Talk:Genetic drift/Archive 2

From Talk:Genetic drift

Before I reitterate my objection to one element of the article, I want to make a comment about process. I realize that in effect there was an "edit war" between 168 and myself. Nevertheless, it should be clear not only that I did not automatically revert any change 168 made; as I worked on the article I incorporated elements of his contributions. I did this several times, and each time 168 reverted automatically. He even criticized me once for using a phrase (gene pool) that he actually introduced to that particular text. Not only did he not recognize that I was trying to work with him; the mere fact that I accepted something he wrote magically invalidated what he wrote. As 168 makes clear above in a remark to Mav, he reverts my contributions before even looking at them. I consider this inappropriate behavior.


 * What I did was to raise a question about the word gene pool. To the extent it was a criticism, it was less of you, SLR, than it was aof me. Obviously I remembered I had used it first. Obviously also I would not describe "the process" as you have. 168...

To put it simply, I do not see this as "my" article; I see all participants in this discussion as contributing to an article that belongs to no one. In contrast, it seems that 168 believes he owns this article; any change of his text is reverted, and he makes no effort at all to introduce or keep points raised by others. 168 says that I was "rocketing away" in my own direction -- a statement that could be made only by someone who does not read my changes and does not see how I have tried to incorporate or keep elements written by others.

For example, he deleted my discussion of Hardy-Weinberg entirely. Now, I have no attachment to what I wrote. If 168 or Mav or Lexor or anyone else thinks they can make the point more clearly, go ahead and edit what I wrote. But why delete content wholesale?

This is my substantive point concerning the article: sentences like the following,
 * Instead these outcomes depend on factors like the weather and being in the wrong place at the wrong time.

and
 * But with a small number of individuals, a lucky break for one or two causes a disproportionately greater deviation from the expected result.

Are at best highly misleading. Yes, outcomes that depend on the weather or being in the wrong place at the wrong time may cause drift (one has to be careful to eliminate natural selection first). Yes, in asexual organisms differential fecundity and morbidity (phrase it however you want to in the article. Mav was right that it was I who introduced these terms to the article, but in a weird way 168 was right to wonder whether Mav was faulting him, as I copied these words from a definition 168 provided in the Talk section) are the primary causes of drift. But in regard to sexual organisms, the primary cause of drift is something else -- something which I tried to explain in the section that 168 deleted. If he thought the section was poorly worded, he should have rewritten it. But if he simply did not understand the section, he should not delete it -- rather than insisting on his own ignorance, he should take an opportunity to learn something. Drift among populations of sexual organisms can be a result of good or bad luck for individuals. But drift is not primarily a phenomena concerning individuals, it concerns the frequencies of genes themselves. This is one problem with the word luck, which we think of as either good or bad. That a couple, each with genes for brown and blue eyes, may have four children with blue eyes may be considered "luck" but in biological terms it is not a matter of good or bad luck; "random" is a better term. Believe it or not, 168 raised this point quite well in an earlier version that used the example of a coin toss. But then I incorporated that analogy into the article -- not realizing that my acceptance of that analogy was the kiss of death. 168 has deleted it ever since. Slrubenstein


 * Amen to that. Tannin

Where to begin? I'm not going to get to all the points SLR made explicitly in one sitting, because there are too many, but our pictures of the situations are different in lots of ways. First, SLR, if drift can occur without sexual reproduction, then it is a distraction from the essence of it to portray sexual reproduction as if it were.168...
 * You should read my most recent revision where -- partially in deference to you -- I forground drift with asexual reproduction.

Secondly, on what basis do you assert that this gamete business is more influential in drift in species that do reproduce sexually? I'm open to believing that it might be, but you cite no source. 168...
 * Well, so far no point in the article invokes a source. But since you ask, read any standard textbook in physical anthropology or population genetics.  By the way, I have no idea what you mean by this "gamete" business.  I never used the term gametes in any of my contributions, only you have used it -- so you should ask yourself what you mean.  But I can tell you this, my point about drift in sexual reproducing populations is not about "gametes."  Perhaps you are a little confused about the terms.  "Gametes" are reproductive cells (in animals, eggs and sperm) developed from precursor cells in ovaries and testes.  "Drift" doesn't have to do with gametes.  Again, please read the article.  If you do not understand it because it needs more development, please tell us and we will work on it. But since you do not understand the science, please defer to scientists who do when it comes to the content.


 * I know what a gamete is. Your eye color allele scenario (which isn't up to date on the genetics of eye-color by the way) has to do with deviations from the odds that alleles of the grandparents of individuals are equally represented in the sperm and eggs (gametes) that come together to make them. Since the father's sperm is diverse (see meiosis) and so are the mother's eggs, equal representation of grandparent alleles in the offspring is a matter of chance. Dimly as I recall it, anyway. I wouldn't have written on this subject off the top of my head if you weren't having trouble imagining what I might be referring to with "gametes." 168... 20:19 20 May 2003 (UTC)

Thirdly, your problems with my invoking individuals suggest to me that your understanding of the statistics here is not very deep. 168...
 * No, your understanding of evolution is not very deep. Darwin's theory of natural selection does indeed focus on evolution acting on individuals. But recent advances in sociobiology (largely made possible by Mendelian genetics which was unavailable to Darwin) present a model where evolution acts on genes, not individuals.  I suppose you can mean "individual genes or alleles" but I think most readers will take "individuals" to refer to individual people.Slrubenstein

Sentences like "Perhaps the most obvious way is accidental death" is also a little worrisome, although perhaps this is semantics. It's misleading to talk describe the drift-pertinent deaths as "accidental." All deaths are pertinent, and the moment and age at which any individual dies is in a sense an accident.
 * Again, you misunderstand the science. All death is pertinent to evolution, but not all death is pertinent to drift.  Slrubenstein


 * If a disproportionate number of individuals of a certain genotype die earlier than the average, fewer than the expected number of such individuals will be present at the census from which allele frequency is calculated. Lifespan also may be the cause for why an individual produces more or less offspring than the average for its genotype. So all lifespans that deviate from the average contribute to drift.  Seems intuitive to me. 168... 20:19 20 May 2003 (UTC)


 * Drift is one evolutionary mechanism, natural selection is another. As the article makes clear -- read it and learn! -- when death occurs one must be careful to distinguish between natural selection at work, or an accident which will result in genetic drift.

Regarding cooperative contributions, yes I see you incorporated lots of my text. Very generous of you. But one can see that I too have worked with your text, after you radically revised and reorganized what had been a far-along article that was largely my doing, until (for reasons evident in the dialogues above, involving jargon, news style, and not responding to the discussion about said) I posted a version tthat I believed conformed more to the desires and interests of the majority. You might not I characterized it as a more news-style starting point. I kept yours and others request that we call it a statistical effect up top. And in subsequent additions, which you pretend do not exist, I have suggested places for all your content. 168... 18:23 20 May 2003 (UTC)
 * Yet you still admit to deleting my material without having read it. As you say, "very generous." Slrubenstein

You misquote me. I said I didn't read the version that _immediately_ preceded the largely reverted article I posted. I did glance at it, and so I knew it represented an iteration of an article that was deeply flawed in its organization. You do sound awfully sure of yourself about the science, but it would be more convincing if you could actually quote or link to a source other than your own writing, as I have done in my posts. I may not have taken a genetics _course_, but I have cloned genes and calculated probability distributions and I wasn't a humanities major. If I ever want to ask a scientist something, I can ask myself. Who do you ask? 168... 20:03 20 May 2003 (UTC)


 * I am disappointed that you want to play the credentials game -- something we seldom do here. I have a PhD. in  anthropology, and was recently promoted to associate professor at my university.  Although I specialize in cultural anthropology I took a course on population genetics as an undergraduate and two courses on physical anthropology as a graduate student, and have taught basics of population genetics and evolution to undergraduates.  I also have a colleague who is a senior professor specializing on genetics who has published in refereed journals, and I have checked my contributions with him before submitting them here.  I also consulted a professor in the biological sciences department at my university.  In addition to those PhD. geneticists, I consulted the following sources (I will spare you the onerous task of doing research using sources other than other web-references, and will quote from some standard books and textbooks):


 * The random factor in evolution is called genetic drift and is due primarily to sampling phenomena (i.e. the size of the population). Since evolution occurs in populations, it is directly tied not only to the nature of the initial allele frequencies of the population, but to the size of the group as well.  If, in a population of 100 individuals, five type O individuals had been killed in an auto accident before they reproduced, they would not have made a genetic contribution to the next generation.  The frequency of the O allele would have been reduced in the next generation, and evolution would have occurred.  In this case, with only 100 individuals in the population, the change due to the accident would have altered the 0 frequency in anoticable way.  If, however, our initial population had been very large (10,000 people), then the evolutionary effect of removing a few individuals would be very samll indeed... When considering genetic drift, we must remember that the genetic makeup of individuals is in no way related to the chance happenings that affect their lives... in Introduction to Physical Anthropology, Robert Jurmain (PhD. Harvard 1975, Professor at San Jose State University), Harry Nelson (Emeritus Professor, Foothill COllege), Lynn Kilgore (Colorado State University), and Wenda Trevathan (New Mexico State University)


 * In the absense of mutation, selection, and migration, the frequencies of genetic variants in a population remain in accord with the Hardy-Weinberg principle, constant generation after generation. This is strictly true, however, only in ideal, infinitely large, populations.  In reality, no population is infinite and many are small...Consider sexual diploid populations of 500,000, 5000, and 50 individuals, respectively.  If the population sizes remain constant, the next generation will come from samples of 1,000,000, 10,000, and 100 gametes from the gene pools of the previous generation.  Suppose that two alleles, A and a, are in some generation equally frequent in these populations (p=q=0.5).  The sampling proces introduces a variance of pq/N and a standard deviation of the square root of pq/N, where N is the number of gametes sampled.  The frequencies of alleles A and a in the next generation will, therefore, be 0.5000 plus or minus 0.0005, 0.500 plus or minus 0.005, and 0.50 plus or minus 0.05, respectively, in the three populations.  This means that in about 95 of the 100 samples, the gene frequencies will be about 0.4990 and 0.5010 in the large, between 0.490 and 0.510 in the intermediate, and between 0.40 and 0.60 in the small population.  The variation in gene frequency may be considered negligible in the large, small in the intermediate, but appreciable in the small population  (Note that it isn't something like a disease that is "sampling" the population, it is the researcher taking a sample of the population to determine gene frequencies -- Slr) in Genetics of the Evolutionary Process by Theodosius Dobzhansky

It shows an incorrect reading of the above text to say that the sampling is being done by the researcher. Rather its the offspring generation that is sampling the parent generation's alleles. The researcher doesn't sample the population in the above scenario, but instead looks at every last individual (i.e. takes a census, not a sample). 168... 02:58 21 May 2003 (UTC)


 * Selection is the most important of the factors that induce evolutionary changes, by affecting the frequency of genes in populations. It is not, however, the only one, as was pointed out at an early stage in the history of evolutionary research.  Among the other possible factors is one, namely, chance, the significance of which is still greatly disputed.  Let me illustrate, on the basis of a single example, how important chance is.  The human male produces many billions of gametes during his lifetime, and the human female many hundreds during hers.  Still, one human couple can produce, at most, only about a score of children.  It is largely a matter of chance which among the countless gametes will form the few successful zygotes.  Since virtually all gametes differe from one another genetically, owing to the almost unlimited number of possible combinations of the parental genes, it is obvious that accident plays an important role in determining the genetic constitution of the F1 of a set of parents...Accidents of sampling occur universally in natural populations.  The evolutionary significance of such random fluctuations depend son the contribution to fitness of the genes involved and on the size of the effective population, its isolation, and its permanence.  from Populations, Species, and Evolution by Ernst Mayr.
 * Mayr is primarily arguing against those who claim that drift can trump natural selection; Mayr is careful to show how the two can work (to use 168's phrasing) in concert. But it is clear that Mayr makes a strong distinction between "selection" and drift/sampling phenomena.  By the way, perhaps this is where you got your idea about gametes -- but Mayr is only using gametes to provide a general example; the real force at work is independent assortment.  Nevertheless, i now understand your reference to gametes.  Still, I wouldn't quite say "Your eye color allele scenario (which isn't up to date on the genetics of eye-color by the way) has to do with deviations from the odds that alleles of the grandparents of individuals are equally represented in the sperm and eggs (gametes) that come together to make them."  I am not sure that it is correct to say "deviation from the odds, isn't it a deviation from the predicted distribution?  In any event, the main point about drift is that this matters in small populations, and not in large populations (go back to the Dobzhansky quote)


 * That should be "deviation from the average". Yes, sorry, my mistake. 168... 21:52 20 May 2003 (UTC)


 * Consider the transmission of genes from parent to offspring. Suppose that a woman is an AO heterozygote at the ABO locus.  If she bears four children, one would predict that she would transmit her O allele to two offspring and her A allele to two the other two children.  That is only a probability, and it is possible that she might transmit the A allele to all four children, solely on the basis of chance.  While exceptions of this sort are true for an individual parent, if we consider a large number of AO genotypes, these exceptions will cancel out themselves out and about half the total number of offspring will will receive the A and the toher half the O allele.  In other words, chance exceptions to expectations are less important in large than in small populations...Genetic drift because of population size will be more important in small groups, and not important at all in larger units... in Physical Anthropology by Francis Johnston (University of Pennsylvania)


 * I hope that these quotes are adequate to justify the material I keep including, which you keep deleting. I hope they also make clear to you the difference between "selection" and "sampling" (and the difference between death as a result of natural selection versus death by accident). Slrubenstein

Thanks for quoting the texts. They don't contradict anything I have been saying, however, nor do they provide the support for your point, which I suggested you quote from an independent source. Above I asked "on what basis do you assert that this gamete business is more influential in drift in species that do reproduce sexually? I'm open to believing that it might be, but you cite no source." Did you not understand the question? I meant "more influential than the sources of randomness I discussed" and the question is a quantitive one. I consider it unanswered, and you do tempt me to conclude that you have no answer. Meanwhile, I believe it was you who started "the credentials" game with such remarks as "But since you do not understand the science, please defer to scientists who do when it comes to the content" and "read it and learn!" in reference to your own text. A university appointment is certainly no guarantee that a person is skilled at explaining things in plain language that most people can understand and will want to read. Posing a geneticist with the question "is this right?" with regard to something you wrote doesn't address the issue of whether it belongs in a certain kind of article in a certain place. Not to mention, you seemed to be confused about how I believe drift to work, since none of my beliefs are contradicted by the texts you seem to think I'd benefit from reading. 168... 21:43 20 May 2003 (UTC)


 * I will let other people judge for themselves. I will also remind you that my revision did not cut most of what you put.  You deleted my discussion of Hardy-Weinberg, and then asked, "why would it be necessary to invoke the HWE to explain drift? And what about the HWE actually applies to the circumstances where drift occurs?" -- see the Donzhansky quote, above, which makes the relevance of HW quite clear.  But your version emphasized mortality, and confused accidental death with selective death.  Above you wrote "So all lifespans that deviate from the average contribute to drift" and that is simply wrong (see the quote from Nelson and Jurmain, above).  Your version also de-emphasized the importance of population size as fundamental to defining drift; again, see the Dobzhansky quote, above. Slrubenstein

It amazes me that anyone could infer from the Nelson & Jurmain quote that lifespan is irrelevant to allele frequency, as you assert,
 * No, I said it is not necessarily relevant to drift. This article is about drift, not evolution in general (i.e. changes in allele frequency)Slrubenstein


 * It was an awkward sentence and you misread it. Read it like this: "It amazes me that anyone could infer from the Nelson & Jurmain quote that lifespan is irrelevant to allele frequency unless it influences numbers of offspring, as you assert." It's largely my fault, but you misunderstand my point. 168... 18:05 21 May 2003 (UTC)

except to the extent that it influences numbers of offspring. Yes, let's let others be the judge. Meanwhile, it seems you have forgotten the context of the discussion and my challenge. I wrote: "what about the HWE actually applies to the circumstances where drift occurs? Perhaps only the principle that the selection of parent alleles in the sex cells that form the zygote is a randomizer." My point was that the HWE didn't belong high up in the article. It's not _general_ to drift, whereas the the chance factors I emphasized were.
 * HW is not a "chance factor," it is the baseline against which chance factors may be compared and is thus generalizable to drift, which is precisely why other discussions of drift by scientists begin with HW. Slrubenstein

"it can't be right to describe it as the essense or the be-all and end all of drift" I wrote. That is my point.168... 22:57 20 May 2003 (UTC)


 * As for the gamete business, which you questioned, again, the Mayr quote (introducing his discussion of genetic drift) makes it clear how it is relevant to drift. Let's be very clear about the way you misquote me,  You wrote above, "on what basis do you assert that this gamete business is more influential in drift in species that do reproduce sexually?" but you were refering to this line in the article "But there is another major cause of drift in populations of sexual organisms, which can be understood only by referring to the Hardy-Weinberg principle of genetic equilibrium as a baseline."  Where do you see the word "more?"  Don't you see the words "another major?"  Are you aware that "more" and "another major" are not synonymous?  SLR


 * SLR, you said HWE was "primary." You wrote "Yes, in asexual organisms differential fecundity and morbidity [sic]...are the primary causes of drift. But in regard to sexual organisms, the primary cause of drift is something else" (i.e. HWE)" 168... 01:50 22 May 2003 (UTC)


 * You ask me to justify my making this statement, appealing to the HW principle, and I provide the wbove quotes and you say that you do not see how they support my point. I just am at a loss right now, since the quote from Dobzhansky appeals to HW by name, and the quotes from Nelson and Jurmain, Mayr, and Johnston rely on the HW principle implicitly.  No, I am not claiming that the HW principle is "more" important, and I didn't write that in the text -- but I do claim that it is very important, and the above four quotes make that very clear.Slrubenstein

You are really painting a misleading picture when you say I "questioned" the gamete business and offer that Mayr extract as a reply. See "That is my point" above.168...


 * They also make clear that drift is primarily a sampling phenomena, and has to do with the size of the population.Slrubenstein

Yes, they did, and how many highschoolers would have understood what they were talking about or bothered to read what they wrote? 168...
 * Hello! In that case, add content to explain "sampling phenomenon" more clearly!  Don't delete the truth because youthink some readers won't understand it; add to it in order to make it more understandable!Slrubenstein


 * And Hello to you too. You seem to forget that I did explain sampling phenomenon for you in your revamped version of the mostly-my version. First I corrected your misuse of the plural form, "phenomena", and then I wrote what was, if I may say so, a nice explanation of the sense in which drift is a sampling phenomenon, which explained at the same time what a sampling phenomenon is. But you did not then understand how the word "sampling" is used in probability and statistics (you provided further evidence for this with your misreading of the textbook you quote above, as I point out in a remark beneath it) and you fussed over the word "select" and rewrote the paragraph to make it an even bigger imposition on the reader's patience (not to mention your paragraph was clunky and less than engaging). Shortly after that, I reverted, but I believe my explanation of sampling remains, for the sake of the otherwise hapless would-be readers of your words "sampling phenomenon", in a section I created for it lower down.168...


 * All of the quotes above come from sections in books on genetic drift, and so are clearly relevant to an encyclopedia on genetic drift; all of the above quotes pertain to material I included which you deleted. I agree with you that it realy doesn't matter what one's credentials are, which is why I never asked you yours.  You say that your understanding of drift is consistent with the above quotes and I am glad.  You know, I really do not care what your understanding of drift is, since this is not "your" article -- nor mine, it is common property.  What I am concerned with is making the article accurate and representative of current scholarship.Slrubenstein

But they pertain equally to the material with which I replaced the material of yours I deleted. You deleted content of mine and wrote your own, because you were under the mistaken belief that I was in error or because (equally mistakenly I believe) you thought you could express the point better. Perhaps I deleted some material that I did not replace in the restructuring I did, but I beleive that where I did so I created a place for you to pop them back in, appropriately tweaked for the new placement. 168... 23:09 20 May 2003 (UTC)


 * Before I made any changes to the article, this was the opening paragraph:
 * In biology, genetic drift refers to the changes in the gene pool and in the characteristics of a species, which arise from the luck of the draw. Luck impacts the commonness or rarity of an allele, because an individual who possesses it and the trait that it confers is not guaranteed either to survive or to produce a particular number of offspring. Instead survival and numbers of offspring depend on chance factors, such as weather and being in the right place at the right time. In other words, even when individuals face the same odds, they will differ in their success.
 * "luck of the draw" is not wrong, but it is vague, and I added material on HW and deviations from it to provide an important example (based on Mayr, Dobzhansky, and Johnston). The second sentence is only partially correct, because the issue -- again, as Mayr and Dobzhansky make clear -- is not primarily about individual luck.  Yes, this can be a factor, and in fact I added the example of the car accident to make this clear (so if you happen to have written this point, I was not rejecting it, I was embellishing it with an example).  I think the car accident is a better example than weather, since weather often is a factor in natural selection (maybe we could put the more specific "getting struck by lightning).  In any event, as my inclusion of the car example shows, I do not disagree per se with this discussion of drift.  But this discussion completely neglects the things Mayr, Dobzhansky, and Johnston write about -- things they consider elemental to a definition of drift.  All I have done is to include them.  But you kept deleting.  Now you suggest that what I wrote may be correct, but is not appropriate to the article on drift.  I just cannot see how you can say that given the quotations I provide.Slrubenstein

You don't need especial weather-fitness or weather-unfitness for your time of death to be determined by the weather and for your lifespan to deviate from your genotypic average.
 * if it is a matter of fitness, then it is not drift. Slrubenstein

You have no basis for repeatedly calling me wrong about the issue of mortality--especially when it figures in the one-word "drift" definitions of people who teach the subject. "Luck of the draw" cost 3 extra monosylabic words vs "luck" and it represented a bit of enrichment for readers intimate with probability and statistics. You indulged a similar temptation when you called drift a "sampling phenomena," but the cost of your indulgence would be the confusion of many readers or a long digression about what sampling is. 168... 23:21 20 May 2003 (UTC)

All that you need for drift is individual luck. Alleles don't die, people do.
 * uh, when a person dies, an allele dies too. But the point is, you do not need "individual luck" for drift to occur.  Review Dobzhansky's discussion of sampling error -- that does not require individual luck.  Review the example from Johnston -- that does not require individual luck. Slrubenstein


 * I think it's fine to use "luck" to describe why you as an individual have your maternal grandmother's allele for eye-color gene #1 and not your maternal grandfather's allele for eye-color gene #1. I think readers are more comfortable to read discussion that is oriented around individuals than around abstract principles of statistics. It is quite possible to construct such a discussion without committing scientific inaccuracies. 168... 18:12 21 May 2003 (UTC)

Alleles may replicate, but they don't reproduce. You don't have biology without individuals. My sentence was not incorrect but slightly misleading to the extent that one could read it as "all drift consists only of the following", and to the extent one says the gamete stuff isn't covered by the word "reproduction." I don't think it was problem, but there is definitely something to your point, and so I have added the word "birth" to survival and reproduction in my latest post. 168... 23:32 20 May 2003 (UTC)


 * In all of my revisions I have strived to include material you wrote when it was correct and relevant. When I have deleted material -- or more importantly, re-inserted material you deleted, it had nothing to do with whatever credentials you may or may not have.  It had to do with your being wrong. Slrubenstein


 * It had to do with you believing I was wrong, but over time it seems progressively fewer of these things have persisted in seeming wrong to you, and you have reinstated them.168...