Haldane's Dilemma
Daniel MacArthur
resolution of 'Haldane's Dilemma'. I understand that there was a discussion
of this topic on this forum some time back, but I have been unable to access
the archives of this discussion through Deja.
[moderator's shudder: Now, let us be very very careful here. There
is a preposterous book which alleges that Haldane's dilemma renders
evolution impossible; this conclusion has been refuted any number of
times in this newsgroup as well as in talk.origins. I invite the
original debunkers (Joe, and others) to post on this thread, but
be warned: if there's any HINT of dragging in creation/evolution
arguments I will delete such articles silently. - JAH]
My question relates to the evolution of the hominid lineage. I have been
told (and many of you can probably guess the ultimate source of this
information) that Haldane's calculations show that only 1,700 or so
mutations could have been fixed in the human lineage since the primate/human
split, even if this event had occurred a generous 10 million years ago. This
means that human and chimp genomes should differ by a total of only 3,400 or
so mutations. Clearly this is not the case - the actual figure, by my
back-of-an-envelope calculations, should be closer to 50 million (assuming
all mutations are point mutations, which is unreasonable but beside the
point for the moment).
My first thought was that the 3,400 or so mutations would represent only the
beneficial mutations being acted on by selection, and that the remaining
genome difference could be accounted for by neutral mutations. I assumed
that neutral mutations would incur no fixation cost as there was no
selection acting to maintain them in the population and thus their fixation
by random drift would incur no reproductive cost. I have been informed that
this is not the case, and that neutral mutations in fact incur the same
fixation cost as positively selected mutations. While I am sceptical of this
notion, I find so much of population genetics so counter-intuitive that I am
unable to definitively deny it.
I figured that if the bulk of the genome difference between chimps and
humans was due to neutral mutations, then a total of 3,400 non-neutral
mutations would be more than enough to account for the morphological
differences between the two lineages (assuming these mutations affected
developmentally active genes like the Hox family this seems more than
generous). But do neutral mutations escape fixation cost? I'm guessing that
if Haldane's Dilemma is still an ongoing issue then this is not the case,
but could somebody explain why not?
And finally, does anyone have a list of possible resolutions for Haldane's
Dilemma? I imagine that there will not be one single definitive solution,
but a myriad of smaller additive solutions. If anyone could provide any of
these solutions I would be deeply grateful.
Oh, and the simpler the better. The person I am discussing this matter is,
as you might imagine, not as well educated in the field of genetics as many
of you in this group. His grasp of Haldane's dilemma is taken pretty much
wholesale from another source (although this source has also provided him
with convenient ripostes to most of my suggestions for a resolution).
If there's any simple way of pointing out possible solutions to the
'Dilemma' I'd love to hear it. Thanks in advance.
Daniel MacArthur.
Jois Lombard
> If there's any simple way of pointing out possible solutions to the
> 'Dilemma' I'd love to hear it. Thanks in advance.
>
> Daniel MacArthur.
Did you look at this:
http://www.gate.net/~rwms/haldane1.html
Jois
Tim Tyler
[Haldane's dilemma]
: I assumed that neutral mutations would incur no fixation cost as there
: was no selection acting to maintain them in the population and thus
: their fixation by random drift would incur no reproductive cost. I have
: been informed that this is not the case, and that neutral mutations in
: fact incur the same fixation cost as positively selected mutations.
[...]
: But do neutral mutations escape fixation cost? I'm guessing that
: if Haldane's Dilemma is still an ongoing issue then this is not the case,
: but could somebody explain why not?
What is "fixation cost"?
Mutations can become fixed simply by the population going through a size
bottleneck. There is not necessarily any "cost" - except in biodiversity.
--
__________ http://alife.co.uk/ http://mandala.co.uk/
|im |yler t...@cryogen.com http://hex.org.uk/ http://atoms.org.uk/
Joe Felsenstein
Daniel MacArthur <dma...@hotmail.com> wrote:
>I was just wondering if anyone knew of any recent developments in the
>resolution of 'Haldane's Dilemma'. I understand that there was a discussion
>of this topic on this forum some time back, but I have been unable to access
>the archives of this discussion through Deja.
>
>[moderator's shudder: Now, let us be very very careful here. There
>is a preposterous book which alleges that Haldane's dilemma renders
>evolution impossible; this conclusion has been refuted any number of
>times in this newsgroup as well as in talk.origins. I invite the
>original debunkers (Joe, and others) to post on this thread, but
>be warned: if there's any HINT of dragging in creation/evolution
>arguments I will delete such articles silently. - JAH]
The discussion on this newsgroup was from late 1997 to mid-1998.
Walter Remine (author of a self-published book "The Biotic Message"
which advertises his views on this) argued that Haldane's Dilemma
shows that one cannot have more than some smallish number of mutations
in evolving humans from their common ancestor witch chimps. This
argument of Remine's is widely touted by creationists on talk.origins.
Remine argues that he has a comprehensive understanding of the
population genetics theory, and that population geneticists don't,
or don't want to.
As has already been mentioned here, the web site by Robert Williams
(http://www.gate.net/~rwms/haldane1.html) takes Remine severely to
task. In our discussion here, I think Remine was fairly thoroughly
demolished. It took a bunch of posters here a while to get it out
that Remine was actually talking, not about Haldane's Dillema but
about an unrelated phenomenon called the Hill-Robertson effect. This
does argue that in a piece of nonrecombining DNA, not too many
favorable mutations can substitute in a period of time. However as
soon as we allow recombination in, the effect dies away. So
Remine's argument works as long as humans and chimps reproduce asexually.
Not otherwise.
As for Haldane's dilemma, it was discussed some in the population
genetics literature in the late 1960's through early 1970's when
papers on it died away, with a variety of views still extant. The
best and clearest paper on it was my own, in American Naturalist in 1971.
In short:
1. If there are neutral mutations happening they cause no
"substitutional load" or "cost of natural selection", and
any number of them can be occurring,
2. If deleterious environmental changes occur, they reduce the
fitness of the organism. Even if it responds by evolving
so as to compensate for them, this takes time. Haldane's
"cost of natural selection" argues that, for a given amount
of reproductive excess available to the species, there is a
limited rate of substitution that can occur this way without
the reduction in fitness pushing the rate of reproduction
below replacement of the population. However ...
3. If the substitions are advantageous they cause no load,
as they increase the reproductive excess more than enough to
compensate. Haldane's cost cannot set a limit on advantageous
mutations. Nor can we necessarily observe the resulting
increase in fitness, because
4. There are also likely to be deleterious changes in the environment
that are *not* compensated for by substitutions, as well.
So basically there is no hard way from Haldane's Dillemma to calculate
any upper limit on total substitutions, or even substitutions brought
about by natural selection in the evolution of humans.
That's a quick summary of what was a fairly discursive, irritable
controversy on sci.bio.evolution. Too bad that deja.com has
"temporarily" made unavailable all archives before May, 1999.
Perhaps someone has an archive of the whole discussion -- I have
parts of it.
--
Joe Felsenstein j...@genetics.washington.edu
Dept. of Genetics, Univ. of Washington, Box 357360, Seattle, WA 98195-7360 USA
phillip smith
G.C. Williams "Natural Selection Domains, Levels and Challenges" pp 143
her it is discussed quite well and the solutions posed by Wallace 1987
examined
Here is a brief summary of Haldane's dilemma as i undestand it.
Say and individual has 10,000 loci. Say at on average 1% of those loci are
of sub optimal fitness say 99% then the fitness on average of a member of
the population is 0.99^100 = 0.36603234127323. So we can see that every
individual is very unfit. And it might be though that they should die
What is the solution to the problem. There are several solutions
1) there is no problem it is just a mathematical artifact
2) The method of calculation is wrong their may be interactions between
genes that change the formula
I think there is a simple explanation. If you use relative fitness, which
is the only one that counts, then the problem goes a way as the mean
relative fitness is the same as the mean fitness. This not a simple trick of
the maths. evolution can only work on changes in gene frequency. You could
argue that the decrease in "absolute fitness" is real because it represents
the ability to compete against other species i.e to avoid predation or
competing for prey or other limited resources. However since all species
would suffer from Haldanes Dilemma this is cancelled out. If I am at 30 %
fitness as long as all my competitors are at 30 % fitness then I am equally
fit. Is haldane's dilemma real? Yes it is genetic load and I have posted
links to my work in this area to show that organisms should organize their
genome so as to reduce genetic load by increasing the effects of mutations.
this is the process of Selfish helping.
The other way to look at this is looking at genome space. Say you have a
genome of 1000 nucleotides ( a little small but its easier to work with)
then there are 4 to power of 1000 possible genomes = 1.148130695274255 e602
thats 1 with 602 zeros behind it. WE have no way of knowing if any species
with a genome of that miniscule size is any where near optimal. So it is
quite likely that extant species are of several magnitudes less fit then the
hypothetical optimum. This shows why species do respond so well to selection
because their is enourmous potential untapped
Phillip Smith
phills@(no spam)ihug.co.nz
http://www.applied-evolution.co.nz
http://www.applied-evolution.co.nz
Daniel MacArthur
Tim Tyler <t...@cryogen.com> wrote:
> Daniel MacArthur <dma...@hotmail.com> wrote:
>
> [Haldane's dilemma]
>
> >I assumed that neutral mutations would incur no fixation cost as
> >there was no selection acting to maintain them in the population and
> >thus their fixation by random drift would incur no reproductive
> >cost. I have been informed that this is not the case, and that
> >neutral mutations in fact incur the same fixation cost as
> >positively selected mutations.
>
>
> Mutations can become fixed simply by the population going through a
>size bottleneck. There is not necessarily any "cost" - except in
>biodiversity.
I'm not entirely sure, to tell you the truth. This is the way it was
explained to me:
For a trait to be fixed in a population by selection there has to be a
selective removal of organisms that do not possess the trait (this
makes sense, at least). Thus the movement of a trait to fixation by
selection incurs a "reproductive cost" on the population equivalent to
the loss of those organisms. This cost is above and beyond the
background death rate due to chance and deleterious mutations.
This more or less makes sense to me (although I, along with many
biologists, would dispute the last point - c.f. "soft selection"). But
the notion that neutral traits would also incur a similar cost seems
rather ludicrous to me, and that is what is being suggested.
Daniel MacArthur.
In Defence of Science:
http://www.stormloader.com/users/mesk/index.html
Sent via Deja.com http://www.deja.com/
Before you buy.
Daniel MacArthur
I'll present it to the aforementioned advocate of Haldane's Dilemma and
see how he responds.
While I'm at it, I guess you're the best person to ask this question:
What is the difference between "cost" and "load"? There is currently a
rather nasty and (on both sides) uninformed debate on this topic on a
discussion board I occasionally visit. One person is claiming
vigorously that the two terms are quite distinct and that confusion
between them is rampant among evolutionary scientists. How would you
define the two terms, and is there a lot of confusion about the
distinction between them?
Daniel MacArthur.
--
Daniel MacArthur
Jois Lombard <lomb...@erols.com> wrote:
> Daniel MacArthur wrote:
>
> > If there's any simple way of pointing out possible solutions to the
> > 'Dilemma' I'd love to hear it. Thanks in advance.
> >
> > Daniel MacArthur.
>
> Did you look at this:
>
> http://www.gate.net/~rwms/haldane1.html
Yes, I found this site on my initial search of the web. Allegedly it
contains a number of factual errors in its critique of Remine's book,
but as I don't have a copy of the book I can't check them out for
myself. This is annoying but unavoidable as I have no intention
whatsoever of forking out the money for it.
It also didn't come across as a decisive and satisfying explanation for
Haldane's Dilemma. I (probably rather naively) imagine that there is an
elegant solution to the problem that will end the issue once and for
all, and this site wasn't it. Maybe I'm asking too much. :-)
Daniel MacArthur.
Christopher Mosley
> That's a quick summary of what was a fairly discursive, irritable
> controversy on sci.bio.evolution. Too bad that deja.com has
> "temporarily" made unavailable all archives before May, 1999.
> Perhaps someone has an archive of the whole discussion -- I have
> parts of it.
These resources may become "permanently" unavailable. Services that
archive usenet are disappearing and deja's archives might also go that
way. So:
Keep the Deja Archive Alive
http://www.PetitionOnline.com/dejanews/
J. W. Edser
Joe Felsenstein wrote:-
Daniel MacArthur wrote:-
> >DM:-
> >I was just wondering if anyone knew of any recent developments in the
> >resolution of 'Haldane's Dilemma'. I understand that there was a
> >discussion
> >of this topic on this forum some time back, but I have been
> >unable to access
> >the archives of this discussion through Deja.
> >[moderator's shudder: Now, let us be very very careful here. There
> >is a preposterous book which alleges that Haldane's dilemma renders
> >evolution impossible; this conclusion has been refuted any number of
> >times in this newsgroup as well as in talk.origins. I invite the
> >original debunkers (Joe, and others) to post on this thread, but
> >be warned: if there's any HINT of dragging in creation/evolution
> >arguments I will delete such articles silently. - JAH]
> JF:-
> The discussion on this newsgroup was from late 1997 to mid-1998.
> Walter Remine (author of a self-published book "The Biotic Message"
> which advertises his views on this) argued that Haldane's Dilemma
> shows that one cannot have more than some smallish number of mutations
> in evolving humans from their common ancestor witch chimps. This
> argument of Remine's is widely touted by creationists on talk.origins.
> Remine argues that he has a comprehensive understanding of the
> population genetics theory, and that population geneticists don't,
> or don't want to.
> As has already been mentioned here, the web site by Robert Williams
> (http://www.gate.net/~rwms/haldane1.html) takes Remine severely to
> task. In our discussion here, I think Remine was fairly thoroughly
> demolished. It took a bunch of posters here a while to get it out
> that Remine was actually talking, not about Haldane's Dillema but
> about an unrelated phenomenon called the Hill-Robertson effect. This
> does argue that in a piece of nonrecombining DNA, not too many
> favorable mutations can substitute in a period of time. However as
> soon as we allow recombination in, the effect dies away. So
> Remine's argument works as long as humans and chimps reproduce asexually.
> Not otherwise.
>snip<
JE:-
My understanding was that Walter ReMine was
arguing that a _minimal_ basic reproductive cost
for any DNA substitution must exist quite independently
to recombination or anything else for that matter,
that must always be met, no exceptions, which
cannot be in this case.
I am not saying I agree/disagree with ReMines
argument, but I am not sure Felsenstein is
correctly representing ReMine's argument, above.
I think certain concepts have to be clarified
so I will ask JF or any others here, if they
could please clarify/correct my understanding/
misunderstanding of the following concepts:-
1. Substitution.
If chimps differ by a certain percentage in
DNA to humans and they are said to have the
same ancestor then that percentage of
the DNA must in each case have been substituted,
nucleotide by nucleotide, by some mechanism,
irrespective of what that DNA does, or does not
do.
Question A:-
Roughly, how many individual
nucleotides are suggested to make up the
estimated percentage difference (and what
is that estimated percentage difference)
between these two species genomes suggested
during Haldane's era and now, today?
What mechanism/s are proposed to supply
this missing number of nucleotides and how
are any of these views tested when:
(a) One substitution calculation is for
one nucleotide substitution only, and not
for one gene's substitution.
(b) Each gene may have from just one to ten, or
hundreds or even thousands (or more) nucleotides
bases that are different per gene, in the stated
percentage that is different.
(c) A substituted gene must be represented
commonly throughout one species and not be found
commonly in the other species, to be considered
one substituted gene.
(d) One substituted nucleotide can only exist
in one fertile organism of the said species
to be counted as one valid substitution event.
If the nucleotide substitution exists in a non
fertile organism, a somatic cell or anywhere
else it cannot be passed on within the _organism_
germ line (chimps and man are defined as organisms),
is at a dead end, and is not one valid substitution
for the argument.
2. Reproduction.
Haldanes dilemma asks if enough time
as passed to allow enough fertile organism
generations in both man and chimp can exist
to reproduce enough germline DNA,
to predict the substituted number of
fixed nucleotides observed per organism
today, as asked in question A.
The disagreement:-
W. ReMine suggests that not enough time
has passed but J. Felsenstein suggested
that is has.
(a) How many fertile organism units
in each case are needed to substitute one
nucleotide as a maximum and minimum measure?
(b) Using (a) what is the maximum number
of fertile organism units that must have
been reproduced as minimum and a maximum
and over what number of organism generations,
to have produced all the nucleotide
substitutions necessary?
(c) Are any of the results of (b) biologically
credible in units of time, if they are compared
to the reproductive rates and lifespans that
are known today, for each species?
John Edser
Independent Researcher
PO Box 266
Church Pt
NSW 2105
Australia
Joe Felsenstein
Daniel MacArthur <dma...@hotmail.com> wrote:
>While I'm at it, I guess you're the best person to ask this question:
>What is the difference between "cost" and "load"? There is currently a
>rather nasty and (on both sides) uninformed debate on this topic on a
>discussion board I occasionally visit. One person is claiming
>vigorously that the two terms are quite distinct and that confusion
>between them is rampant among evolutionary scientists. How would you
>define the two terms, and is there a lot of confusion about the
>distinction between them?
I am not aware of any distinction between them. The load due to some
phenomenon is the excess reproduction needed (above mere replacement
of the population) needed when the phenomenon is present, to prevent
extinction.
--
laser...@my-deja.com
misrepresentations, there isn't enough space to cover them all. Here
are some:
>As has already been mentioned here,
>the web site by Robert Williams
>(http://www.gate.net/~rwms/haldane1.html)
>takes ReMine severely to task.
On the contrary, Williams's website misrepresents the situation
grievously. It pretends to critique my book, when he has never read my
book. I pointed this out repeatedly on the Internet newsgroups. His
website abundantly misrepresents my book (by commission and omission).
Misrepresentation does not get bolder than this.
[moderator's note: Groannnn. Here we go again. Walter, the fact that
you constantly refer people to discussions in your book, which you
sell yourself, rather than making your case in a public forum, does
not prove you to be a charlatan. But it certainly leaves the door
open. I hereby grant you the right to quote sufficiently large swaths
from your book if you think it'll make your argument - but of course
this will allow critics to refute said argument. I'm tired of
references to this mysterious book as providing the answer; it's
time for you to put up or shut up. Call it censorship if you like,
but if you want to argue about the topic, then ARGUE ABOUT THE
TOPIC. Don't plant thinly-veiled advertisements with a paragraph
of hand-waving. I await constructive, lucid, COMPLETE postings from
you; otherwise, don't waste my time. - JAH]
>In our discussion here, I think ReMine was fairly
>thoroughly demolished. It took a bunch of posters
>here a while to get it out that ReMine was actually
>talking, not about Haldane's Dilemma but about an
>unrelated phenomenon called the Hill-Robertson effect.
Not true. In our previous discussion here on sci.bio.evolution, I
pursued the central concept in Haldane's Dilemma -- the cost of
substitution. I demonstrated that evolutionists do not agree among
themselves on its fundamental meaning or what the actual problem is. I
did this with a key example -- Joe Felsenstein himself, who holds that
beneficial substitution, even taken one at a time in isolation, can
incur ZERO cost of substitution. Stripped of the extraneous matters
(sexual reproduction, diploidy, various selection models, etc.) the
central issue doesn't get any cleaner, or simpler, than this.
I say Felsenstein is wrong. Nothing can ever go from few to many,
without reproductive excess. Absolutely, positively, no exceptions.
This is the central point captured, and quantified, in the cost of
substitution concept, (though still frequently garbled in the
literature). Substitution (i.e. going from few to many) incurs a cost
of substitution. Felsenstein is wrong.
This point is fundamental, nothing in Haldane's Dilemma is more
fundamental than this. Yet evolutionists here on this newsgroup
contradicted each other on the matter, most of them implicitly
contradicting Felsenstein (and his published papers). Despite my
repeated encouragement and provocation, they failed to pursue a
resolution on this simple matter, and let it drop, much as they have for
the last forty-some years. I kept bringing the discussion back to the
cost of substitution. Felsenstein took an evasive posture, repeatedly
trying to divert to other topics, such as asexual versus sexual
reproduction, Hill-Robertson effect, etc. He continues that posture
today.
I encourage evolutionists to resolve this simple matter: What is the
cost of substitution? What does it mean? And, for basics, in the
simplest possible model (say, an asexual haploid, and a single
beneficial substitution occurring in isolation) can it incur zero cost?
While I have arguments on this matter (given in my book), I point out
that evolutionists themselves do not resolve these simple issues. It
nicely demonstrates my case -- Haldane's Dilemma was never solved, it
was merely garbled and brushed aside for forty years.
======================================
> ReMine argues that he has a comprehensive understanding
>of the population genetics theory, and that population
>geneticists don't, or don't want to.
I did not say any such thing. Felsenstein made that up.
-- Walter ReMine
_The Biotic Message_
http://www1.minn.net/~science
http://www.deja.com/getdoc.xp?AN=679857459
Tim Tyler
:> >I assumed that neutral mutations would incur no fixation cost as
:> >there was no selection acting to maintain them in the population and
:> >thus their fixation by random drift would incur no reproductive
:> >cost. I have been informed that this is not the case, and that
:> >neutral mutations in fact incur the same fixation cost as
:> >positively selected mutations.
:>
:> What is "fixation cost"?
:>
:> Mutations can become fixed simply by the population going through a
:> size bottleneck. There is not necessarily any "cost" - except in
:> biodiversity.
: I'm not entirely sure, to tell you the truth. This is the way it was
: explained to me:
: For a trait to be fixed in a population by selection there has to be a
: selective removal of organisms that do not possess the trait (this
: makes sense, at least). Thus the movement of a trait to fixation by
: selection incurs a "reproductive cost" on the population equivalent to
: the loss of those organisms. This cost is above and beyond the
: background death rate due to chance and deleterious mutations.
Whatever the "cost" of fixation, it can be afforded. All you need to
change genes is some radiation. All you need to fix most of the changed
copies is to wipe out all but two individuals.
The former process takes time - if it is to be done without
causing irreparable damage to the rest of the genome - but the
latter process can be close to instantaneous.
The former process is what should be concentrated on if a claim that
certain genetic changes could not have happened in the available time
span is made. How many genes are fixed seems irrelevant.