I've prepared a bunch of exam questions for students in my molecular evolution course. I gave them out two weeks before the exam and I promised them that I would post some of these questions on my blog to see how you would answer them. I'm hoping that you, dear readers, will show my students that there really is some controversy.Here's the third question.
Read the following statements by Richard Dawkins from his latest book, The Greatest Show on Earth (2009, pages 332 and 333).My students have read the Spandrel's paper so they are aware of the arguments made by Gould & Lewontin. Some of you may not be as familiar with those arguments so let me remind you of what Gould & Lewontin said back in 1978-79.
When the neutral theory of molecular evolution was first proposed by, among others, the great Japanese geneticist Motoo Kimura, it was controversial. Some version of it is now widely accepted and, without going into the detailed evidence here, I am going to accept it in this book. Since I have a reputation as an arch-“adaptationist” (allegedly obsessed with natural selection as the major or even the only driving force of evolution) you can have some confidence that if even I support the neutral theory it is unlikely that many other biologists will oppose it!Dawkins doubts that any mutation giving rise to a visible phenotype can be neutral ("ultra-Darwinists like me incline against the idea"). Such mutations are only important in molecular evolution. Do you agree with him? Does Neutral Theory only apply to invisible mutations that can only be detected by molecular geneticists? Be sure to bring up the enormous variations in phenotypic characteristics among different human populations.
... When a gene mutates into one of its synonyms, you might as well not bother to call it a mutation at all. Indeed, it isn’t a mutation, as far as the consequences on the body are concerned. And for the same reason it isn’t a mutation at all as far as natural selection is concerned. But it is a mutation as far as molecular geneticists are concerned, for they can see it using their methods.
At this point, some evolutionists will protest that we are caricaturing their view of adaptation. After all, do they not admit genetic drift, allometry, and a variety of reasons for non-adaptive evolution? They do, to be sure, but we make a different point. In natural history, all possible things happen sometimes; you generally do not support your favored phenomenon by declaring rivals impossible in theory. Rather, you acknowledge the rival but circumscribe its domain of action so narrowly that it cannot have any importance in the affairs of nature. Then, you often congratulate yourself for being such an undogmatic and ecumenical chap. We maintain that alternatives to selection for best overall design have generally been relegated to unimportance by this mode of argument. Have we not all heard the catechism about genetic drift: it can only be important in populations so small that they are likely to become extinct before playing any sustained evolutionary role?To which I would add the following argument: "We've all heard about Neutral Theory but it only applies to inconsequential mutations detectable only by molecular geneticists."
Larry, how about giving some good old examples? It would be very forthcoming of you if you would share some phenotypic traits that have been shaped by neutral evolution. Specifically, not molecular ones, but major traits that have a clear effect on the phenotype (but not, of course, in terms of selection), such as eyes, lungs, blood-clotting cascade, limbs, wings, blood-types, eye-color, ear-lobes. Have any been shown to be neutral?
ReplyDeleteLarry, I just enjoyed reading you piece on genetic drift from 2006: as quoted in Random Genetic Drift, v1.3 ©2006 Laurence A. Moran. In that you quote Harrison et al., 1988:
ReplyDeleteThe key question at stake is whether the immense genetic variety which is observable in populations of all species is inconsequential to survival and reproduction (ie. is neutral), in which case drift will be the main determinant, or whether most gene substitutions do affect fitness, in which case natural selection is the main driving force.
Is that really the key question? Yes, if that is what you're interested in explaining is. And then genetic drift does indeed plausibly explain lots of it (the majority, maybe). However, if you consider "the key question" to be how complex traits evolve, then genetic drift does not seem to be a very good explanation.
One of the major controversies in evolutionary biology is the neutralist-selectionist debate over the importance of neutral mutations. [Ibid.]
Importance for what? Explaining how eyes, evolved? I don't think so. You get my drift...
Bjørn Østman asks,
ReplyDeleteLarry, how about giving some good old examples? ... Have any been shown to be neutral?
It's shocking to me that you would have to ask such a question in 2011. As a graduate student interested in evolution you should be very familiar with the common examples and the debate.
Lewontin likes rhinoceros, I like tongue rolling. Some people can roll their tongues and some can't? Is this an adaptive trait?
I admit that these particular examples aren't going to change the course of evolution but it's all that's required to establish the real possibility that alleles with visible phenotypes can be neutral.
Once we've agreed on that point we can discuss whether the evolution of blood types in different human populations is due to selection or drift. (The answer is drift.)
Your other examples are more complicated but many of them arise from a combination of all the important mechanisms of evolution, not just natural selection.
I anticipate that you are going to fall back on the old canard that none of these traits have been proven beyond all reasonable doubt to be the result of random genetic drift acting on nearly neutral alleles, therefore it's permissible to say they all must be adaptations. If you don't see the flaw in that logic then there's no point in continuing.
Visible Mutations and Evolution by Natural Selection
Calling All Adaptationists (Again)
Darwin and Design by Michael Ruse
Bjørn Østman says,
ReplyDeleteImportance for what? Explaining how eyes, evolved? I don't think so. You get my drift...
This is old hat. If the only thing that interests you is adaptation then it's quite obvious that adaptationism is the only explanation for the things you are interested in.
Just don't confuse your interest in adaptation with an interest in evolution. And don't dismiss the rest of us who might not share your tunnel-vision "fixation" on adaptation.
And please make sure that the things that interest you are really adaptations and not accidents that are tricking you into believing that they are adaptations. That's a tough thing for adapationists to deal with because they don't like to question whether their favorite characters really are adaptive.
They don't have the tools to ask/answer questions like that.
It's shocking to me that you would have to ask such a question in 2011. As a graduate student interested in evolution you should be very familiar with the common examples and the debate.
ReplyDeleteI am familiar, but am making a point with the question. Also, I am no longer a graduate student.
I admit that these particular examples aren't going to change the course of evolution but it's all that's required to establish the real possibility that alleles with visible phenotypes can be neutral.
And that point I wanted to make is that while some traits are most probably neutral (I purposefully chose traits some of which are probably neutral), but that arguably a very large fraction of traits are adaptive, and that I think you underplay this to make your point. I work at a place with a very large group of evolutionary biologists (BEACON), and I am pretty sure that not one of them does not understand drift and accept that many traits are neutral.
I anticipate that you are going to fall back on the old canard that none of these traits have been proven beyond all reasonable doubt to be the result of random genetic drift acting on nearly neutral alleles, therefore it's permissible to say they all must be adaptations. If you don't see the flaw in that logic then there's no point in continuing.
You anticipate wrong, and I find it condescending. You have shoehorned me faster than the speed of thought. EOF.
My understanding is that the crux of the debate is quantitative - just how important is neutral evolution in comparison to adaptive? It's not that Dawkins is not aware of tongue rolling in humans, he just doubts there are many traits like this. Personally I am leaning toward very important, but I think the experimental answer to this question is not yet available. Without solid experimental foundation, the debate is essentially theological.
ReplyDeleteLarry Moran said:
ReplyDelete"It's shocking to me that you would have to ask such a question in 2011. As a graduate student interested in evolution you should be very familiar with the common examples and the debate.
Lewontin likes rhinoceros, I like tongue rolling. Some people can roll their tongues and some can't? Is this an adaptive trait?"
"Shocking"? Isn't that rather a patronising and arrogant response to Mr Østman?
Lewontin's guess that the difference between two-horned and one-horned rhinoceroses is non-adaptive is just that: a pure guess. Here is what Lewontin wrote:
"For example, the Indian rhinoceros has one horn and the African rhinoceros has two. Horns are an adaptation for protection against predators, but it is not true that one horn is specifically adaptive under Indian conditions as opposed to two horns on the African plains." (cited, together with my reply, in The Extended Phenotype, the chapter called 'Constraints on Perfection':
Actually, it seems unlikely that horns are an adaptation against predators. Lewontin is being too adaptationist here. Horns are more likely an adaptation against other rhinos, both as weapons and as displays. In which case it could very well be the case that one horn is best for Indian rhinos and two horns best for African rhinos (because a rhino has to match whatever is the norm for its competitors in its own region).
Never mind Lewontin's ignorance about the biology of rhinoceroses. The point is that Moran has no more right to assume, with Lewontin, that the number of rhinoceros horns is neutral than anybody else has the right to assume that they are under selection. Even if I could not invoke sexual selection (which makes it positively likely that Lewontin is wrong) Moran would have no right to assume that this is a good example of a neutral trait, still less would he have the right to patronise Bjørn Østman in that way.
Tongue-rolling is a polymorphism (I can 't do it, you very possibly can) which makes it more complicated. In this case we are not seeking an adaptive advantage in tongue-rolling per se. We are seeking some kind of balance between the two sides of the polymorphism. I don't know enough about it (nor does Moran) but I would keep an open mind, and the first place I would look would be for some kind of balancing selection (or perhaps heterozygous advantage) with tongue-rolling as a pleiotropic byproduct. Once again, the last thing I would do is simply assume in my ignorance (and Moran's ignorance) that selection is not involved.
Richard Dawkins
I'm kind of amazed at fairly adept scientists running with the idea that the natural world works with mainly this or that, but sometimes other things. It seems to me that when you see something like this, you're not looking at all the evidence you need to make reasonable conclusions.
ReplyDeleteAn example of this is how so many thought that neanderthals did not mate with cro magnon. Common sense and a visit to a dog park will tell you that if they lived in the same environment, there was cross mating.
If the information at one level does not make sense but is less nonsensical at a lower level, go a couple of more levels down to see how that makes sense. While we still do not understand what all that 'junk dna' is for, any change should be considered important enough to study. To understand how adaptation works, you need to understand the underpinnings of it, and the underpinnings of that, and of that and so on as far as you can go. Show that these lower/smaller changes actually do NOT have an affect and then you're good. Dismissing it without knowing is ... sloppy? It is not impossible that string theory has something to do with mutation. It is unfortunately not possible to say it does not if you ignore or dismiss things which may well have an effect that you had not considered.
I don't mean to push infinite regress, but it is reasonable to investigate any changes that you can see at some level unless you wish to limit the impact of your work to some subset of the possible.
Can't genetic drift be one of the sources available (maybe the main source) so that natural selection has more to work with if the environment changes? Alleles can sit there benignly until their importance (positive or negative) changes. Tongue rolling may not be important now but if all vegetables change into access only by tubes then it gets promoted ( just a silly example). Isn't this just dealing with the details and semantics?
ReplyDeleteRookie question: wouldn't "circle species" be examples of neutral evolution?
ReplyDeleteRather than polarizing the discussion I'll take the middle ground:
ReplyDeleteA mutation that changes phenotype can evolve 'as if it were neutral' even though its effect on fitness is non-zero. Said another way, non-selective effects on allele frequency can be more important than small selective effects.
So the variation in ability to roll one's tongue could exist because of non-selective effects on the alleles responsible, even if this ability has a non-zero effect on fitness.
such as eyes, lungs, blood-clotting cascade, limbs...
ReplyDeleteWhat about the initial duplications of HOX genes that created extra limbs wayyy back in the Cambrian, eventually allowing those extra limbs to turn into, e.g., wings, mouth parts, etc.? The initial duplication of limbs might well have been neutral or even negative (think of the effect on the ability to propagate of polydactyly in humans, for example).
@NewEnglandBob
ReplyDeleteCan't genetic drift be one of the sources available (maybe the main source) so that natural selection has more to work with if the environment changes?
No, it can't. Genetic drift neither introduces not maintains genetic variation in a population. In fact, much like natural selection, genetic drift eliminates variation within a population. In other words, if tongue rolling were neutral, it would still either become fixed or go extinct in a given population.
Bjørn Østman says,
ReplyDeleteYou anticipate wrong, and I find it condescending. You have shoehorned me faster than the speed of thought. EOF.
Point taken. What, exactly, is your position? Do you agree with me that there are many examples of neutral alleles that express a visible phenotype?
Why did you ask for examples? And why did you say the following?
Is that really the key question? Yes, if that is what you're interested in explaining is. And then genetic drift does indeed plausibly explain lots of it (the majority, maybe). However, if you consider "the key question" to be how complex traits evolve, then genetic drift does not seem to be a very good explanation.
Are you suggesting that natural selection is the only mechanism required to explain complex traits? Is that a correct interpretation of what you are saying?
Do you think that the organization of the human genome is complex?
DK says,
ReplyDeleteWithout solid experimental foundation, the debate is essentially theological.
Not quite. I say that many phenotypic traits seem to have no effect on fitness while some have an extremely important effect on fitness.
Others say that, to a first approximation, all phenotypic traits effect fitness.
I do not those as equivalent positions. One of them makes a very direct claim that is not supported by the evidence.
Richard Dawkins says,
ReplyDelete"Shocking"? Isn't that rather a patronising and arrogant response to Mr Østman?
Perhaps it's patronizing but I really do think it's shocking that thirty years after the Spandrels paper students are still unaware of this very important controversy.
Students are still being taught evolutionary biology as though natural selection were the only game in town. They don't know anything about random genetic drift or Neutral Theory.
Biology students at my university are graduating without ever being told that the vast majority of alleles that are fixed during evolution are fixed by random genetic drift.
Worse than that, they can't even define evolution correctly. They've been taught that "evolution" and "natural selection" are practically synonyms.
I don't care whether you are a pluralist or an adaptationist. If you are teaching evolution you have an obligation to teach both sides of this controversy. If teachers don't even know there's a controversy then that' a serious problem.
Richard Dawkins wrote,
ReplyDeleteLewontin's guess that the difference between two-horned and one-horned rhinoceroses is non-adaptive is just that: a pure guess.
It's an informed opinion that's every bit as valid, scientifically, as that of an adaptationist. Lewontin is not wedded to the idea that every phenoptically visible mutation is adaptive because he knows that there are perfectly reasonable alternatives.
He knows that there are well-supported examples of phenotypic traits that behave as if they were neutral.
I loved your adaptationist response in The Extended Phenpotype. That's a book that should be required reading for every biology students who wants to understand evolution.
The point is that Moran has no more right to assume, with Lewontin, that the number of rhinoceros horns is neutral than anybody else has the right to assume that they are under selection.
I do not ASSUME that the number of horns is non-adaptive. I raise the question. I challenge the assumption that everything like that is necessarily adaptive. I seek other explanations. The idea that the difference between Indian and African rhinos might just be due to accident seems quite reasonable to me so I agree with Lewontin that this presents a challenge adaptationist thinking.
Richard, do you agree that the presence or absence of an extra horn might not confer any selective advantage?
Point taken. What, exactly, is your position? Do you agree with me that there are many examples of neutral alleles that express a visible phenotype?
ReplyDeleteYes, I most definitely agree that they must be there. Tongue rolling seems plausible, as do earlobe shape.
Are you suggesting that natural selection is the only mechanism required to explain complex traits? Is that a correct interpretation of what you are saying?
Yes, that is in fact my position, with a qualification. I would be much surprised if we ever learned that traits that are functional and yet whose fixation is only ever affected by drift. I do not believe on theoretical grounds that drift can do (all) that. This answer does arguably depend on what one understands a complex trait to be, and I emphatically do not consider the human genome to be a trait (and in the case of some genome or string of four letters that has only ever been drifting, then that genome/string is not complex at all, but easily described as 'just random').
As for tongue rolling, tongular dexterity is likely adaptive, I would think, though a few tongular abilities more or less might not affect fitness, and could be drifting (or, as suggested above, under balancing selection due to pleiotropy).
And again, the "key question" is only "whether the immense genetic variety which is observable in populations of all species is inconsequential to survival and reproduction" if that's one's interest. If one is interested in explaining how complex traits evolve, then it seems really foolish not to look towards adaptive explanations from the get go (but also not to keep neutrality in mind, of course).
I do not ASSUME that the number of horns is non-adaptive. I raise the question. I challenge the assumption that everything like that is necessarily adaptive. I seek other explanations. The idea that the difference between Indian and African rhinos might just be due to accident seems quite reasonable to me so I agree with Lewontin that this presents a challenge adaptationist thinking.
ReplyDeleteI totally agree with you here, Larry. I also do not assume that this or that trait is adaptive, but raise the question (presently I am using dN/dS for one bacterial gene to assess that). However, I do not know anybody working in evolution who thinks that everything "like that" is adaptive. Who on Earth are these people who don't know and teach about genetic drift, for chrissakes? As for U of Toronto, hopefully not a single student graduates without learning about genetic drift. That would be intolerable.
Richard Dawkins says,
ReplyDeleteTongue-rolling is a polymorphism (I can 't do it, you very possibly can) which makes it more complicated. In this case we are not seeking an adaptive advantage in tongue-rolling per se. We are seeking some kind of balance between the two sides of the polymorphism. I don't know enough about it (nor does Moran) but I would keep an open mind, and the first place I would look would be for some kind of balancing selection (or perhaps heterozygous advantage) with tongue-rolling as a pleiotropic byproduct.
Like you, I keep an open mind. In the absence of any obvious selective advantage that I have over you (I can roll!) I'm prepared to entertain the idea that there is none.
That's why I react strongly when people try to tell me that (almost) every phenotypic trait has to be an adaptation.
Richard, do you believe that the ability to roll or not roll your tongue MIGHT have absolutely nothing to do with fitness? Do you agree that this is a perfectly reasonable hypothesis that's just as valid as the one assuming that it has to be an adaptation? Or, do you insist that the adaptationist assumption is a better way to go about answering the question?
Once again, the last thing I would do is simply assume in my ignorance (and Moran's ignorance) that selection is not involved.
We are both ignorant. Unlike you, I do not assume a priori that this particular phenotypic trait has to affect fitness. I'll wait for evidence that tongue rolling (or not rolling) is adaptive. Until then, my working hypotheses it that it's probably not adaptive.
At the very least, I'd like you stop implying that my position is any less reasonable than yours.
Rosie Redfield says,
ReplyDeleteRather than polarizing the discussion I'll take the middle ground:
What's the middle ground? Is it that some visible phenotypes may have no effect on fitness? That's what I'm saying so it can't be a middle ground.
Is it that all visible mutations affect fitness in some way, even if it's very subtle? That's not a middle ground either.
This may be one of those situations where you can't pretend to be neutral! :-)
The problem with asking whether or not limbs or wings are adaptive traits is that these aren't truly "traits" as I think of it. Wings and limbs are collections of traits. Do we know that every feature of the limb or wing was adaptive, or that these features evolved initially under selection?
ReplyDeleteMeh, all you guys ever focus on is the 1% of total biodiversity anyway. I'll follow mainstream evolutionary theory when:
ReplyDeletea) it properly involves subjects other than ecology and behaviour – molecular and cell biology being big ones (more than just giving them a passing mention); and
b) when you actually look at what most life is: prokaryotes and protists.
Animal evolution is largely irrelevant to my interests. Besides, we must first have a decent grasp of developmental genetics and cell biol prior to philosophising about large scale 'adaptations' or 'spandrels' or whatever. It's difficult to talk about something as ill-defined as tongue rolling. Developmental genetics is progressing quite well, but it has yet to reach a level where we can sensibly talk about complex phenotypes and behaviours, and it has yet to be understood by evolutionary biologists, too many of whom have a pathetically depressing understanding of how genes actually work. This is why molecular evolution is a good start – we actually know it well enough to theorise about its evolution; though still have lots of room for improvement.
For now, I'll stick to the works of people like Mike Lynch and Arlin Stoltzfus...
Larry, that's not what I said.
ReplyDeleteI said that visible phenotypes may affect fitness and still evolve neutrally. I think that qualifies as 'middle ground'.
Bjørn Østman says,
ReplyDeleteHowever, I do not know anybody working in evolution who thinks that everything "like that" is adaptive. Who on Earth are these people who don't know and teach about genetic drift, for chrissakes? As for U of Toronto, hopefully not a single student graduates without learning about genetic drift. That would be intolerable.
It's encouraging that you work in an environment where everyone understands evolution. Where is this wonderful place?
I hate to break it to you but students are graduating from universities all over the world with only a rudimentary—and mostly incorrect—understanding of evolution. It's not just at the University of Toronto.
Not only that, there are students getting Ph.D.s in biological sciences who don't know anything about population genetics, random genetic drift, or neutral theory. We have quite a few in my department.
Rosie Redfield says,
ReplyDeleteI said that visible phenotypes may affect fitness and still evolve neutrally. I think that qualifies as 'middle ground'.
I gotta hand it to you, that certainly looks like a middle ground.
It's called having your cake and eating it too.
Are you ready to write up this extension to Neutral Theory that includes alleles affecting fitness? :-)
(I assume you're not referring to nearly neutral alleles, right?)
Not quite. I say that many phenotypic traits seem to have no effect on fitness while some have an extremely important effect on fitness.
ReplyDeleteOthers say that, to a first approximation, all phenotypic traits effect fitness.
I do not [see] those as equivalent positions. One of them makes a very direct claim that is not supported by the evidence.
I do see them as equivalent. Strictly speaking, your "seem to have no effect on fitness" may only mean that we don't know or understand much (and that's not unreasonable expectation). The way I see it, until you and Richard can look at rhino's second horn and objectively decide if it is adaptive or not, the debate about its adaptive value will be strictly in the realm of articles of faith.
Personally, I find the idea that tongue rolling might be adaptive pretty laughable but I fully realize that I can't prove it one way or another.
It's encouraging that you work in an environment where everyone understands evolution. Where is this wonderful place?
ReplyDeleteClose by you, actually. Michigan State University, BEACON Center for the Study of Evolution in Action (beacon-center.org). Of course, I cannot guarantee that everyone here are well-versed in population genetics, but I'm willing to bet that no one here with a PhD is unfamiliar with genetic drift. By the way, I TAed population genetics with John Endler (of Natural Selection in the Wild) when I was at UCSB, so I'm okay. ;)
Not only that, there are students getting Ph.D.s in biological sciences who don't know anything about population genetics, random genetic drift, or neutral theory. We have quite a few in my department.
PhDs in nothing to do with evolution, I must presume, then. Right? Otherwise, WTF!?
Rosie Redfield:
ReplyDeleteI said that visible phenotypes may affect fitness and still evolve neutrally.
Actually, how does that work? With s<1/N, you mean?
Larry Moran wrote:
ReplyDeleteI do not ASSUME that the number of horns is non-adaptive.
Yes, Larry, that is exactly what you do. Bjørn Østman asked you to "share some phenotypic traits that have been shaped by neutral evolution. . . Have any been shown to be neutral?" That's "shown to be", Larry, as in "evidence". And instead of giving him evidence you (a) patronised him and (b) gave no evidence, but offered a pair of plausibility arguments.
The idea that the difference between Indian and African rhinos might just be due to accident seems quite reasonable to me so I agree with Lewontin that this presents a challenge adaptationist thinking.
Richard, do you agree that the presence or absence of an extra horn might not confer any selective advantage?
Of course I agree. The mythical 'adaptationist' that you have constructed doesn't exist. The hypothesis that horn number is selectively neutral is a null hypothesis waiting to be tested. That's what you do with null hypotheses. You test them: glue extra horns on Indian rhinos, Tinbergen-style, or something like that.
In the absence of experimental tests on rhinos, I met your plausibility argument head on. Lewontin forgot sexual selection. The moment you remember it, the adaptation hypothesis starts to look a lot more plausible.
The Bishop of Birmingham (admittedly with a different agenda from yours) wrote "As for camouflage, this is not always easily explicable on neo-Darwinian premises. If polar bears are dominant in the Arctic, then there would seem to have been no need for them to evolve a white-coloured form of camouflage."
My satirical translation (The Blind Watchmaker) of his 'Argument from Personal Incredulity' might just as well apply to you: "I personally, off the top of my head, sitting in my study, never having visited the Arctic, never having seen a polar bear in the wild . . . have so far not managed to think of a reason why polar bears might benefit from being white."
I went on to point out that predators need to be camouflaged from their prey. I could have mentioned possible selection pressures involving solar reflectance. But the fact that I can reel off counter plausibility arguments is not the point. The point is that we should keep an open mind until further research is done.
You, Larry, claim to have an open mind, but it is not much in evidence. You have invented a category of people called 'adaptationists' who insist that all observable phenotypes are naturally selected. All I would really insist is that it is worth testing the null hypothesis that they are not naturally selected.
It is you, Larry, who are prepared to insist that particular examples are neutral, when you have no evidence, nothing but a plausibility argument like the Bishop of Birmingham.
You said, on this thead:
. . . we can discuss whether the evolution of blood types in different human populations is due to selection or drift. (The answer is drift.)
"The answer is drift". Larry, you actually said, "The answer is drift." Not, "My personal belief (based on what I find plausible, and in spite of the published evidence for selective effects on blood group polymorphisms) is that the answer is drift." You came right out and said, "The answer is drift." How dare you? How do you KNOW? How would you prove your null hypothesis that no selection is involved? Are you aware that some of our human blood group polymorphisms are shared by chimpanzees, suggesting that they are extremely ancient? Doesn't that suggest to you that there just might be a frequency-dependent selection based on disease-resistance, a frequency-dependent selection strong enough to have persisted through 6 million years?
Richard
Wow, what a fascinating discussion, and to have Richard Dawkins actually involved in the debate really makes reading this a priviledge.
ReplyDeleteI am very much a layman trying to understand these issues and do have a question to ask.
Prof Moran paraphrases Prof Dawkins in the original post as follows:
"Dawkins doubts that any mutation giving rise to a visible phenotype can be neutral ("ultra-Darwinists like me incline against the idea")."
May I just seek some clarification. My understanding is that "nearly neutral theory" says that traits can be fixed not only when they are selectively advantageous, or even neutral, but even when they negatively affect survival.
This is when pure random chance simply outweighs the decrease in fitness the trait causes, and by fluke allows it to spread throughout the population.
Most people in this thread are sticking to purely neutral changes, but isn't it the case that Random Genetic Drift can fix negative traits, as long as they are not so negative to outweigh the steer RGD can give them?
So my understanding is that the genome consists of at least 3 parts - firstly truly junk DNA which has no phenotypical effect and so is invisible to selection which is propogated by RGD.
Secondly parts which encode phenotypically useful traits which are fixed by selection.
Thirdly parts which are neutral or even negative which have been fixed despite this by RGD.
Reading Prof Moran's posts on Lewontin I'll add a fourth area - parts which may be postive, neutral, or even negative, but which are fixed because their expression is liked to the expression of a phenotypically important area (is this a Spandrel?!).
Is this correct - at the layman level?
Prof Moran, seems to be quoting Prof Dawkins (as an ultra-Darwinist) as saying neutral traits don't exist phenotypically, maybe we are balancing on a razors edge and that is true, nothing is ever truly zero. But surely the point is even things that are negative can be fixed by RGD.
Isn't this debate about traits which are negative or only weakly selective being fixed not by selective pressure, but by RGD?
Nit picking about true neutrality isn't the issue. Lets instead concentrate on traits which decrease fitness, but become fixed.
Does Prof Dawkins really say they don't exist?
Prof Moran is your championing of these ideas solely down to theoretical ideas, or are there real examples of weakly negative traits being fixed - could you give them?
Richard Dawkins wrote:
ReplyDeleteAll I would really insist is that it is worth testing the null hypothesis that they are not naturally selected.
Dr. Dawkins, you say that neutrality is your null hypothesis. That's a bit surprising because your writings don't leave this impression on me. So let me ask you a simple question:
If you had to bet, what would you say is the proportion of alleles that evolve neutrally?
Thanks.
RD: Are you aware that some of our human blood group polymorphisms are shared by chimpanzees, suggesting that they are extremely ancient? Doesn't that suggest to you that there just might be a frequency-dependent selection based on disease-resistance, a frequency-dependent selection strong enough to have persisted through 6 million years?
ReplyDeleteThis knife cuts both ways. Since the difference between selection and drift is the rate of fixation of alleles, the long-term persistence of multiple polymorphic alleles could also be used to argue the absence of selective pressure. One would have to know the degree to which the changing allele frequencies over time could be modeled by a random walk vs a relatively constant allelic distribution consistent with selective forces.
I'm way over my head here, but I just want to say, even if we assume that there is no benefit of one horn vs. two horn, not even sexual selection taking place, you could still argue that both traits could have been adaptive. For simplicity's sake, say a single gene has a no-horn variant, a one-horn variant, and a two-horn variant. If any-horns is advantageous over no-horns, but one-horn is neutral to two-horn; and if an ancestral population of no-horns is isolated into two sub-populations, in one of which appears the one-horn mutation and in the other of which appears the two-horn mutation... then both traits are adaptive rather than neutral, and there is no need to even invoke genetic drift, at least not beyond the trivial sense that natural selection can only operate on the mutations that happen to randomly become available.
ReplyDeleteI've blathered on at length about it here, though as a layperson I'm probably not making much sense :)
"Dr. Dawkins, you say that neutrality is your null hypothesis. That's a bit surprising because your writings don't leave this impression on me.
ReplyDeleteIt is not MY null hypothesis. That isn't how null hypotheses work. It is THE null hypothesis, that which an experiment seeks to disprove. Perhaps you are unfamiliar with the technical term 'null hypothesis'?
So let me ask you a simple question:
If you had to bet, what would you say is the proportion of alleles that evolve neutrally?
How I would bet should be of no interest to you. But if you force me to guess, I would guess well over 50%, depending on how you measure it. That's looking at change at the molecular level, which is very different from saying that well over 50% of phenotypically manifested mutations are selectively neutral. My gut feeling there (once again, nobody should care about my gut feelings, nor about Larry's or Lewontins' or anybody else's, but you did ask) is that the number is much smaller than 50%.
The reason I am sceptical about the neutrality of mutations that are phenotypically realised is twofold:
1. Most mutations have pleiotropic effects. Even if one of a mutation's effect is neutral, it is unlikely that all of them are neutral.
2. Theoretical calculations (e.g. by J B S Haldane) show that selection pressures can be exceedingly small -- even too small for us to measure under ordinary field conditions -- yet still have an evolutionary impact. Natural selection, in other words, can be a more sensitive guage than human measurement!
Richard
1. Most mutations have pleiotropic effects. Even if one of a mutation's effect is neutral, it is unlikely that all of them are neutral.
ReplyDeleteThis is just hand-waving. You are essentially saying that, even if the allele is neutral with respect to the trait under examination, it is not neutral with respect to some other unspecified trait, so the fate of the allele is dependent on natural selection and not genetic drift.
2. Theoretical calculations (e.g. by J B S Haldane) show that selection pressures can be exceedingly small -- even too small for us to measure under ordinary field conditions -- yet still have an evolutionary impact. Natural selection, in other words, can be a more sensitive guage than human measurement!
You have a citation for this?
Your summary seems to contradict most of population genetics.
Quoting Chinahand:
ReplyDeleteMay I just seek some clarification. My understanding is that "nearly neutral theory" says that traits can be fixed not only when they are selectively advantageous, or even neutral, but even when they negatively affect survival.
Yes, this is true. Approximately, if the selection coefficient (the effect of a mutation, s=w'/w-1, where w' is fitness after the mutation, and w is fitness before) is less than 1/N (N is the population size), then the effect of the mutation cannot be 'detected' by selection, and the mutation is effectively neutral. Thus, the smaller the population is, the larger the selection coefficient can be without selection able to affect whether the mutation will go to fixation or not. The smaller the population is, the more genetic drift controls what happens to alternative alleles. Slightly deleterious mutations may go to fixation in this way.
James, it is true that it may be as you imagine it, that there may indeed never have been any drift involved in rhino horn evolution, and that both are adaptive. One can still ask what would happen to individuals born with two horns in the one-horn population, and vice versa, and it may of course be that it makes no difference, and so the new trait will drift. Yet, for the reason you have laid out, I don't really think this trait is a good example for this discussion. I personally prefer ear-lobe types.
It is not MY null hypothesis. That isn't how null hypotheses work. It is THE null hypothesis, that which an experiment seeks to disprove. Perhaps you are unfamiliar with the technical term 'null hypothesis'?
ReplyDeletePerhaps I am? Perhaps you know that null hypotheses are not something god-given and that they are *formulated* by humans? As in, based on their available knowledge? As in, perhaps you could imagine that maybe neutrality wasn't Darwin's null?
How I would bet should be of no interest to you.
Since we are on a shaky ground where experimental evidence is minimal or nonexistent, an exchange of gut feelings can be of interest.
I am sceptical about the neutrality of mutations that are phenotypically realised
What exactly is "phenotypically realized"? Such a term relies entirely on the senses used to evaluate phenotype. TGT --> GGT (Cys --> Ser, a mutation that is so near neutral in some contexts that it is almost impossible to find any biological consequences of it) is no less observable today than, say, a hyperdactyly. Hence, by definition, it is a phenotype, isn't it?
@Richard Dawkins,
ReplyDeleteRichard, I'm not sure I understand the logic of your argument. My position is that adaptationists should not assume that all phenotypic variation is subject to natural selection. I maintain that such an assumption is unwarranted in light of what we know about evolution. Adaptationist reasoning seems to reject, a priori, any non-adapatationist hypotheses.
Your response seems to be that pluralists are just as bad because they assume, without evidence, that some traits are non-adaptive. If that were true then surely that's not a defense of adaptationism but rather a condemnation of both kinds of dogmatic thinking. Is that the point you are making—that both sides are equally wrong?
If so, what is the appropriate way to think about evolution? When confronted with a characteristic should we consider all possibilities—adaptive and non-adaptive—and reserve judgment until we have evidence? Or do you believe that 30-year-old statements like the following are still reasonable and consistent with the scientific evidence in 2011?
The adaptationism controversy is quite different. It is concerned with whether, given that we are dealing with a phenotypic effect big enough to see and ask questions about, we should assume that it is the product of natural selection. The biochemist's "neutral mutations" are more than neutral. As far as those of us who look at gross morphology, physiology and behavior are concerned, they are not mutations at all. It was in this spirit that Maynard Smith (1976b) wrote: "I interpret 'rate of evolution' as a rate of adaptive change. In this sense, the substitution of a neutral allele would not constitute evolution..." If a whole-organism biologist sees a genetically determined differences among phenotypes, he already knows he cannot be dealing with neutrality in the sense of the modern controversy among biochemical geneticist.
This, as you know, is from The Extended Phenotype (p. 32). Does it still represent how you view evolution, phenotypes, and neutral mutations?
Richard Dawkins says,
ReplyDeleteYou, Larry, claim to have an open mind, but it is not much in evidence. You have invented a category of people called 'adaptationists' who insist that all observable phenotypes are naturally selected.
I don't think I've invented such a strawman. There may actually be such people.
However, the most common kind of adaptationist is the kind who are very skeptical about the existence of non-adaptive phenotypic variation. They don't necessarily INSIST that it's all under selective pressure.
All I would really insist is that it is worth testing the null hypothesis that they are not naturally selected.
Then we agree. We'll consider all hypotheses, both adaptive and non-adaptive, until we have evidence of adaptation.
Glad we cleared that up.
It is you, Larry, who are prepared to insist that particular examples are neutral, when you have no evidence, nothing but a plausibility argument like the Bishop of Birmingham.
Oh dear. That's pretty insulting.
I do, in fact, insist that there are some clear-cut cases of the non-adaptive evolution of alleles with visible phenotypes. The evidence strongly points in that direction so I accept it as a tentative fact.
On the other hand, there are some examples where the question is still open. There's no solid evidence for adaptation but it's still a possibility. We pluralists use some of these examples (e.g., rhinoceros horns) to stir up trouble among the adaptationists. We like to provoke them into defending the adapaptationist explanation in the absence of any evidence for adaptation.
It's fun, but don't make the mistake of assuming that Dick Lewontin is wed to the idea that the second rhinoceros horn is non-adaptive.
When writing the test for this course, are the students arguing for/against Larry or Richards view of Adaptationism?
ReplyDeleteIn reading, two completely different views of what it is, but I guess Larry is marking... :p
@Richard Dawkins,
ReplyDeleteEarlier in this thread I said, "... we can discuss whether the evolution of blood types in different human populations is due to selection or drift. (The answer is drift.)"
and Richard reacted,
The answer is drift". Larry, you actually said, "The answer is drift." Not, "My personal belief (based on what I find plausible, and in spite of the published evidence for selective effects on blood group polymorphisms) is that the answer is drift." You came right out and said, "The answer is drift." How dare you? How do you KNOW? How would you prove your null hypothesis that no selection is involved?
Let's take the ABO blood group as an example, although the same reasoning applies to all the other human blood groups. Different human populations have different frequencies of the various alleles at the ABO locus [OMIM]. There are about 50 different alleles segregating in our species. Most of them are rare.
There are several native populations in South America where the only alleles in the population are O-type alleles (typically a mixture of O(01) and O(02). Among native North Americans the frequencies of O alleles range from 0.70 to 0.98.
The best explanation seems to be a combination of founder effect and random genetic drift. It is estimated that only a few thousand people migrated to the Americas from Asia. The frequency of O-type alleles in the Siberian population is high and it's likely that the founding group, by chance, had an even higher frequency.
Balancing selection is clearly not a possiblity in populations that have no A or B alleles. Selection for the O-type allele is a possibility but as Halverson and Bolnick (2008) point out, there are several other genetic loci that show a similar enrichment in specific alleles and they can all be explained by the founder effect and random genetic drift.
There are lots of differences in allele frequencies among different human populations. You can see a Table on the Wikipedia site ABO blood groups. The frequency of B-type blood, for example, ranges from 7% in the Netherlands, to 20% in Poland, to 31% in India. The frequences of various A and O alleles also vary over a wide range.
Although we can't prove beyond a shadow of doubt that these frequency differences are due to random genetic drift it seems like the best explanation.
It would take a lot of special pleading to account for all that variation between populations by invoking natural selection.
Are you aware that some of our human blood group polymorphisms are shared by chimpanzees, suggesting that they are extremely ancient?
The data that I'm aware of suggests that the B and O alleles arose independently in the chimpanzee and hominid lineages. The data also shows that the ancestor for O(01) group of alleles is very ancient reaching almost as far back as the divergence between chimps and humans.
Doesn't that suggest to you that there just might be a frequency-dependent selection based on disease-resistance, a frequency-dependent selection strong enough to have persisted through 6 million years?
No. That data is perfectly consistent with neutral alleles segregating in a population by random genetic drift.
There may be some evidence of balancing selection in some human populations but that seems to play a minor role if the allele frequency distributions in various human populations.
@ Michael M
ReplyDelete2. Theoretical calculations (e.g. by J B S Haldane) show that selection pressures can be exceedingly small -- even too small for us to measure under ordinary field conditions -- yet still have an evolutionary impact. Natural selection, in other words, can be a more sensitive guage than human measurement!
You have a citation for this?
Your summary seems to contradict most of population genetics.
It does by no means contradict population genetics, but is indeed exactly what it says. Selection coefficients can be so small that we humans can't detect them (s=0.001 means a point one percent increase in fitness - very hard to measure in the wild - and yet selection can 'respond' to such a low s as long as the population size is larger than only 1000!). This is standard issue and can be found in any decent textbook on population genetics.
@ Larry Moran,
Adaptationist reasoning seems to reject, a priori, any non-adapatationist hypotheses.
I know this to be true in evolutionary psychology, but not anywhere in biology have I ever seen people refusing to admit that some trait under study could have evolved neutrally. It may not be said explicitly in many cases, but that is certainly not the same as not considering it the null hypothesis.
I do, in fact, insist that there are some clear-cut cases of the non-adaptive evolution of alleles with visible phenotypes. The evidence strongly points in that direction so I accept it as a tentative fact.
I would still like to know what those are, because I don't know them (but don't doubt they exist if you tell me).
Bjorne, you've been right on everything so far in this thread, (and very interesting, thanks), EXCEPT your assertion that evolutionary psychology researchers "reject, a priori, any non adaptationist hypothesis." With all due respect, that is nonsense. Evopsych researchers use the same null hypothesis as Prof. Dawkins cited: that a particular behavior is NOT influenced by any evolved psychological mechanism. If they didn't, they wouldn't be scientists. Just because their work focuses on generating and testing adaptationist hypotheses doesn't mean they believe all behaviors are the product of adaptations, in whole or in part. If you know some who do, they are not representative of the field. Every field has its share of better and worse scientists.
DeleteChinahand says,
ReplyDeleteMost people in this thread are sticking to purely neutral changes, but isn't it the case that Random Genetic Drift can fix negative traits, as long as they are not so negative to outweigh the steer RGD can give them?
You are correct. Not only that but most beneficial alleles are lost by random genetic drift long before they become fixed.
According to population geneticists, the probability of fixing a beneficial allele is about 2s, where "s" is the selective coefficient. That means that an allele conferring a 5% advantage (s=0.05) will only be fixed 10% of the time. In 90% of all cases it will be lost.
The stochastic nature of natural selection is rarely mentioned by adaptationists. They tend to think and say that as long as an allele confers any selective advantage—no matter how small—it will sweep to fixation.
Your point illustrates the difference between accepting Neutral Theory and understanding the role of random genetic drift. They are not the same thing at all.
Understanding random genetic drift means understanding that beneficial and detrimental alleles are also affected by drift.
It looks to me like we have two different theoretical models vying for empirical data!
ReplyDeleteOn the one side we've claims that a selective advantage of s = 0.001 can be fixed in a population of over 1000, while on the other we've got only a 10% chance of a selective advantage of 0.05 being fixed.
Bjorn Ostman what is the probability of s = 0.001 being fixed in a population of 1000 - Prof Moran's theory would seem to indicate it is tiny, but maybe population size is what is creating the differences.
Prof Moran are your calculations based on finite populations?
As a layman I have no ability to know what is the cause of what seem to be very different theoretical claims.
In realistic conditions can very slight selective advantages really be "reliably" fixed?
And what about neutral or slightly deleterious ones?
Both sides seem to be saying the maths is on their side - the only explanation I can think for this is the assumptions the models have.
Can anyone shed any light on this.
I'm fascinated and really enjoying learning. Thank you to all involved!
@ Larry and Bjorn: Why wouldn't I mean nearly neutral alleles?
ReplyDeleteLarry, I'm envious of your students. I wonder how many professors manage to engage in an argument with the actual protagonists of the discussions they have in class.
ReplyDeleteToo bad you didn't criticise Lewontin, he might be posting here if you did!
Chinahand, there is no disagreement between what Larry and I say. Larry is talking about the probability that a mutant allele will go to fixation, and I am talking about whether selection can detect the effect of it at all.
ReplyDeleteYes, the probability that a mutation with s=0.001 will fix is tiny, but it does depend on the population size. If you're really interested, this is the time to pick up a textbook.
Richard Dawkins said "The reason I am sceptical about the neutrality of mutations that are phenotypically realised..."
ReplyDeleteOne example that comes to mind of neutral mutation (or drift) being possibly important as an evolutionary force that affects phenotype is the process of gene duplication and subsequent creation of gene families, either within a genome or as part of whole genome duplication event (eg polyploidy). I'm thinking of the time period between a gene being duplicated and when it adopts a new function- what was happening in the interim?
The change from a functioning gene, to a not-needed, 'extra' gene that is no longer under selection and begins to drift by way of random base pair mutations until a new (probably related to the original) useful function is taken on. For those genes that don't become pseudogenes but take on new functions,they are probably producing gene products that don't affect things much one way or the other throughout the transition, or are selected against if deleterious, but otherwise are not under selection until a useful new function brings the gene under positive or purifying selection again. At what point is the gene changing (evolving) in a neutral way and when does it come under selection again?
I'm not sure how many base pairs are needed on average to take on a new function (or lay the groundwork for one - and now I am curious to look into it). As this is a common process in plants, and has been especially important in the diversity of angiosperms, 'drift' in this case could be a significant, almost necessary contribution, sort of a way to get to the next adaptive peak, or gene function, which is then fine-tuned by positive selection.
Bjørn Østman said:
ReplyDeleteI know this to be true in evolutionary psychology, but not anywhere in biology have I ever seen people refusing to admit that some trait under study could have evolved neutrally. It may not be said explicitly in many cases, but that is certainly not the same as not considering it the null hypothesis.
Few biologists will deny that genetic drift could play a role when you confront them with it, but my personal experience is that this possibility is not seriously entertained otherwise. My experience is mainly with ecologists (some of which tend to have an aversion to genetics), but some PhD students in evolutionary biology have a similar blind spot.
It is just gut feelings of course, but I believe Larry is correct when he claims that there is a bias towards adaptationist thinking.
BTW: I love this comment thread. Fascinating
It's not true in evolutionary psychology. As with biology grad students, some evopsych grad students may have a blind spot, but to characterize the field itself as a priori rejecting ANY plausible causal mechanism is false, and if you think about it for a moment, you'll realize it's quite insulting as well.
DeleteDawkins: "When a gene mutates into one of its synonyms, you might as well not bother to call it a mutation at all. Indeed, it isn’t a mutation, as far as the consequences on the body are concerned. And for the same reason it isn’t a mutation at all as far as natural selection is concerned. But it is a mutation as far as molecular geneticists are concerned, for they can see it using their methods."
ReplyDeleteMoran: "Dawkins doubts that any mutation giving rise to a visible phenotype can be neutral [...]"
By using "synonym", Dawkins was talking about base-pair substituions that resulted in no change in the resulting protein, I figure. Switching from this to a discussion of changes in visible phenotypes suggests that this interpretation was accidentally lost or misplaced.
"Once we've agreed on that point we can discuss whether the evolution of blood types in different human populations is due to selection or drift. (The answer is drift.)"
ReplyDeleteWhoah! What? You do realise, I presume that blood types are under the influence of stabilising selection pressures - due to their role in immunity to disease? Blood type variation helps to prevent disease spread. This is not remotely a neutral trait. Yet how can your comment possibly be reconcilled with that?
Bjorn:
ReplyDeleteIt does by no means contradict population genetics, but is indeed exactly what it says. Selection coefficients can be so small that we humans can't detect them (s=0.001 means a point one percent increase in fitness - very hard to measure in the wild - and yet selection can 'respond' to such a low s as long as the population size is larger than only 1000!)
How often do actual populations exceed 1000? Maybe for Gene Simmons.
Quoting Corneel:
ReplyDeleteFew biologists will deny that genetic drift could play a role when you confront them with it, but my personal experience is that this possibility is not seriously entertained otherwise. My experience is mainly with ecologists (some of which tend to have an aversion to genetics), but some PhD students in evolutionary biology have a similar blind spot.
That is indeed sad that these people are oblivious to the effects of drift. However, ecologists who don't care about genetics... hmm. And PhD students. Do these PhD students graduate and become postdocs and professor in evolution and still don't understand drift? If so, where are they? (I am implying here that if one does understand drift, then it must be taken seriously.)
anonymous says,
ReplyDeleteIt does by no means contradict population genetics, but is indeed exactly what it says. Selection coefficients can be so small that we humans can't detect them (s=0.001 means a point one percent increase in fitness - very hard to measure in the wild - and yet selection can 'respond' to such a low s as long as the population size is larger than only 1000!)
What you say is, strictly speaking, correct but it is also misleading. When adapationists refer to this fact they imply that all you need to do is postulate a very tiny selective advantage and this will quickly lead to the allele becoming fixed in the population.
They conveniently forget to add a few things like ...
1. The probability of such an allele becoming fixed in real populations is extremely small. Most alleles with tiny selection coefficients will be lost by random genetic drift before they even come close to fixation.
2. Neutral alleles (s = 0.0) can also become fixed and it's extremely hard to distinguish between alleles with tiny selection coefficients and those with zero selection coefficients. To all intents and purposes they are equivalent as far as evolution is concerned.
3. Slightly deleterious alleles (s=-0.001) can be fixed in a population by random genetic drift. Adapaptionists tend not to construct hypotheses based on the fact that an allele is slightly deleterious even though it's not much different from one that has a tiny positive selection coefficient.
The situation is even worse among scientists who aren't knowledgeable about evolution. Most of them think that as long as an allele confers even a tiny advantage it will become fixed. This popular misconception goes a long way toward explaining the behavior of evolutionary psychologists.
Tim Tyler says,
ReplyDeleteWhoah! What? You do realise, I presume that blood types are under the influence of stabilising selection pressures - due to their role in immunity to disease? Blood type variation helps to prevent disease spread. This is not remotely a neutral trait. Yet how can your comment possibly be reconcilled with that?
Have you asked some of the natives of South American about this? Those populations have lost all A and B alleles and they only have various pseudogenes (O alleles) that are non-functional.
I'd love to hear your explanation of this based on stabilizing selection.
Waiting ....
This is just hand-waving. You are essentially saying that, even if the allele is neutral with respect to the trait under examination, it is not neutral with respect to some other unspecified trait, so the fate of the allele is dependent on natural selection and not genetic drift.
ReplyDeleteThat's a very uncharitable reading. Dawkins specifically used the word "unlikely," which is not the same thing as saying that it is definitely the case that it is not neutral with respect to some other trait, which appears to be how you've read it.
Tim Tyler says,
ReplyDeleteBy using "synonym", Dawkins was talking about base-pair substituions that resulted in no change in the resulting protein, I figure. Switching from this to a discussion of changes in visible phenotypes suggests that this interpretation was accidentally lost or misplaced.
Here's the full quotation from page 334.
Not all neutral mutations are quite so neutral as that. Sometimes the new gene translates into a different protein, but the 'active site' ... of the new protein remains the same as the old one. Consequently, there is literally no effect on the embryonic development of the body. The unmutated and the mutated form of the gene are still synomyms as far as the effects on their bodies are concerned. It is also possible (although 'ultra-Darwinists' like me incline against the idea) that some mutations really do change the body, but in such a way as to have no effect on survival one way of the other.
Bjørn Østman said...
ReplyDeleteDo these PhD students graduate and become postdocs and professor in evolution and still don't understand drift?
I am not aware of any, so I guess not. Like the measles, most people get over it :-)
That's a very uncharitable reading. Dawkins specifically used the word "unlikely," which is not the same thing as saying that it is definitely the case that it is not neutral with respect to some other trait, which appears to be how you've read it.
ReplyDeleteExcept the argument is essentially a (blank)-of-the-gaps argument. Dawkins is proposing that, since we don't know all of the possible pleiotropic effects of a mutation, it is more likely that there is some hidden selective advantage to the mutation and that the population only appears to evolve neutrally with respect to the mutation. Dawkins doesn't tolerate God-of-the-gaps arguments from theists, so why should we tolerate an adaptationism-of-the-gaps argument from him?
Anonymous said, How often do actual populations exceed 1000? Maybe for Gene Simmons.
ReplyDeleteThe (effective) population size if often greater than 1000. It is not a number of how many offspring one individual has (like Gene Simmons), but how many individuals that makes up the population. The human population size is about 7 billion, while the effective human population size has been estimated to be in the order of 10,000 (ref 1, ref 2).
Larry said, What you say is, strictly speaking, correct but it is also misleading. When adapationists refer to this fact they imply that all you need to do is postulate a very tiny selective advantage and this will quickly lead to the allele becoming fixed in the population.
So it is (correct but) misleading because these adaptationists, who I still don't know who are except for some biochemistry PhD students at Toronto, misconstrue what it means?
1) Yes, that's true,
2) that was my point,
3) true again, and also what I said above:
Yes, this is true. Approximately, if the selection coefficient (the effect of a mutation, s=w'/w-1, where w' is fitness after the mutation, and w is fitness before) is less than 1/N (N is the population size), then the effect of the mutation cannot be 'detected' by selection, and the mutation is effectively neutral. Thus, the smaller the population is, the larger the selection coefficient can be without selection able to affect whether the mutation will go to fixation or not. The smaller the population is, the more genetic drift controls what happens to alternative alleles. Slightly deleterious mutations may go to fixation in this way. [Emphasis added.]
Larry said
ReplyDeleteHave you asked some of the natives of South American about this? Those populations have lost all A and B alleles and they only have various pseudogenes (O alleles) that are non-functional. I'd love to hear your explanation of this based on stabilizing selection. Waiting ....
That functional alleles were lost in some populations does not imply that there isn't stabilizing selection in other populations. Especially if these O-population went through a bottleneck perhaps during their descent into South America. (Apart from that, I don't know anything about blood types conferring immunity.)
That functional alleles were lost in some populations does not imply that there isn't stabilizing selection in other populations.
ReplyDeleteThat's not correct. The absence of stabilizing selection in some populations does carry implications for the general hypothesis.
What you mean to say, I think, is that if you are wedded to the stabilizing selection scenario then you can always make up stories that rationalize some of the evidence against you.
But that's not really the point. I claimed that genetic drift was the best explanation for the difference in allele frequencies in different populations. How does stabilizing selection explain the different frequencies of the B alleles in various populations? How does it explain that different populations have different ratios of the A1 and A2 alleles?
You can't just say "stabliizing selection" and leave it at that.
What you mean to say, I think, is that if you are wedded to the stabilizing selection scenario then you can always make up stories that rationalize some of the evidence against you.
ReplyDeleteI wouldn't ever say that, no. I am not "wedded" either way, but pointing out that stabilizing selection can explain the data. You seem to take this whole thing extremely personally (e.g., "you are wedded" and "against you"). Let me emphasize that in this particular case, I can see ways to explain the data that involve selection or no selection (i.e., drift), and that's all I know. You seem to be the one wedded to one particular hypothesis in this case.
But that's not really the point. I claimed that genetic drift was the best explanation for the difference in allele frequencies in different populations. How does stabilizing selection explain the different frequencies of the B alleles in various populations? How does it explain that different populations have different ratios of the A1 and A2 alleles? You can't just say "stabliizing selection" and leave it at that.
I say that if blood-types confer immunity differently, then balancing selection can explain the polymorphism within a (sub)population. That different populations have different ratios does not imply that it can't be balancing selection. Different populations may have experienced slightly different selection pressures from different pathogens (or whatever it is that different blood-types may have to do with). It is curious, after all, that A or B or O hasn't gone to fixation in all populations. That would have to be explained, too, in the absence of selection.
I am still not sure why it's neutral drift and not natural selection which is the null hypothesis.
ReplyDeleteBjørn Østman says,
ReplyDeleteI wouldn't ever say that, no. I am not "wedded" either way, but pointing out that stabilizing selection can explain the data.
And I pointed out that stabilizing selection can NOT explain the data.
You seem to take this whole thing extremely personally (e.g., "you are wedded" and "against you").
I'm sorry. I did not direct this at you personally. I've been reading your comments and I'm aware of our position. Next time I'll be careful to say, "If ONE is wedded to the stabilizing selection scenario ..." or "If an adaptationist is wedded ..."
The "you" was not meant to refer to you personally. Sorry if you took it so personally. :-)
No problem, of course. And thanks for the whole discussion. I'll make sure the post gets in the next Carnival of Evolution.
ReplyDeleteMany populations are structured into fairly small,local breeding pools, or demes. That suggests genetic drift and selection may both play important, concurrent roles.
ReplyDeleteThat's a tough thing for adapationists to deal with because they don't like to question whether their favorite characters really are adaptive.
ReplyDeleteThey don't have the tools to ask/answer questions like that.
You're not only a snide, smug, half-ignorant brickwall on this subject, Moran, but also a liar.
I seek other explanations.
ReplyDeleteNo you don't. You never do.
You just endlessly wave around your dog-eared copy of Gould & Lewontin, make an argument from ignorance, and then make fun of those silly 'adaptationists'.
Rinse.
Repeat.
It's boring.
Wave function says,
ReplyDeleteI am still not sure why it's neutral drift and not natural selection which is the null hypothesis.
I'm preparing a separate posting on that question since it seems to come up so often.
Sven says,
ReplyDeleteYou just endlessly wave around your dog-eared copy of Gould & Lewontin, make an argument from ignorance, and then make fun of those silly 'adaptationists'.
Rinse.
Repeat.
It's boring.
Sorry you feel that way, Sven.
Just out of curiosity, are you equally upset with the adaptationists? Or do you think they have a much more interesting point of view?
Richard said (albeit in a book, and not in this forum):
ReplyDelete... When a gene mutates into one of its synonyms, you might as well not bother to call it a mutation at all. Indeed, it isn’t a mutation, as far as the consequences on the body are concerned. And for the same reason it isn’t a mutation at all as far as natural selection is concerned. But it is a mutation as far as molecular geneticists are concerned, for they can see it using their methods.
Kimura came up with this theory well before I was born, so that explains why the extensive regulatory elements that act on DNA and RNA aren’t included in there. In addition, we now know that “silent” mutations can affect translation rates, thus affecting phenotype. Therefore, I don’t see how you can argue that “...it isn’t a mutation at all as far as natural selection is concerned.”
Larry said:
Be sure to bring up the enormous variations in phenotypic characteristics among different human populations.
The implicit assumption in doing so as a way of advocating for neutral theory implies that human populations are not subject to environmental differences (read: differences in selection). See CCR5-Δ32 differences, or sickle cell anaemia. I am not attempting to argue that these examples of selection are the rule; rather, you can’t claim that alleles have not been selected for at some time point.
Which brings me to the reason I first sought to post: In the pluralist/adaptationist debate, has the moot been clearly worked out? Are pluralists arguing that drift is the predominant factor in fixing alleles, or in altering allelic frequencies, and if so, at which time points, what of synergistic effects where neither ? Sure, there may be polymorphic traits whose frequencies are fluctuating randomly, but what if they were selected for at some point in the past? Who gets the point for that one?
I am still not sure why it's neutral drift and not natural selection which is the null hypothesis.
ReplyDeleteBecause of the definition of selection. It's a correlation between phenotypic variation and fitness. The null hypothesis, again by definition, is the lack of such a correlation.
This comes in very handy rhetorically to 'Pluralists' because it essentially means that their claim can only ever be rejected for particular cases, and even then only probabilisticly. Therefore, it need never be abandoned as a general claim.
@Susan: I'm not sure about an on average value, but shifts in function after gene duplication events can be obtained through as little as a single amino acid substitution. Looking at experimentally validated datasets it looks like on average the number of substitutions tends to be quite small, usually only a handful are needed to change binding affinities.
ReplyDeleteOf course more ancient duplications tend to have accumulated larger numbers of substitutions and comparatively larger shifts in function.
@Richard Dawkins While you seem to be defending a pluralist paradigm why is it that if you accept that Neutrality is the appropriate Null Hypothesis that even in the absence of evidence you, and many others, will still always come up with adaptationist scenarios for those changes? While those scenarios may be plausible, it seems premature to use them in a way that to many readers comes across as the explanation, as opposed to an alternative hypothesis that hasn't been shown to be true?
ReplyDeleteI think it is that sort of description of Evolution that has Larry and so many others (like myself) a little irritated at times.
And as @PsiWavefunction pointed out, the majority of the time when we are talking about these phenotypic differences it is important to remember that we are looking at a small fraction of biological diversity on this planet, and a small fraction of evolutionarily important changes that have come up over the time life has existed on this planet.
The vast majority of evolutionary innovations have been molecular, biochemical, metabolic, and cell biologic.
The A and B blood types diverged at least 13 million years ago. The A, B and O blood types persisted in Baboons, Orangutans and Humans - who all lived in Africa for most of that time. Explaining this using drift is like trying to pretend that three drunkards staggered between two electric fences for 13 million years and no-one died.
ReplyDeleteThe continued persistence and relative uniformity of these blood groups among modern populations is much better explained by frequency-dependent selection, caused by diseases that are sensitive to blood type, than by "The answer is drift".
Tim Tyler says,
ReplyDeleteThe A and B blood types diverged at least 13 million years ago. The A, B and O blood types persisted in Baboons, Orangutans and Humans - who all lived in Africa for most of that time.
I haven't done a thorough literature seach (have you?) but I believe that the B and O variants have arisen independently in these lineages. The B allele converts the enzyme activity into making a slightly difference product and the O allele is an inactive pseudogene.
Explaining this using drift is like trying to pretend that three drunkards staggered between two electric fences for 13 million years and no-one died.
If the A gene is non-essential then it's not surprising that variants and null mutations arise frequently. If these alleles were neutral with respect to fitness they will persist in the population for a long time. That's what we see when we look at all other kinds of known neutral alleles.
I'm not saying that is the only possible explanation. I'm saying that the data is consistent with multiple explanations for the existence of multiple alleles.
For those populations where the O allele has become fixed it seems difficult to explain this by any selection model unless you assume that the creation of a pseudogene confers a selective advantage over the primitive A allele.
The continued persistence and relative uniformity of these blood groups among modern populations is much better explained by frequency-dependent selection, caused by diseases that are sensitive to blood type, than by "The answer is drift".
I disagree. I think the fact that there are populations where everyone is homozygous for the O allele and the huge differences in frequencies of the B allele in different populations can only be reasonably explained by random genetic drift.
One of the problems with balanced selection as an explanation is that it requires the demonstration of an association between the ABO alleles and some measure of fitness. That association seems to be far from proven.
Here's what they say on the NCBI database for The ABO Blood Groups System.
Disease association
None; The A, B alleles are dispensable and no apparent physiological disadvantages are noted for the null phenotype O. ABO incompatible transfusion may result in death of patients. HDN (hemolytic disease of the newborn) due to ABO incompatibility is rare. A large number of studies have examined associations between ABO blood groups and a variety of diseases or conditions, but so far consistent observations have been few. Those include the association of blood group O individuals with increased incidence of ulcers and gastric carcinoma and increased susceptibility to Helicobacter pylori. Levels of von Willebrand factor, a coagulation glycoprotein, seem to be consistently about 25% lower in blood group O individuals (Jenkins and O'Donnell, Transfusion 2006 46 1836).
Tim, perhaps you could explain to us how this balancing selection is supposed to work?
Sven says,
ReplyDeleteBecause of the definition of selection. It's a correlation between phenotypic variation and fitness. The null hypothesis, again by definition, is the lack of such a correlation.
This comes in very handy rhetorically to 'Pluralists' because it essentially means that their claim can only ever be rejected for particular cases, and even then only probabilisticly. Therefore, it need never be abandoned as a general claim.
It's amazing how few cases of natural selection have actually been proven. There's lots of speculation and lots of just-so stories but not much evidence. Most people don't realize this. They assume that there are tons of examples of natural selection that are well supported by evidence.
I'm not saying that natural selection doesn't happen and I'm not saying there aren't some obvious examples that we should accept even though they may not be proven. What I'm saying is that it's a bit hypocritical of the adaptationists to be demanding proof of random genetic drift when we suggest that it's a distinct possibility in certain cases.
I said, "Be sure to bring up the enormous variations in phenotypic characteristics among different human populations."
ReplyDeleteAdrian replied,
The implicit assumption in doing so as a way of advocating for neutral theory implies that human populations are not subject to environmental differences (read: differences in selection). See CCR5-Δ32 differences, or sickle cell anaemia. I am not attempting to argue that these examples of selection are the rule; rather, you can’t claim that alleles have not been selected for at some time point.
I was referring to something else. Look around you the next time you are in a crowd of people. They all look different and many of those differences obviously have a genetic basis—children resemble their parents.
The differences may not be great but they are distinct enough that you have no trouble recognizing people you know and people from different parts of the world. I'm talking about the shapes of chins, and noses, and ears, and the shape of one's head and eyes.
A huge amount of that variation is due to different alleles segregating within the population. Those alleles are responsible for phenotypic as well as genetic variation.
Now the question is, how much of that variation is under selection?
I think it strains credibility to imagine that most of it confers some fitness value. Most adaptationists disagree. They assume that the differences between me and my children mean that one of us has to be more fit than the other. Otherwise they'd be forced to admit that the differences are neutral.
I can't imagine what the adaptationists think of the Asians, Africans, Europeans, and Native Americans walking the streets of Toronto. They are phenotypically different therefore selection must be acting. I wonder which group is destined to survive and which ones will become extinct?
It's amazing how few cases of natural selection have actually been proven. There's lots of speculation and lots of just-so stories but not much evidence.
ReplyDeleteHow can you say this? Every protein coding sequence conserved at greater than the background mutation frequency is evidence of natural selection. Even if it is not clear which phenotype caused by the protein expressed from the conserved sequence is the dominant factor in selection, the selection is assuredly there; sequence conservation is prima facie evidence.
Larry Moran said:
ReplyDelete"I was referring to something else. Look around you the next time you are in a crowd of people. They all look different and many of those differences obviously have a genetic basis—children resemble their parents.
The differences may not be great but they are distinct enough that you have no trouble recognizing people you know and people from different parts of the world. I'm talking about the shapes of chins, and noses, and ears, and the shape of one's head and eyes.
A huge amount of that variation is due to different alleles segregating within the population. Those alleles are responsible for phenotypic as well as genetic variation.
Now the question is, how much of that variation is under selection?
I think it strains credibility to imagine that most of it confers some fitness value. Most adaptationists disagree. They assume that the differences between me and my children mean that one of us has to be more fit than the other. Otherwise they'd be forced to admit that the differences are neutral.
I can't imagine what the adaptationists think of the Asians, Africans, Europeans, and Native Americans walking the streets of Toronto. They are phenotypically different therefore selection must be acting. I wonder which group is destined to survive and which ones will become extinct?"
You're falling in a trap here, a trap you should be aware of since one of your recent posts was related to it.
Many of the traits you talk above may have very well been under selection, in our recent past.
But don't refer to our modern society, with its inherent lack of selection for who survives and who doesn't.
It's a fallacy to use modern human society as an example.
Even more, when talking about traits that differentiate between "races", sexual selection is inevitable.
I think it strains credibility to imagine that most of it confers some fitness value. Most adaptationists disagree. They assume that the differences between me and my children mean that one of us has to be more fit than the other. Otherwise they'd be forced to admit that the differences are neutral.
ReplyDeleteAgain, I'm going to have to ask who the hell these "adaptationists" are. As far as I can tell (and let that be the null hypothesis), they are strawmen invented by Gould and Lewontin. At least, if they were around in 1978, by now they have been purged from the pool of professional evolutionary biologists. (And I will not count biochemsitry students, or any other students for that matter.)
I can't imagine what the adaptationists think of the Asians, Africans, Europeans, and Native Americans walking the streets of Toronto. They are phenotypically different therefore selection must be acting. I wonder which group is destined to survive and which ones will become extinct?"
Honestly, here you sound like a clueless creationist. In fact, it is the other way around than what you argue. If all differences were neutral, then they should eventually disappear, with all alleles becoming lost or going to fixation. With selection, this is not the case. Selection can have that effect, but if polymorphisms persist for very long times (i.e., seemingly indefinitely), then genetic drift cannot explain that, while stabilizing selection can: frequency-dependent selection, sexual selection. Also, what samwho said about selection pressures changing over time, except there is no "inherent lack of selection" in out modern society; people with pale skin living at the equator are less protected against skin cancer, and people with dark skin living in Toronto don't make as much vitamin D. Yes, both effects can be mitigated by sunscreen and milk fortified with vitamin D, but that really is a different story.
Lastly, I still am unaware of traits or trait values that have actually been shown to be neutral (and again, not that I doubt they exist, of course).
samwho says,
ReplyDeleteYou're falling in a trap here, a trap you should be aware of since one of your recent posts was related to it.
Many of the traits you talk above may have very well been under selection, in our recent past.
But don't refer to our modern society, with its inherent lack of selection for who survives and who doesn't.
In other words, in 2011 we have plenty of examples of phenotypic differences that have are completely neutral. Is that what you mean to say?
Will the adapationists agree with you?
It's a fallacy to use modern human society as an example.
Why? And just what is your evidence that selection was more important 100,000 years ago?
Even more, when talking about traits that differentiate between "races", sexual selection is inevitable.
Ahhhhh yes, sexual selection rears it's ugly head. The last refuge of the adaptationists who see everything in terms of animals.
Let's talk about dandelions instead. There are several thousand variants divided into several dozen different species depending on who does the classification. Are all those different-looking dandelions subject to selection for fitness?
It's much better to talk about plants or bacteria in these debates 'cause it means the adaptationists can't fall back on their favorite cop-out to explain why some phenotypes don't seem to have obvious fitness differences.
I said, "It's amazing how few cases of natural selection have actually been proven. There's lots of speculation and lots of just-so stories but not much evidence."
ReplyDeleteand anonymous replied,
How can you say this? Every protein coding sequence conserved at greater than the background mutation frequency is evidence of natural selection. Even if it is not clear which phenotype caused by the protein expressed from the conserved sequence is the dominant factor in selection, the selection is assuredly there; sequence conservation is prima facie evidence.
Point taken. I was thinking of demonstrated cases of positive selection leading to the fixation of new alleles.
However, as you point out, negative selection is occurring all the time to eliminate new deleterious alleles that arise by mutation.
Point taken. I was thinking of demonstrated cases of positive selection leading to the fixation of new alleles.
ReplyDeleteHowever, as you point out, negative selection is occurring all the time to eliminate new deleterious alleles that arise by mutation.
Larry, are you familiar with measuring dN/dS? If so, are you satisfied with the interpretation that a ratio higher than 1 identifies positive selection? If so, are you familiar with cases demonstrating positive selection?
Mimicking Cato the Elder: What are some examples of traits or trait values that have been shown to be neutral?
(I'm deliberately not using 'just-so stories' and 'proven', as I think neither belongs anywhere in this discussion - try 'hypothesis' and 'shown'.)
If you toss a coin ten times and it comes up 6 heads and 4 tails, does that mean that the coin is biased to coming up heads?
ReplyDeleteLarry Moran said: "In other words, in 2011 we have plenty of examples of phenotypic differences that have are completely neutral. Is that what you mean to say?"
ReplyDeleteYes, there should be many instances of neutral characters, in our specie, living in a modern society (where differential survival is greatly reduced). But this tells nothing about how this traits "fixed" in the first place.
"Will the adapationists agree with you?"
:-?? I don't know them to ask them.
I said: "Many of the traits you talk above may have very well been under selection, in our recent past.
ReplyDeleteBut don't refer to our modern society, with its inherent lack of selection for who survives and who doesn't..."
Not just: "...It's a fallacy to use modern human society as an example.
Larry Moran answered: "Why? And just what is your evidence that selection was more important 100,000 years ago?"
I said -selection for who survives and who doesn't-.
Why did you cut that part and asked me about selection?
Is it because the difference is enough to answer your request?
Modern societies take out differential survival, and fitness if mostly the result of cultural factors. It's hardly the ideal example when looking for rules (to explain why should drift be the null hypotheses),
Bjørn Østman asks,
ReplyDeleteLarry, are you familiar with measuring dN/dS? If so, are you satisfied with the interpretation that a ratio higher than 1 identifies positive selection?
Not always. It depends on whether you are analyzing a single sequence or a multiple sequence alignment. I'm also skeptical about applying it to sequences within a population.
If so, are you familiar with cases demonstrating positive selection?
Yes, I'm aware of the fact that there are many papers claiming to have demonstrated positive selection based on the ratio dN/dS.
The same data shows that many sites within a gene are NOT under positive selection and fixation must be due to random genetic drift. If you accept that all those sites demonstrate drift then the ration of selected to nearly neutral sites is very small.
I wasn't thinking of sequences when I made my earlier statement. Everyone agrees that random genetic drift dominates at the level of nucleotide and amino acid substitution.
I was thinking of cases where there is a visible phenotype and the case for selection has been demonstrated by showing that a certain allele confers a selective advantage.
Application of the dN/dS ratio has been abused in the past—I'm sure you're aware of the papers. I assume that you are an expert so any examples of positive selection you come up with are likely to be real.
Re: "Disease association: None [...]" That information is simply wrong. For instance:
ReplyDelete"There is an inherited predisposition to infection, and individuals with blood type O are more often infected,[a][g] while blood types B and AB can confer partial protection against symptomatic infection." - http://en.wikipedia.org/wiki/Norovirus
Individuals' susceptibility to cholera is affected by their blood type, with those with type O blood being the most susceptible - http://en.wikipedia.org/wiki/Cholera
"Several neutrality tests clearly demonstrated deviations compatible with balancing selection, peaking at several locations along the gene. The time depth of the genealogy was also incompatible with neutral evolution, particularly in the region from exons 6 to 7, which codes for most of the catalytic domain." - http://www.ncbi.nlm.nih.gov/pubmed/18629539
ReplyDeleteBjørn Østman
ReplyDeleteMimicking Cato the Elder: What are some examples of traits or trait values that have been shown to be neutral?
Like Cato, I suspect you have an fixation. :-)
You seem to be unaware of any adaptationists and seem willing to challenge every example of random genetic drift.
You have undoubtedly discovered dozens of examples on your own—all those genes where the magical dN/dS ratio is less than one.
I don't have any more time to defend examples of drift. Why not just look up the standard examples in the textbooks and try and refute those?
Tim Tyler says,
ReplyDeleteIndividuals' susceptibility to cholera is affected by their blood type, with those with type O blood being the most susceptible -
I think you have to be a little more skeptical of those correlations, especially when they haven't been reproduced. The information on the NCBI website was contributed by Fumiichiro Yamamoto who is probably the world's leading expert in this field.
If type O blood is so bad then why do we see populations where the O allele has become fixed? How does that jibe with balanced selection?
More molecular fingerprints of selection:
ReplyDelete"Thus, both an accumulation of mutations as well as an assortment of the mutations by recombination seems to be responsible for the ABO gene diversity. The prevalence of replacement mutations indicates positive selection for allelic diversity."
- http://www.ncbi.nlm.nih.gov/pubmed/12829588
Point taken. I was thinking of demonstrated cases of positive selection leading to the fixation of new alleles.
ReplyDeleteWhat's old today was new yesterday. All those highly conserved sequences today exhibiting such strong negative selection against mutation were new once upon a time. It seems reasonable to posit positive selection at that time.
In any event, the entirety of cancer is one big exercise in positive selection leading to the fixation of new alleles.
The same data shows that many sites within a gene are NOT under positive selection and fixation must be due to random genetic drift. If you accept that all those sites demonstrate drift then the ration of selected to nearly neutral sites is very small.
ReplyDeleteNo, that's only assuming that fixation took place along the branch that is being tested, and that is of course not true. Many codons across different branches code for the same amino acid because the folding of the protein is important (I'm aware you know all this, of course), but synonymous changes are neutral. However, there was likely (and by that I mean to the best of our theoretical understanding) selection acting at some point, don't you think?
But, often the ratio is 1, which means that changing codons to code for other amino acids is fine, and those sites are likely drifting, and consequently so are the various alleles.
I know there are cases that would break the simple interpretation of dN/dS, but it's still a pretty powerful method to infer something about selection pressures.
Has anyone read Yang and Nielsen (2002)?
ReplyDeleteThey estimate the selection coefficients of various mutations based on the dN/dS ratio and, at least for humans, determine that the vast majority of mutations are neutral or nearly neutral.
Bjørn Østman says,
ReplyDeleteHonestly, here you sound like a clueless creationist. In fact, it is the other way around than what you argue. If all differences were neutral, then they should eventually disappear, with all alleles becoming lost or going to fixation. With selection, this is not the case. Selection can have that effect, but if polymorphisms persist for very long times (i.e., seemingly indefinitely), then genetic drift cannot explain that, while stabilizing selection can: frequency-dependent selection, sexual selection.
Bjørn, all of your comments strongly support the adapationist position. For example, you claim that you have never met a scientist who isn't a pluralist (denying that there's a problem). You claim that you are unaware of any demonstrated examples of random genetic drift. You dispute all non-adapationist hypotheses by offering adaptationist explanations. And now you are defending the selectionist view of diversity.
You may just be playing devil's advocate but it's hard for me to tell the difference.
Please read this short posting: The Cause of Variation in a Population.
Tell me honestly what your personal opinion is. Which of the following theories (hypotheses, actually) best explain the diversity in most populations in most species: Classical Theory, Balance Theory, Neoclassical Theory?
I'm not asking you to identify the exceptions—we all know about exceptions in biology. I'm asking you to identify the generality. How do you, personally, think about diversity? What causes it and why does it exist?
Oh for fuck's sake, people, adaptation and non-adaptive phenome are NOT mutually exclusive. It's like asking: "What's more important, gravity or electro-magnetic interaction?". It DEPENDS. You CANNOT prove gravity over electromagnetism and vice versa; likewise, asking whether something proves or disproves adaptation/drift is absolutely pointless and foolish. On the other hand, you can show the gravity is a more important force in certain cases (eg. when modeling a solar system), and cases where gravity becomes almost irrelevant relative to electromagnetism (eg. when designing electronic equipment). In the same way, adaptation becomes more likely to be prominent in some areas, and drift is more powerful in others.
ReplyDeleteIt's an utter waste of time to treat them as mutually exclusive, and to even discuss whether one has been 'proven' or 'disproven' in general.
Perhaps behavioural ecologists and physiologists of large mammals emphasise adaptation so much because they are akin to astronomers exploring systems largely dominated by forces of gravity. Problems would arise if said astronomers proceeded to insist their models were general to all physics and relevant at all levels, regardless of scale and context. That said, to keep with the analogy, gravity is not ultimately irrelevant in electronics, nor is electromagnetism ultimately irrelevant in astronomy; but their relative significance in the given contexts varies. Furthermore, neither system may contradict the more generalised theories of physics, and if, say, a strange EM phenomenon is observed by astronomers, it must be accounted for even if it is of little relative relevance within that context. Ie, it cannot be entirely ignored. And vice versa.
I insist we are far from ready to formulate a standardised, generalised theory of evolution. (Modern Synthesis and its reincarnations fail to cut it as many areas of biology are simply ignored there, both in terms of discipline and phylogeny), and thus we must explore all approaches and compare observations from our respective fields PRIOR to enforcing any conclusions upon each other. It's as if we were busy working out Maxwell's Laws and some oldschool Newtonian physicists walked up to us and insisted their theory explains everything, and our observations are irrelevant in the grand scheme of things. As a biologist of a smaller scale, this is how I feel some days when confronted with vicious arguments from the ecologist side.
tl;dr: fundamental evolutionary processes (say, selection, mutation, recombination and drift; adaptation being a product of selection and mutation for the sake of the argument) are akin to the fundamental 'forces' in physics, in that while differing in relative strength dependent on context, each of these forces still exists and plays a role; it is meaningless and foolish to assume them to be mutually exclusive.
PS: please don't read in to the physics-biol analogy too deeply; just an example of theories, nothing to do with actual parallels or anything. I fucking hate it when physical theories are pompously shoved into biology to underline some deep metaphysical 'truth' or whatever...
as neither random mutations nor neutral evolution have ever shown to generate new anatomy, this whole conversation is pointless.
ReplyDeleteBjørn, all of your comments strongly support the adapationist position. For example, you claim that you have never met a scientist who isn't a pluralist (denying that there's a problem).
ReplyDeleteHmm. I have several times emphasized that not every trait value is adaptive; I do not support the position that all trait values are adaptive. I do claim that I have never met a scientists who isn't a pluralist - if a not-pluralist is someone who believes all traits and trait values are adaptive (I don't deny there is a problem, but I have never encountered it - and I have asked who these people are, and I still don't know). I also still don't know of examples of traits (as opposed to trait values) that have been shown with a good level of confidence to be neutral, but - crucially - don't deny their existence, but just asking about them. Let me do that again: examples?
You dispute all non-adapationist hypotheses by offering adaptationist explanations. And now you are defending the selectionist view of diversity.
All? I dispute the ABO blood-type example. That is all I have done. And yes, I do defend that selection can explain diversity. I also think drift can do it. I am very familiar with models in which either selection or drift results in diversity.
Tell me honestly what your personal opinion is. Which of the following theories (hypotheses, actually) best explain the diversity in most populations in most species: Classical Theory, Balance Theory, Neoclassical Theory?
Neoclassical theory. Happy? (I am not sure what the relative importance is, and am unable to make anything but a qualified guess.)
The reason I am not happy just saying that is that "the cause of variation in a population" is not the only question that matters, as I have pointed out, but which you have never addressed. If I ask another question, then the answer changes: What are the causes of the major traits of species? By this I mean, for one way of putting it, traits that clearly have a function that is important to the organism and loss of which would be lethal or significantly reduce fitness. For example, heads, brains, eyes, blood-clotting, wings, limbs, DNA replication, citric-acid cycle, citrate metabolism, etc. (examples chosen so Psi Wavefunction can play along).
In summary, I second Psi Wavefunction's analogy about what one is trying to explain.
Here's a lesson from population genetics: The effects of drift and (straight-forward) natural selection is the same: within populations they both decrease genetic variation, and between populations they both increase variation (mutation increases variation in both comparisons, and gene flow increases variation within a population, and decreases it comparing populations).
anonymous says,
ReplyDeleteas neither random mutations nor neutral evolution have ever shown to generate new anatomy, this whole conversation is pointless.
How in the heck do you get "new anatomy" without ransom mutations?
And what, exactly, do you mean by "new anatomy"? Are the various differences between dandelion species examples of new anatomy? If so, what makes you so certain that none of them could be due to random genetic drift of nearly neutral alleles?
@Bjorn
ReplyDeleteCheck out this article, if you can:
"Irremediable Complexity?"
"Constructive neutral evolution" provides a possibility for complex traits to have evolved through non-adaptive processes. Examples include RNA editing, intron splicing, and duplicate gene retention. See Stoltzfus 1999 for more details. (Selection's role is negative here - it prevents reversals a la Muller's ratchet.)
These have not been proven to be a result of non-adaptive processes, but they provide alternatives for some major "traits."
Another possible example is codon bias. Many assume that codon bias reflects selection for optimal codons and weakens the Neutral Theory's proposal that many/most mutations are neutral. This paper which I stumbled upon a few days ago says that most codon bias is a result of GC mutation bias. http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0013431 I cannot say for sure whether the authors are right, but again, it's an alternative possibility. (My evidence that "many" believe this is selective is because this is what my evolution textbook, Evo Analysis by Freeman & Herron says.)
Again, I don't know which of these is true, but there are at least other possibilities that I think should be explored.
@anonymous: False.
ReplyDeleteLukes et al. 2009 PNAS* – intricate RNA pan-editing and minicircles in the trypanosome kinetoplast (mitochondrial DNA)
Lambowitz intron in Neurospora that lost its ability to self-splice following an initially fortuitous interaction witha tRNA loading enzyme that eventually took on the role of a spliceosome – definitely some new anatomy there (explained in Stoltzfus 1999 J Mol Evol**)
See Grey et al. 2010 Science for a concise explanation of constructive neutral evolution, or a mechanism by which complexity can ratchet up without net fitness gain. A more thorough explanation of CNE can be found in Stoltzfus 1999 JME.
For something more macroscopic, I'd argue the following to be a solid case of net non-adaptively increased complexity: There was a case in the 60's where a parasitic ('X') bacterium plagued a culture of Amoeba proteus (D-strain). After some time, some individuals of the D strained evolved resistance to the X bacterium, or somehow the X bacterium became less virulent to it – getting called 'x' strain at this point. This step is presumably adaptive for the amoeba, for obvious reasons. Now, here comes the twist – the amoebae actually became -dependent- upon the x-bacterium, which by this point has become a mutualistic endosymbiont. Removing this bacterium from the amoeba resulted in its ultimate demise (the bacteria were already obligate [parasitic] endosymbionts.****
It gets better – the symbiont produces several large proteins that are targeted towards the host nucleus. Furthermore, in 2000's it was revealed the host lost the ability to transcribe some critical metabolic gene.
Here we have a case of essentially a new organelle in the process of being obtained, all within a few decades in the lab. What is incredibly interesting in the perspective of non-adaptive evolution, however, is that these new 'xD' amoebae are actually physiologically -worse off- than their regular D counterparts. They are more sensitive to starvation and overfeeding, more sensitive to crowding, grow slower and generally can be argued to have lower fitness. Here we have a process of increased anatomical complexity with a net *decline* in fitness, showing that overall, this process is probably non-adaptive.
I'd argue this is also a case of constructive neutral evolution, where a previously irrelevant or slightly deleterious interaction ultimately becomes essential to the organism by allowing it to degrade in some function that this interaction makes redundant (ie 'presuppression' sensu Stoltzfus 1999).
There are doubtlessly multitudes more of such cases of net non-adaptive complexity gain, it's just that few have tested this framework outside molecular biology. Of course, on my scale, the genome IS part of anatomy, and increases in genome complexity are anatomical – introns, for example, are new anatomy. Spliceosomes are new anatomy (initially, introns were self-splicing.) For a good introductory overview of non-adaptive genome evolution and importance of population genetics, see Lynch 2007 PNAS; complexity gain in genet pathways: Lynch 2007 Nature Rev*****.
Care must be taken to not confuse the current function of some structure with its initial function (or lack thereof). Non-adaptive theories do not preclude subsequent 'exaptation' of initially pointless structures for something adaptively useful. But they allow one to explain the rise and persistence of net deleterious changes.
Links to refs:
*Lukes et al 2009: http://www.cmde.science.ubc.ca/pdf/PNASEugDino.pdf
** Stoltzfus 1999 J Mol Evol: http://bioinformatics.bio.uu.nl/BINF/pdf/Stoltzfus.jme99.pdf
*** Grey et al. 2010: http://www.bc.cas.cz/doc/news/Gray-Science-2010.pdf
**** Jeon 1995 J Euk Micro http://onlinelibrary.wiley.com/doi/10.1111/j.1550-7408.1995.tb01532.x/abstract (sadly, paywalled. I'll blog about it sometime...)
***** Pdfs of Lynch 2007 PNAS, Nature Rev Genet and others can be found here: http://www.indiana.edu/~lynchlab/LynchPublications.htm
C'mon, Larry. Anonymous is a creationist, already. New anatomies are designed, for sure.
ReplyDeleteFollowing Micheal M's comment from Friday, February 25, 2011 10:18:00 PM,
ReplyDeletedN/dS actually performs quite poorly when the moelcular details of the gene are known.
See;
http://www.ncbi.nlm.nih.gov/pubmed/18768803?dopt=Abstract
and http://www.ncbi.nlm.nih.gov/pubmed/18768804?dopt=Abstract
Research paper's abstract;
Vertebrate ancestors appeared in a uniform, shallow water environment, but modern species flourish in highly variable niches. A striking array of phenotypes exhibited by contemporary animals is assumed to have evolved by accumulating a series of selectively advantageous mutations. However, the experimental test of such adaptive events at the molecular level is remarkably difficult. One testable phenotype, dim-light vision, is mediated by rhodopsins. Here, we engineered 11 ancestral rhodopsins and show that those in early ancestors absorbed light maximally (lambda(max)) at 500 nm, from which contemporary rhodopsins with variable lambda(max)s of 480-525 nm evolved on at least 18 separate occasions. These highly environment-specific adaptations seem to have occurred largely by amino acid replacements at 12 sites, and most of those at the remaining 191 ( approximately 94%) sites have undergone neutral evolution. The comparison between these results and those inferred by commonly-used parsimony and Bayesian methods demonstrates that statistical tests of positive selection can be misleading without experimental support and that the molecular basis of spectral tuning in rhodopsins should be elucidated by mutagenesis analyses using ancestral pigments.
dN/dS actually performs quite poorly when the moelcular details of the gene are known.
ReplyDeleteHow does this affect the estimation of the scaled selection coefficients (S=2Ns)?
There are draw-backs to every test of neutrality, but merely pointing out that the particular test used isn't perfect doesn't necessarily invalidate the results. In particular, I'm puzzled by your citation, because, as far as I understand the abstract, the investigators report that dN/dS overestimates the strength of positive selection on genes of know function, implying that the the number of neutral substitutions is actually underestimated.
As much as I dislike arguments of importance based on magnitude, don't the findings in the article you cited actually support the idea that neutral and nearly neutral evolution are more important that adaptationists claim?
Micheal M says,
ReplyDeletedon't the findings in the article you cited actually support the idea that neutral and nearly neutral evolution are more important that adaptationists claim?
Yes. I was actually agreeing with you.
I think both of those papers are accessible (not behind the PNAS pay wall). The "comment on" paper really drives that point home.
Dave Speijer presents an argument against neutral evolution:
ReplyDeletehttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/
Earlier I had posted:
ReplyDelete"Dave Speijer presents an argument against neutral evolution:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/"
So we have the interesting situation where adaptationist explanations fail, and neutral evolution explanations also fail.
Perhaps the real explanation lies elsewhere.
ReplyDeleteEarlier I had posted:
"Dave Speijer presents an argument against neutral evolution:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/"
So we have the interesting situation where adaptationist explanations fail, and neutral evolution explanations also fail.
Perhaps the real explanation lies elsewhere.
And Keeling et al responded to that criticism here: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840371/
The original criticism seems to be flawed in its interpretation of exactly what Constructive Neutral Evolution really means, and the nature of evolutionary convergence at the molecular level.
Personally I find the CNE explanation for the convergent evolution of needlessly complex molecular systems like what are discussed in that article far more compelling then simply invoking positive selection with zero evidence to support it.
It always seems to be missed that under situations like CNE selection isn't being ignored entirely, purifying selection still plays a role, just not in the fixation of the original mutations.
How does constructive neutral evolution differ from "God did it"?
ReplyDelete
ReplyDeleteHow does constructive neutral evolution differ from "God did it"?
For starters it is a hypothesis put forward based on observations of evolutionary changes at the molecular level, the majority of which appear to be selectively neutral. None of the mechanisms that are invoked are at all controversial or surprising unless you think complexity can only be the product of positive selection.
"For starters it is a hypothesis put forward based on observations of evolutionary changes at the molecular level, the majority of which appear to be selectively neutral. None of the mechanisms that are invoked are at all controversial or surprising unless you think complexity can only be the product of positive selection."
ReplyDeleteThe "God did it" explanation is exactly the same.
It is an "hypothesis put forward based on observations of evolutionary changes at the molecular level, the majority of which appear to be selectively neutral. None of the mechanisms that are invoked are at all controversial or surprising unless you think complexity can only be the product of positive selection."
In the neutral evolution theory there is no explanation whatever of the mechanism involved. Or why things happen as they do.
At least in the "God did it" explanation there is a form of explanation. The changes which appear to be neutral are brought about by a level of intelligence such as the intelligence of Nature. You can disagree with that explanation, but at least it is an explanation, compared to no explanation whatever with the neutral theory.
Right?
From the article by Dave Speijer:
ReplyDeletehttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
This objection does not seem to have been addressed.
ReplyDeleteFrom the article by Dave Speijer:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
This objection does not seem to have been addressed.
It is addressed through simple observation of molecular sequence data. It would appear that the majority of fixations, even at the amino acid level, in most organisms are effectively neutral, and not the product of selection.
And this goes beyond simple dN/dS ratios as tests for positive selection, many non-synonymous changes will be neutral in most molecular contexts. Tests for functional importance of residues tends to bear out this observation, as do many mutagenesis studies. Many sites are under little to no purifying selection, may even be experiencing high indel rates, or will tolerate a wide range of amino acids within broad property absed characterizations (such as dayhoff class).
So it isn't surprising for neutral mutations to be fixed again and again at the molecular level, it seems to happen rather frequently.
From the article by Dave Speijer:
ReplyDeletehttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
DG has responded by saying that
"it seems to happen rather frequently."
That is not an explanation.
Let's use an analogy.
A set of people win the lottery everyday for a year. Some people explain this by saying that this "seems to happen rather frequently" as if that were an explanation. Others think that there is actually an explanation - for example that there is cheating going on.
Neither selection nor "neutral changes" that just happen by chance, explains what is going on. The changes cannot just be happening by chance and they cannot just be going to fixation by chance. There must be intelligence involved since what is occurring cannot happen simply by chance.
That is what underlies the point that Dave Speijer is making when he says:
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
I am not looking for an unpleasant argument.
Can anyone see the point I am making?
I'm not going to bother engaging in pointless debate with someone clearly advocating intelligent design creationism while using a paper that invokes selection as better explaining the data over CNE.
ReplyDeleteThere are lots of great posts on this blog and others for just why the "this can;t be chance it must be intelligence" is complete hogwash.
From the article by Dave Speijer:
ReplyDeletehttp://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
This point by Dave Speijer has not been addressed.
DG seems not willing to address it.
Can anyone else address this?
Let's use an analogy.
ReplyDeleteA set of people win the lottery everyday for a year. Some people explain this by saying that this "seems to happen rather frequently" as if that were an explanation. Others think that there is actually an explanation - for example that there is cheating going on.
Your analogy is broken. Evolution is not a series of independent lotteries.
Your mistake is #4 in the "Five Major Misconceptions about Evolution" list.
http://www.talkorigins.org/faqs/faq-misconceptions.html
The Other Jim is missing the point.
ReplyDeleteNeutral changes do not rely on selection. (They are neutral).
I hope that people see the point I am making. Point #4 in the talkorigins article is irrelevant to what I am saying.
ReplyDeleteWe are left with the objection that Dave Speijer has made:
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2840355/
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
This shows that neutral theory cannot account for the fixation of neutral changes.
But we also see that selection cannot account for what we observe.
We have to face up to the fact that NEITHER explanation works.
The answer must lie elsewhere.
anonymous says,
ReplyDeleteWe are left with the objection that Dave Speijer has made: "Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
This shows that neutral theory cannot account for the fixation of neutral changes.
It is true that the fixation of a particular neutral allele is a low probability event. However, there are so many of them that fixation of neutral alleles dominates molecular evolution.
Speijer is addressing some "elaborate idiosyncrasies" in Alveolata and Euglenozoa. He is expressing some skepticism over the claim that these "elaborate idiosyncrasies" could be entirely due to fixation of neutral alleles. As part of his objection he points out, quite rightly, that the fixation of particular neutral alleles is an improbable event.
He is NOT saying that fixation by random genetic drift doesn't happen. That's why he writes,
Obviously, evolution results from the interplay between variation (chance processes) and selection. Contributions of neutral processes, such as random mutation, genetic drift, and chance fixation have long been underestimated. Theoretically, slight differences in likelihoods of molecular occurrences and their reversals (such as nucleotide deletion vs. insertion in genes having to be restored at the RNA level) could lead to accumulation of widespread instances of “useless” processes.
HTH
You have said:
ReplyDelete"It is true that the fixation of a particular neutral allele is a low probability event. However, there are so many of them that fixation of neutral alleles dominates molecular evolution."
Can we analyze that a bit please?
Do we find a scattering of particular neutral alleles? One here, a different one there? Another one here, another different one there etc?
Do we find sets of alleles continuing?
Do you see what I am asking?
Here is a clearer version of what I am saying:
ReplyDeleteYou have said:
"It is true that the fixation of a particular neutral allele is a low probability event. However, there are so many of them that fixation of neutral alleles dominates molecular evolution."
Can we analyze that a bit please?
Do we find a scattering of particular neutral alleles? One here, a different one there? Another one here, another different one there etc?
Or do we find SETS of neutral alleles continuing from generation to generation?
Do you see what I am asking?
anonymous asks,
ReplyDeleteDo we find a scattering of particular neutral alleles? One here, a different one there? Another one here, another different one there etc?
Or do we find SETS of neutral alleles continuing from generation to generation?
Do you see what I am asking?
If you're talking about neutral substitutions in proteins then there's no evidence that I know of for coevolution of sets of neutral alleles.
How does one neutral allele here and another unrelated one there and another here and another there, result in anything?
ReplyDeleteWhat effect is being claimed for the sprinkling of these neutral alleles all over the place?
Can you give an example or two?
It looks like this discussion has dried up.
ReplyDeleteWe are left with the objection that Dave Speijer has made: "Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
This shows that neutral theory cannot account for the fixation of neutral changes.
I emailed the following to Dave Speijer:
ReplyDelete"You have said that:
> "Neutral changes have to take over the complete population by chance
> alone. It is highly unlikely but has to occur again, and again ."
> This seems to me to be correct and is a general principle related to
> the claims of neutral theory.
>
> Is that your thinking also?"
He replied by saying:
"Yes'.
This confirms that the objection he is raising is not just in regard to the specific experiment by Lukes et al, that he is discussing, but is a general objection that:
Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
ReplyDelete"You have said that:
> "Neutral changes have to take over the complete population by chance
> alone. It is highly unlikely but has to occur again, and again ."
> This seems to me to be correct and is a general principle related to
> the claims of neutral theory.
>
> Is that your thinking also?"
He replied by saying:
"Yes'.
This confirms that the objection he is raising is not just in regard to the specific experiment by Lukes et al, that he is discussing, but is a general objection that:
Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
And your point is? It is clear Dr. Speijer prefers adaptationist explanations to neutral ones and much prefers to focus on positive selection judging by his recent publication on the topic in Bioessays.
None of this is going to help your argument, which seems to be that neither neutrality, selection, or the combination of the two (which in reality, is what is always going on) can explain complexity.
Bringing up someone arguing for selection to dismiss neutral explanations in your bid to insert intelligent design seems more than a little deceptive.
Dave Speijer has said:
ReplyDelete"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
DG is having trouble understanding this point.
Assuming the mutation rate, estimated in Genetics, Vol. 156, 297-304, 2000 (2.75x10^-8 per base), we each have about 175 variations from our parent's DNA (and I mean the chromosome that our DNA was copied from).
ReplyDeleteBy the time I type this message, this number will be out-of-date, but there are about 6879114400 people on the planet.
You are correct that variation #1 on my DNA has little chance of fixing. But there are 1.2x10^12 new variants in the human population (relative to their parental "reference sequence" - this is complicated by 2-3 generations of a human lineage co-existing at the moment, but this is "back of the napkin" math for illistration).
These will not be the lethal mutations, as the person is alive, so purifying selection has reduced the number of target sites possible. What are the chances that some of these neutral alleles can fix?
Remember that there are only about 5.8x10^10 bases in our diploid genome, and the effective population size in humans is estimated around 10,000.
I'll save you cracking open the pop-gen textbook and running the simulation. The chances are not that bad that some neutral variant will fix. This is different than asking if "that specific one" will fix.
And can I just add here, in response to the person quoting the article from Speijer, that the author is merely asserting his personal opinion that neutral evolution is "unlikely". For about a century, population geneticists have been building on a formal theory for estimating the probabilities of things such as the fixation of a neutral allele. Speijer ignores that and goes with his gut. His more recent piece in BioEssays does the same thing-- he merely asserts his personal opinion that neutral evolution is unlikely, giving no reasons and citing no references.
ReplyDeleteArlin
Arlin can you describe an example of neutral alleles going to fixation please?
ReplyDeleteThat would be very helpful.
Anonymous, you asked for an example of neutral alleles going to fixation. This process was demonstrated by Buri (1956), whose experiments are famous, and whose results fit nicely with the mathematical theory.
ReplyDeleteIf you just google "buri drift 1956" you will find lots of pages that can explain these results. You may find it rewarding to study foundational works like this.
I had said earlier:
ReplyDeleteArlin can you describe an example of neutral alleles going to fixation please?
That would be very helpful.
I find that people will make general comments but do not back them up with examples.
I wonder if anyone could give a reference more current than 1956.
ReplyDeletePlease provide an actual link.
you want me to "describe the example", in the sense of spoon-feeding it to you?
ReplyDeleteYou're going to have to work harder than that. The Buri experiment is described in section 8.3 "Empirical examples of genetic drift" on p. 138 of Neal, Introduction to Population.
The book is available on scribd here:
http://www.scribd.com/doc/50358194/Dick-Neal-Introduction-to-Population-Biology-Cambridge-University-Press-2004
Is there anyone else who can give a link to a more current reference than 1956?
ReplyDeleteanonymous, I just *did* give you a more recent reference. I even provided a link.
ReplyDeleteIf you don't read about the Buri experiment there, then I'm going to conclude that you are not interested in learning enough to understand how your questions can be answered, and instead are just trying to bait me.
ReplyDeleteArlin said...
anonymous, I just *did* give you a more recent reference. I even provided a link.
If you don't read about the Buri experiment there, then I'm going to conclude that you are not interested in learning enough to understand how your questions can be answered, and instead are just trying to bait me.
I'm pretty sure the latter is actually the case. It's not like those results have been overturned. It's a pretty foundational experimental proof of the math, which is solid, and fairly intuitive all on its own.
Arin I don't care enough about you to "bait" you.
ReplyDeleteI am trying to get info.
If I offered up data from 1956 I would be laughed out of town.
You have not helped at all and we both know it.
Is there anyone else who can provide a link to a reference of data more current than 1956?
Anonymous:
ReplyDeleteYou would be laughed out of town if the information you gave was not just old, but outdated and incorrect. The experiment that Arlin has given you is not outdated and incorrect, it is foundational and represents a clear and early bit of experimental evidence for the math behind population genetics.
He also provided a more recent book reference that explains that work. Perhaps you should check the book and start following references to other experiments that may be contained in it.
Anonymous - it would depend on what you "accept" as evidence for neutral mutations going to fixation.
ReplyDeleteFor example, I or any number of people here could refer you to papers showing elevated rates of substitution in pseudogenes, for example, that are untranslated and have no known function.
Under an evolutionary framework they would be assumed to evolve neutrally or close to it, and their accelerated rate of substitution would be evidence that not only do neutral mutations fix but they can do so in great numbers when freed from the constraints of purifying selection.
But then, if you were to have certain creationist leanings, you might not accept that pseudogenes are most likely functionless. If this were the case, there would be simply no point in linking you to those papers.
It seems this might be the case. You have presented no reason to suspect anything about the classic experiment that Arlin referred you to (and no age is not a 'reason').
Anonymous said:
ReplyDelete"I am trying to get info.
If I offered up data from 1956 I would be laughed out of town.
You have not helped at all and we both know it. "
Data are data, form 1956 or from 1828: it is the quality of the researcy, not the year. The experiment you are referred to is a classic. Classic experiments remain valid data at any time scale.
If you don't know that, you shouldn't be commenting.
Okay, it looks like nobody can provide a reference to data more current than 1956.
ReplyDeleteSo be it.
Anonymous:
ReplyDeleteThe Book Arlin linked to also discusses many other issues related to Drift, Neutrality, Selection, and the Fixation of alleles besides just the classic 1956 experiment.
Kimura 1991 is a review paper talking about many experimental observations in support of neutral theory (not just fixation of neutral alleles but neutral theory in general)
Much of the work on fixation of neutral alleles has to do with synonymous versus non-synonymous sites, non-synonymous changes at the amino acid/protein level, etc. Pick a protein of interest and look for papers on its evolution and you have a wealth of analyses on this topic. Hemoglobin (alpha and beta), myoglobin, and any other protein family you can think of have been favourite choices in molecular evolution.
Much of the rest of the experimental studies on drift focus on what happens to beneficial, deleterious, slightly deleterious, and nearly neutral mutants under drift.
The experimental argument over the last 60 years or more has been on the relative importance of drift versus selection. That a mutant will eventually be fixed or lost in a population is a given. Its a question of how long and what effect the population size and selective coefficient have on that. The basic foundational literature is old because that is when the work was done. More recent work has gone on to investigate other parts of the story.
You've been given some good starting points to read from. That's academic research. Find a good starting point and branch out.
We have kind of veered off the point I was making:
ReplyDeleteDave Speijer has said:
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
My questions about the fixation of neutral alleles was to get a simple example of the fixation of neutral alleles so we could actually evaluate Speijer's point.
Anonymous Said...
ReplyDeleteWe have kind of veered off the point I was making:
Dave Speijer has said:
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
My questions about the fixation of neutral alleles was to get a simple example of the fixation of neutral alleles so we could actually evaluate Speijer's point.
And you got it. The older paper is a classic experimental example. If you look at molecular work on almost any protein of interest you have a wealth of examples.
Most (but not all) Synonymous substitutions at the DNA level are neutral and represent neutral fixations.
All fixations in non-transcribed pseudogenes represent fixation of neutral alleles.
Etc
Anonymous;
ReplyDeleteThe 1956 experiment demonstrated the phenomenon, and the data we observe continues to support their conclusions. So, Lets generate some new data on the topic that behave exactly as predicted by fixation of neutral alleles.
I downloaded the mitochondrial COX1 protein sequences (NP_008188.1 and NP_008199.1) from Genbank for the chimp and the bonobo. These two genes are of identical length (513 amino acids) and differ from each other by 2 amino acids (position 24 T or A; position 155 V-I). We have no experimental evidence to discuss whether these two AA changes are neutral or not, so we will just ignore them.
I also downloaded the corresponding DNA sequences. They vary from each other at 60 sites. 2 of them (nt 70 and nt 463) account for the two amino acid changes. This means 58 of them do not change the amino acid sequence (genetic code redundancy). Almost all are 3rd codon position changes that utilize the same tRNA for translation, so arguing a fitness difference between the two codons is a little difficult. These 58 changes represent neutral changes in this single gene since the split of the chimp and bonobo over the last ~2 million years.
Please also note that no one is dismissing General Relativity (1915) or the Avery–MacLeod–McCarty experiment (1944) based on how old the publications are.
The issue is not with the data. The issue is with the interpretation.
ReplyDeleteIt is commonplace in science that data from an earlier period is re-interpreted in a later period.
the other Jim posted:
ReplyDelete"The 1956 experiment demonstrated the phenomenon, and the data we observe continues to support their conclusions. So, Lets generate some new data on the topic that behave exactly as predicted by fixation of neutral alleles.
I downloaded the mitochondrial COX1 protein sequences (NP_008188.1 and NP_008199.1) from Genbank for the chimp and the bonobo. These two genes are of identical length (513 amino acids) and differ from each other by 2 amino acids (position 24 T or A; position 155 V-I). We have no experimental evidence to discuss whether these two AA changes are neutral or not, so we will just ignore them.
I also downloaded the corresponding DNA sequences. They vary from each other at 60 sites. 2 of them (nt 70 and nt 463) account for the two amino acid changes. This means 58 of them do not change the amino acid sequence (genetic code redundancy). Almost all are 3rd codon position changes that utilize the same tRNA for translation, so arguing a fitness difference between the two codons is a little difficult. These 58 changes represent neutral changes in this single gene since the split of the chimp and bonobo over the last ~2 million years."
I appreciate that you have taken the time to do this analysis work.
I would be very interested in analyzing this further with you.
Concerning the "58".
Is there any way that we could see that they have a role? For example that they perform a "regulatory" function?
This is the kind of thinking that people are beginning to use.
Your thoughts?
Concerning the "58". Is there any way that we could see that they have a role?
ReplyDeleteI picked a mitochondrial gene on purpose. There is currently no evidence that I am aware of that suggests a within-gene regulation of that genome.
But there is codon bias, but no clear idea of what that issues that would cause in such closely related species. But there is only a single tRNA for the codon family, so "tRNA choice efficiency" is not behind it.
The other Jim wrote:
ReplyDeleteThis means 58 of them do not change the amino acid sequence (genetic code redundancy).
To which anonymous replies:
I would be very interested in analyzing this further with you.
Concerning the "58".
Is there any way that we could see that they have a role? For example that they perform a "regulatory" function?
They don't change the result. Therefore, there is no regulation taking place. (If you wave your hand and the burner setting on your stove doesn't change, would you say waving your hand "regulates" the burner?)
The other Jim posted
ReplyDelete"These 58 changes represent neutral changes in this single gene since the split of the chimp and bonobo over the last ~2 million years."
Does this mean that those "58" went to fixation in the entire population and remained in place for up to 2 million years (even though they are not beneficial?
Does this mean that those "58" went to fixation in the entire population and remained in place for up to 2 million years (even though they are not beneficial?
ReplyDeleteMutations have been occurring all of this time, since the populations split. Many vanished from the population by stochastic sampling, and others spread through and fixed in the population (such as these ones).
And remember, this is one gene only.
I have a question.
ReplyDeleteWhy would these "58" neutral alleles go to fixation in the entire population?
What are the chances of that? Given that there is no survival reason for them.
How do you visualize that process?
Anonymous:
ReplyDeleteThe process is called Genetic Drift, the book Arlin linked to is an intro to Population Genetics book that covers the topic. If you want to understand you can read it there.
DG has chosen not to answer my questions.
ReplyDeleteCan anyone else?
Here is an absolutely fascination entry from wikipedia:
ReplyDeletehttp://en.wikipedia.org/wiki/Fixation_(population_genetics)
"Because the effect of natural selection is stipulated to be negligible, the probability at any given time that an allele will ultimately become fixed at its locus is simply its frequency p in the population at that time. For example, if a population includes allele A with frequency equal to 20% and allele a with frequency equal to 80%, there is an 80% chance that after an infinite number of generations a will be fixed at the locus (assuming genetic drift is the only operating evolutionary force)."
This is clever. It describes what happens after an infinite number of generations. That is not the question.
The question is - what happens after the number of generations that we would estimate as being reasonable to have actually occurred?
They don't talk about that, because then they would confirm the point that Dave Speijer is making:
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
Anonymous:
ReplyDeleteI answered the question by telling you that you have already been given the answer you seek. The wikipedia entry on Genetic Drift also has an excellent example using random sampling of paint cans.
During reproduction alleles in a population are being randomly sampled from one generation to the next. Repeat this sampling over and over again and eventually allele variants will be lost from the population, with only one remaining (being fixed in the population). That is genetic drift in a nutshell. You can set up any number of "experiments" yourself to simulate how this process works. The paint can and marble examples being too of the more well known ones from a pedagogical perspective.
Selection and population size come in on top of this random sampling effect.
Anonymous writes:
ReplyDeleteDG has chosen not to answer my questions.
Can anyone else?
Religious books contain a great deal of wisdom.
In the Talmud, volumes of interpretive commentaries on the Torah, there is the story of the man who asked Hillel if the great sage could teach him the whole meaning of the Torah while the questioner stood on one foot.
Hillel considered, then said: "That which is hateful to you, do not to your neighbor. That is the whole Torah; the rest is commentary. Go and study it."
Your questions appear to be no more sincere than those of the doubter who demanded to be taught the whole of the Torah while standing on one foot.
Nevertheless, like Hillel, DG and the others here have told you the "whole of the law" and more. They have told you about population genetics and how neutral alleles come to fixation in a population through genetic drift, and they have given you detailed examples. They have told you where to learn more. Now, as Hillel told the doubter, it is time for you to "go and study it."
The question you are asking is often not answered until the upper levels of a biology undergraduate degree. 10 lines in a blog comment field will not cut it.
ReplyDeleteThe reader's digest version is on wiki.
http://en.wikipedia.org/wiki/Genetic_drift
If you actually want to learn it, at least go to wiki. Plugging your ears and screaming "I can't hear you" only enforces our suspicions that you are trolling.
Anonymous:
ReplyDeleteYes, that is the probabilistic definition (and a somewhat simplified explanation of it at that). I have already pointed you to excellent examples that you could set up and run yourself to prove that in a finite population, within a finite number of generations, you will have fixation of an allele.
http://www.biology.arizona.edu/evolution/act/drift/drift.html
Is a site designed for teaching evolutionary concepts to school children, and even includes a simple simulation program to illustrate drift.
http://en.wikipedia.org/wiki/Fixation_(population_genetics)
ReplyDelete"Because the effect of natural selection is stipulated to be negligible, the probability at any given time that an allele will ultimately become fixed at its locus is simply its frequency p in the population at that time. For example, if a population includes allele A with frequency equal to 20% and allele a with frequency equal to 80%, there is an 80% chance that after an infinite number of generations a will be fixed at the locus (assuming genetic drift is the only operating evolutionary force)."
AND
http://en.wikipedia.org/wiki/Genetic_drift
"P2 takes his 20 cans, now full of paint, off to the back of his own truck. P3 awaits him with 20 empty cans. Continue ad infinitum. Eventually, a painter will walk away with 20 cans all of the same color.
However, evolution theory claims to be working in a finite time.
And concerning finite time (ie. the real world) here is what Dave Speijer has said:
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
Insulting me with these ad hominem attacks does not change the facts.
http://en.wikipedia.org/wiki/Fixation_(population_genetics)
ReplyDelete"Because the effect of natural selection is stipulated to be negligible, the probability at any given time that an allele will ultimately become fixed at its locus is simply its frequency p in the population at that time. For example, if a population includes allele A with frequency equal to 20% and allele a with frequency equal to 80%, there is an 80% chance that after an infinite number of generations a will be fixed at the locus (assuming genetic drift is the only operating evolutionary force)."
AND
http://en.wikipedia.org/wiki/Genetic_drift
"P2 takes his 20 cans, now full of paint, off to the back of his own truck. P3 awaits him with 20 empty cans. Continue ad infinitum. Eventually, a painter will walk away with 20 cans all of the same color.
However, evolution theory claims to be working in a finite time.
And concerning finite time (ie. the real world) here is what Dave Speijer has said:
"Neutral changes have to take over the complete population by chance alone. It is highly unlikely but has to occur again, and again …"
Insulting me with these ad hominem attacks does not change the facts.
You are once again hung up on the probability of one specific allele fixing. So is Dave Speijer.
ReplyDeleteWe are trying to point out that only a very tiny fraction of the millions that are generated every generation fix.
the other Jim, you are misrepresenting what I am saying.
ReplyDeleteAnd misrepresenting what Speijer is saying.
Nobody has answered what I said earlier:
Why would these "58" neutral alleles go to fixation in the entire population?
What are the chances of that? Given that there is no survival reason for them.
How do you visualize that process?
But I am being tired of the run-around from people here.
You don't actually know what you are talking about.
Anonymous:
ReplyDeleteYou're not getting the run around. But you're providing very little motivation for anyone to want to take the time to give in-depth explanations of evolutionary biology given you're apparent unwillingness to read material that has been linked to you beyond Wikipedia.
Did you go to the tutorial that I linked to? The one with a simulator (limited though as it is and meant only for simple educational purposes) built in? It provides the explanation for why, in the absence of any selective value, alleles will eventually be fixed or lost in a population and that is a pretty simple one. Even in the absence of any selective value there will still be differential reproduction in populations due to random events.
Animals get caught in a forest fire, bugs get stepped on, etc. There are any number of reasons, besides selection, for why organisms will leave behind different numbers of offspring. This differential reproduction means that every generation alleles are being randomly sampled from one population to the next, and eventually alleles will be lost during this process, meaning that some will therefore become fixed.
We talk about infinite time because that is the correct probabilistic way to talk about that process, that doesn't mean it actually takes an infinite amount of time for an allele to become fixed. Like I said, you can do the experiment yourself with a small population, and fixation will occur in a relatively short amount of time.
Even Speijer, who prefers adaptive explanations, wouldn't argue that the fixed mutations in pseudogenes (there are lots) weren't neutral, or that many (but not all) synonymous substitutions weren't selectively neutral.
I have not been asking for in-depth explanation of evolutionary biology.
ReplyDeleteI am just wanting to analyze a real world example of the fixation of neutral alleles.
But I realize that people do not know this material at the level of a real world example.
I would have to say that my March 25th post is a "real world example" ;-)
ReplyDeleteYesterday I had begun to type up a rather lengthy answer and decided to not to bother.
ReplyDeleteHere's the long and short of it anyway:
There is a robust set of mathematical equations and models that describe the fluctuation of alleles in a finite population over time under genetic drift, and include parameters for the strength of selection on an allele (neutral, purifying, positive), as well as the effective population size.
Even in the absence of selection, there will be differential reproduction due to chance forces. Organisms with the same fitness will not all leave the same number of offspring, and the alleles present in the next generation will be a random sample of those present in the parental generation. Because we are dealing with finite populations this sample will not be the same as the frequency of the alleles, that is just simple statistics.
So under pure drift (and even with selection present) we have certain predictions that fall out of what should be happening under genetic drift. For alleles under no selection we expect a loss of diversity (heterozygosity in the two allele case), over time because of this random sampling process (fixation of one of the two alleles).
The Buri experiment (as well as a bunch of other experiments in the 50's by Wright) aptly demonstrate this process.
The process is essentially the same whether or not selection is present. Strong positive selection means that alleles are less likely to be lost under drift, strong purifying selection and the allele is less likely to be fixed. But there is no gaurentee, sometimes less fit alleles get fixed, sometimes beneficial ones are lost, this is especially true in small populations (bottlenecks and founder effects).
If you really want to understand this, please... pick up one of the many population genetics books out there and read through it a little, it goes into all of this in much more detail.
Many of the more recent experimental studies that focus on drift don;t necessarily focus on fixation of neutral alleles. Many look at the fixation of poor alleles in fact under small populations. But many do focus on things like the loss of within population heterozygosity and increase in between population differentiation when larger populations become divided.
Templeton has a series of articles dating from something like 1990 to the present on populations of collared lizards in the Ozarks. There are a ton of experimental studies on the effects of population divisions on plants. Young is one researcher to check out and to start from.
DG it is very good of you to take the time to present info about the topic of fixation of neutral alleles but I am not putting any more study into something that I know is incorrect.
ReplyDeleteThat is why I want to just work through a real life example (that is claimed to be the fixation of neutral alleles) and we will all see that the idea is incorrect.
We could work with the "58" but someone will have to provide more info about it.
How do you see the process? For example, did it begin with a sprinkling of alleles with each one having a .0000001 per cent frequency?
How do you visualize it?
http://www.ncbi.nlm.nih.gov/About/primer/genetics_genome.html
ReplyDelete"There are four basic blood types, and they are O, A, B, and AB. We know that our blood type is determined by the "alleles" that we inherit from our parents. For the blood type gene, there are three basic blood type alleles: A, B, and O. We all have two alleles, one inherited from each parent. The possible combinations of the three alleles are OO, AO, BO, AB, AA, and BB. Blood types A and B are "co-dominant" alleles, whereas O is "recessive". A codominant allele is apparent even if only one is present; a recessive allele is apparent only if two recessive alleles are present. Because blood type O is recessive, it is not apparent if the person inherits an A or B allele along with it. So, the possible allele combinations result in a particular blood type in this way:
OO = blood type O
AO = blood type A
BO = blood type B
AB = blood type AB
AA = blood type A
BB = blood type B
You can see that a person with blood type B may have a B and an O allele, or they may have two B alleles. If both parents are blood type B and both have a B and a recessive O, then their children will either be BB, BO, or OO. If the child is BB or BO, they have blood type B. If the child is OO, he or she will have blood type O."
Would I be right in thinking that within some finite time one of these alleles will go to fixation?
It looks like the folks here do not know the subject of neutral alleles very well.
ReplyDelete