. Is someone forging his posts? After
all the detailed explanation you've gone through this retort is just
pathetic.
I give you credit for trying. Sean is just incompetent when it comes
to this issue.
It is pretty sad when the insanity defense is the only legitimate
defense these people have.
Ron Okimoto
>
> I'm sorry, but you don't even understand what the issues are. And I
> dare say I've published more papers about selection than you have.
>
> > Exactly . . . How are you so sure you're right on this one? Where are
> > your references to back up your most unusual notions?
>
> Start by reading _Fifty years of genetic laod_, by Bruce Wallace.
>
RobinGoodfellow
[snip]
PS: Sorry if this appears multiple times - Google's being capricious
today.
Steve Schaffner
> Steve Schaffner wrote:
> > "Seanpit" <seanpi...@naturalselection.0catch.com> writes:
> >
> > > Steve Schaffner wrote:
> > >
>
> SNIP:
>
> > >
> > > Well, you clearly don't even understand the basic issues involved and
> > > aren't contributing to this discussion at all. Sorry, but even Leonid
> > > is doing a far better job here than you are.
>
> I can't believe Sean wrote this. Is someone forging his posts? After
> all the detailed explanation you've gone through this retort is just
> pathetic.
He wrote this before my detailed explanations (or at least before most
of them). I think you must have received this post out of order.
Walter Bushell
"Ron O" <roki...@cox.net> wrote:
> I can't believe Sean wrote this. Is someone forging his posts? After
> all the detailed explanation you've gone through this retort is just
> pathetic.
>
> I give you credit for trying. Sean is just incompetent when it comes
> to this issue.
>
> It is pretty sad when the insanity defense is the only legitimate
> defense these people have.
>
> Ron Okimoto
It's a shame there isn't a sanity clause
--
"The power of the Executive to cast a man into prison without formulating any
charge known to the law, and particularly to deny him the judgement of his
peers, is in the highest degree odious and is the foundation of all totali-
tarian government whether Nazi or Communist." -- W. Churchill, Nov 21, 1943