Re: Haldane's Dilemma - clarifications - and Felsenstein
- From: joe@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx (Joe Felsenstein)
- Date: Sat, 17 Jun 2006 13:27:00 -0400 (EDT)
In article <e6v5gh$682$1@xxxxxxxxxxxxxxxxxxx>,
Walter ReMine <science@xxxxxxxx> wrote:
The paper clarifies many long-standing confusions about the cost of
substitution, which is the fundamental concept behind Haldane's
Dilemma. The paper was peer-reviewed by evolutionary geneticists,
including James Crow and Warren Ewens, who acknowledge it is correct.
Walter fails to mention that they also recommended rejection, and gave
reasons for doing so. The above makes it sound as if they urged publication.
The paper was also peer-reviewed (at a different journal) by Joe
Felsenstein, who (mistakenly) claimed it is incorrect. Felsenstein
published his peer-review (at ARN.org, through an intermediary). A
response to Felsenstein's review is available here:
http://SaintPaulScience.com/Felsenstein_comments.htm
The "response" continues to repeat the same mistaken assertions that
I complained of in the review. It continues to conflate different
notions of costs and claim the paper clarified the issue.
I stand by my review as written. I see nothing in Walter's response
that would cause me to change anything in it.
And by the way, the mysterious process of publishing "through an
intermediary" (Allan Force) was simply that
1. Walter was complaining about my review and characterizing it, but
did not make it available. He apparently thought that copyright
issues would prevent him from doing so.
2. I made my review available as a PDF to Allan Force, who was debating
Walter in that forum, and authorized him to post it.
3. Allan made a couple of typos in posting it, Walter complained loudly
and made it sound sinister that Allan hadn't posted the exact review.
4. So I sent Allan the text used to create the PDF and he posted it.
Walter suddenly stopped referring to the accuracy issue.
If someone here other than Remine (or one other poster to whom I am not
responding for reasons of his past insults) feels that his response to my
review needs discussion, I'll discuss it.
Oh, and here's the mysterious review, written originally as a review of
substantially the same paper that Walter just posted links to, when it was
submitted to Journal of Theoretical Biology in 2003:
Review of "Cost Theory and the Cost of Substitution
-- a clarification" by Walter Remine.
This paper is written in a contentious, supercilious, nonacademic style, and
argues that the notion of a "cost of substitution" has been misunderstood and
misapplied by many workers, who have been confused. It argues that the correct
definition is the reproductive excess required by a given situation.
The paper has the defect that it assumes that all authors have been attempting
to describe the same notion, and that their differences have resulted from
misunderstandings of the correct concept. Remine argues strenuously against
the idea that if there is too little reproductive excess the population will
go extinct. Instead, he argues that too little reproductive excess to pay the
cost implies that the scenario is implausible. Remine also argues that any
trait that increases in frequency in the population implies a cost, even if
its increase is purely the result of genetic drift.
In fact, it is not at all clear that different researchers had the same
objective in mind. Haldane (1957) did not use the phrase "cost of
substitution" (his title mentions "cost of natural selection"). Ewens put
forward a cost that was nonzero for pure genetic drift, other authors' costs
were zero for change by genetic drift.
Remine's treatment is thus inadequate in its historical treatment of others'
work, and is also inadequate internally. Take the idea that susbtitution
requires reproductive excess. Suppose that we have a (haploid) species with
no reproductive excess. Suppose that the environment changes so that all
individuals have 20% less reproduction, except for 10% of them who have a
particular allele, and those continue to barely replace themselves. A little
consideration will show that the substitution will happen, and the population
will end up 90% smaller. But if there is even a slight reproductive excess,
then with enough time the population will ultimately recover its numbers.
There is then no lower limit on the reproductive excess needed.
However, Remine defines the cost in terms of the number of copies of the
allele. He adds up the reproductive excess over the generations involved, so
that in the above case he would arrive at a finite total of the reproductive
excess. The population may be in no difficulty, but if the researcher thinks
that the change of number of copies of the gene is possible in that amount of
time, they must be able to argue that there is that much reproductive excess.
The problem for the population in such a scenario comes when we have, not one
substitution, but a regular flow of such events. Then, if there is no
reproductive excess, the population drops by 90% each time, and ends up going
extinct. It is thus possible to define a cost that is the cost of avoiding
extinction, in spite of Remine's strenuous arguments against such a notion. A
certain reproductive excess *is* [italics in original] necessary when there
are recurring events, but it is to avoid extinction, not to allow
substitution. Remine mentions that repeated substitutions are involved in
several other people's definitions, but then simply declares that "that
interpretation is faulty", without saying why. He sounds as if someone has
issued him a certificate stating that his interpretation of the cost is the
correct one and all others are wrong.
Remine's treatment of selective neutrality is also confusing. Aside from
declaring that in such cases there is a cost but that there are "special
mechanisms that allow high *overall* [italics in original] rates of
substitution". The reader may be confused, as they may imagine that neutral
substitutions just happen, as a result of randomness of mutation, birth,
death, and genetic segregation. Saying that there are special mechanisms
present implies that these are adaptations (or divine interventions) designed
to make Motoo Kimura feel happy.
One minor matter should also be mentioned: Remine has the habit of presenting
a table of frequencies of different genotypes (as on pages 11 and 17) as an
equation. Instead of writing p and q as the two frequencies that also
add to 1, he writes p+q = 1 . This is trivially true for any two frequencies
of alternative genotypes, irrespective of whether they are the correct
frequencies. (For example, 2p + (1-2p) = 1 , and that statement is correct
even when the frequencies of the two alleles are not 2p and 1-2p ). This
is easily corrected by replacing the summations by tables.
I conclude that, although Remine's algebra is correct, his description of how
his work relates to the costs and loads defined by others is seriously wrong.
Stylistics aside, it thereby makes a negative contribution to discussion of
these issues, and I cannot recommend its publication.
I should be identified to the author.
Joe Felsenstein
----
Joe Felsenstein joe@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
Department of Genome Sciences and Department of Biology,
University of Washington, Box 357730, Seattle, WA 98195-7730 USA
.
- Follow-Ups:
- Re: Haldane's Dilemma - clarifications - and Felsenstein
- From: Perplexed in Peoria
- Re: Haldane's Dilemma - clarifications - and Felsenstein
- References:
- Haldane's Dilemma - clarifications - and Felsenstein
- From: Walter ReMine
- Haldane's Dilemma - clarifications - and Felsenstein
- Prev by Date: Re: Paper: Environmental Coupling of Selection and Heritability
- Next by Date: DHEA and Brain Development
- Previous by thread: Re: Haldane's Dilemma - clarifications - and Felsenstein
- Next by thread: Re: Haldane's Dilemma - clarifications - and Felsenstein
- Index(es):
Relevant Pages
|
|
