Thursday, March 13, 2014

The "selfish gene" is not a good metaphor to describe evolution

Last December David Dobbs wrote an article entitled "Die, selfish gene die!" in which he promoted some really stupid ideas about evolution. Jerry Coyne wasted little time in showing why Dobbs was way off the mark.

Unfortunately, Dobbs didn't make the best case against the concept of the selfish gene and Jerry Coyne didn't recognize that there was a real problem.

Now the issue has resurfaced because Aeon has announced another dead horse that needs beating [Dead or Alive? Is it time to kill off the idea of the ‘Selfish Gene’? We asked four experts to respond to our most controversial essay].

Once again Jerry Coyne is up to the challenge, taking on four "experts" without breaking a sweat [The “selfish gene” redux: Aeon magazine collects opinion on the metaphor].

And, once again, everyone misses the real point.

Here's what Jerry says in his latest post.
I won’t reprise my criticisms, except to say that the metaphor of genes acting as if they are "selfish" when subject to natural selection remains perfectly good, whether or not those genes (or any bit of DNA) are part of the genome that makes proteins, regulates other genes, or comprises any bit of DNA that has the ability to get itself replicated more often than its competitors.
This is correct as far as it goes. The "selfish gene" is a reasonable metaphor if you want to think about natural selection and adaptation. It's quite reasonable to metaphorically describe genes as "selfish" in such cases.

However, I think that "selfish gene" is often used as a metaphor for all of evolution and not just for natural selection. I think that most people who read Dawkins' book take it to be about EVOLUTION and not just natural selection. They understand that Dawkins is promoting a gene-centric view of evolution—and that's okay—but they come away from reading the book by thinking that all genes are selfish.

As most of you know, I prefer to emphasize Evolution by Accident. That's not to say that natural selection (selfish genes) isn't important, it is. What I believe is that there is a lot more to evolution than just selfish genes and we should not use the selfish genes metaphor as a stand-in for all of evolution.

Once you grasp that idea, it becomes much less useful to use the term "selfish gene" as a metaphor for anything, even natural selection and adaptation. That's why I think we should stop using the term "selfish gene."


46 comments :

  1. I have always thought that only valid use of the word "selfish" is in the context of the "selfish genome", It's still not an entirely accurate metaphor for evolution but it captures a lot more of the complexities...

    ReplyDelete
    Replies
    1. I don't think so. The genome goes into a scale where the explanatory power of the selfish gene fades off. Selfish DNA is an offspring of the selfish gene in the sense that it's DNA that succeeds out of being "catchy," rather than for being useful to the host. Selfish DNA would contradict a selfish genome unless you have some idea that solves this apparent contradiction.

      Delete
    2. (The contradiction being that selfish DNA propagation serves the selfish DNA, but does not serve the genome.)

      Delete
    3. You can have the "selifsh gene" if genes were inherited as individual units. But they're not. Even in Oxytricha and the likes, that's only the case in the somatic genome

      Delete
    4. Viral insertions, transposons, recombination. I think there's some stuff going on that makes the genome an unstable unit of selection for the word "selfish" to be useful and focus our attention.

      Delete
    5. This is what I wrote:

      It's still not an entirely accurate metaphor for evolution but it captures a lot more of the complexities...

      You can look at transposons and viruses as genomes themselves (even if embedded within other genomes) - a transposon is more than a gene...

      Delete
    6. I think I'm the one mistaken because we are talking about two very different things. I am talking about a metaphor helping us focus on some particular as a unit of selection, while you are talking about a metaphor for evolution. I still think that selfish genome would not be good as either, but I started complaining without reading carefully enough.

      Sorry.

      Delete
    7. This conversation seems to be implying, on both sides, that selfish genes would not cooperate with other genes. Are you both envisioning a genome red in tooth and claw?

      Delete
    8. Negative EntropyThursday, March 13, 2014 9:24:00 PM
      I think I'm the one mistaken because we are talking about two very different things. I am talking about a metaphor helping us focus on some particular as a unit of selection, while you are talking about a metaphor for evolution.


      Yes, but:

      1) The unit of selection is not really the individual gene because there is this thing called linkage disequillibrium.

      2) Some genomes contain a lot more than genes. What about mutations outside of genes? There is this famous Shh enhancer that's 1Mb away from the Shh gene and located within the intron of another gene, LMBR1 - how do we classify mutations in it within the "selfish gene" framework?

      Delete
    9. how do we classify mutations in it within the "selfish gene" framework?

      It's a slightly different 'gene concept' in use. Some bits of DNA enhance their own survival by virtue of their effect on other bits of DNA. The fact that promoter and 'gene' may sit some map distance apart isn't an issue. All DNA must be at some remove from its phenotypic effect. Many interacting systems sit on multiple chromosomes.

      Delete
    10. Hi Georgi,

      1. Linkage disequilibrium could be described as a "you scratch my back, I scratch yours" situation, for example.

      2. As Allan said.

      Delete
    11. Hi John,

      My part of the conversation was not about uncooperative genes, was more about the wide range of things that can be explained via the selfish gene metaphor. Red-and-claw metaphors would not be able to explain too much. Maybe a few things. What I like about the selfish genes is that this metaphor helps explain more things than other metaphors. Even gene cooperation.

      Delete
  2. Larry,

    I'm not sure what is going on with you, or who you are trying to impress, but if you continue like that, I'm pretty sure some of your "friends" are going to think you are leaning toward creationism... There is nothing wrong with that, but you are going to become the object of satan's targets.....hiihihi

    ReplyDelete
  3. Quest demonstrates once again that she/he/it has no understanding of anything Larry posts on his blog.

    ReplyDelete
    Replies
    1. What else is knew. As Ernst Mayr put it in his book, What Evolution Is, there are 5 elements in evolution. The there is no controversy on the first 4, namely old earth, extinction, appearance of new genomes, and common descent. The main controversy is over the mechanism, particularly natural selection vs random genetic drift. Now quite clearly, Prof. Coyne and Prof. Moran are not in complete agreement on this point but their disagreement is over the relative importance of the two mechanisms. It's a little like the disagreement between Huygens and Newton over whether light was a particle or a wave. Turns out they were both 1/2 right as it is both.

      Delete
  4. Larry,

    But the selfish genes was always presented to me as a metaphor putting the emphasis of natural selection on the gene rather than on the organism. I never heard of it as being a metaphor for the whole of evolution. I would say that to stop using a useful metaphor because some use it in the wrong context would be a huge mistake. What am I missing?

    ReplyDelete
    Replies
    1. P.S. I read the book too. And I don;t remember reading that the selfish gene was intended to explain the whole of evolution, or to be a metaphor for the whole of evolution. I might have missed that part, but if so, my point stands that as metaphor for the focus of natural selection it works quite well.

      Delete
    2. Having read the book more than once (though I admit it's been a few years since I last read it), I always recognized (as it was clarified several times within the book), that the selfish-gene term was meant as a metaphor for natural selection, but not evolution as a whole.

      Delete
    3. Thanks Saul, that was my recollection too.

      Delete
    4. I agree with this, where it does apply, it is a very good metaphor, just because some people are misapplying it doesn't make it bad in situations where it does apply perfectly well.

      Delete
    5. @William Bell,

      I have nothing against metaphors if they illuminate understanding. However, the more I think about it the less I'm convinced that "selfish gene" is useful. Could you give me a specific example of how the term could be correctly used today in order to enhance our understanding of evolution?

      What I'm looking for is a few sentences that describe an example of evolution where the you use the term "selfish gene" in an appropriate context.

      Delete
  5. I think the selfish gene is a useful metaphor for teaching that evolution via natural selection is about what is good for the genes, not what is good for the individuals. It does a great job of correcting this common mistake.

    And then we go on to other issues, like neutral changes other and random changes. The selfish gene metaphor doesn't help there, but so what?

    ReplyDelete
    Replies
    1. I agree. I see the metaphor working for the layperson (or even a beginning scientist - in terms of refocusing the attention of what natural selection means, or at least on the level that it operates on) - but that if a person learns more, and specializes in the fields of genetics or evolution, as they encounter other "forces" driving evolution - that the selfish-gene metaphor begins to break down. But as you say, "so what?". If a person as at the point that they are trying to learn about the other forces of evolution, then, they'll be able to, or should be able to, know when to put aside the metaphor, and focus on the topic at hand.

      Delete
    2. Aside from the historical order in which ideas arose, the issue is that natural selection has always been the easiest aspect of evolution to describe in a way that is likely to be intuitively grasped by people. I think it does have value (esp for lay audiences) but I suppose one of the problems is that for many people it became the whole story of evolution which is what Larry is trying to combat.

      Delete
  6. I read a report over at Discovery Institute that Larry had lost his scientific employment and is currently working McDonalds. Does anyone know what location, I'd love to buy him a coffee.

    ReplyDelete
  7. Is there any genetic evidence using genes etc to back up evolutionary biology??
    I mean evidence a creationist can sink his teeth into. not extrapolation musings?
    In thirty words or less please.

    ReplyDelete
    Replies
    1. http://www.nature.com/nature/journal/v461/n7268/full/nature08480.html

      Delete
    2. Larry's summary.
      http://sandwalk.blogspot.de/2013/12/lenskis-long-term-evolution-experiment.html

      Delete
    3. I mean evidence a creationist can sink his teeth into.

      That's a contradiction in itself.

      Delete
    4. You wasted 32 words to ask this question.

      Delete
    5. How do you expect people to answer a question is less words than it took to ask?

      Here are some search terms if you're looking to actually learn something:

      - Human Chromosome 2
      - Endogenous Retroviruses
      - Nested Hierarchies

      Delete
  8. In thirty words or less please.

    And that, ladies and gentlemen, is the kind of commitment to detailed research that gives creationism its well-earned reputation.

    ReplyDelete
  9. It's a perfectly respectable metaphor. Some of the distaste for it among biologists may come from the use of the term 'gene', which they will recognise in different ways depending on whether they are molecular biologists, geneticists or whatever.

    The Selfish Linkage Unit (Where Self-Interest Does Not Exclude Co-Operation And Can Only Be Pursued Where There Is A Copying Differential To Be Exploited). Snappy!

    ReplyDelete
    Replies
    1. Neat. Though I think that the distaste comes more from the word "selfish" than anything else. People tend to have an allergic reaction to this word.

      Delete
  10. Interestingly ... (I hesitate, because I know how well critique of experts goes down in these here parts) ... in view of the discussions on sex, ISTM that a significant amount of the confusion stems from a misunderstanding of the SG concept and effect of the sex/asex boundary on levels of selection. I don't mean everyone, so individuals need not bristle. But any treatment that begins 'Consider a gene for asexuality' needs to be greeted with a cold, hard stare.

    ReplyDelete
  11. Re: Once you grasp that idea, it becomes much less useful to use the term "selfish gene" as a metaphor for anything, even natural selection and adaptation.

    How so? That certainly isn't a very coherent argument. That the idea of a "selfish gene" might be even more useful to a confused person is a weak argument against its current utility.

    ReplyDelete
  12. they come away from reading the book by thinking that all genes are selfish

    Only by skipping most of the words in it!

    ReplyDelete
    Replies
    1. Here are some of the words you must not skip.

      Evolution works by natural selection, and natural selection means the differential survival of the "fittest." (p. 7)

      ... I must argue for my belief that the best way to look at evolution is in term of selection occurring at the lowest level of all. (p. 11)

      Darwin's theory of evolution by natural selection is satisfying because because it shows us a way in which simplicity could change into complexity, how unordered atoms could group themselves into ever more complex patterns until the ended up manufacturing people. Darwin provides a solution, the only feasible one so far suggested, the the deep problem of our existence. (p. 12)

      A gene is defied as any portion of chromosomal material that last for enough generations to serve as a unit of natural selection. (p. 28 - this implies that all genes are subject to natural selection)

      There's lots more where that came from. The problem isn't that Dawkins fails to associate selfish genes with natural selection—he does that. The problem is that he fails to differentiate between natural selection and evolution. You'd be hard pressed to find any passage in the book where he says that his description of selfish genes only applies to that part of evolution that's due to natural selection.

      Can anyone find a passage that talks about evolution in terms of genes that aren't selfish?

      Delete
    2. My reading - it's been a while - is that genes are intrinsically 'selfish' (they will increase in the population on their own account if they can find levers to do so) and yet we see a substantial amount of altruistic, co-operative and sociable behaviours at several levels. So the book is (to me) about setting up and resolving that apparent paradox.

      Certainly genes only take on the character of being self-interested when they exploit a mechanistic opportunity to increase, and evolution is not solely conducted by genes.operating in 'selfish' mode. The book is light on things that are somewhat peripheral to its main theme: modes of selection.

      Delete
    3. I don't think anyone is talking about "selfishness" as a metaphor for all of evolution. Not Dawkins. Not Coyne. It's about selection, period. And it seems to me that it's a fine metaphor. When genes are displaying "altruistic, co-operative and sociable behaviours", they're being just as selfish as when they compete; cooperation is one way to increase in frequency, right? Drift is the main "non-selfish" process.

      Delete
    4. @John, @Allan,

      Imagine that a mutation creates a new allele that has a 1% fitness advantage over all other alleles in the population. I assume you are both happy to refer to that new allele as a "selfish gene." Right?

      Now imagine that the new allele is lost before it becomes fixed. That's what will happen approximately 98% of the time. How does the metaphor work in that case? Do we say that the selfish gene was outcompeted by other selfish genes that were less fit?

      Imagine that you discover a gene for coat color that has become fixed in a species of mouse. Do you refer to that gene as a selfish gene in the absense of evidence that the coat color is adaptive?

      The problem with the selfish gene metaphor is that there are very few times you can profitably use it once you understand that natural selection is not the most common mechanism of evolution. I bet neither of you can come up with a scenario where the term "selfish gene" is really useful and accurate.

      Delete
    5. I don't see the problem. You keep introducing drift into it, when drift isn't what the metaphor is about. I think everyone is willing to stipulate that "the selfish gene" is about selection and nothing else. Anyway, your argument here seems to be against selection in general, not against a term we use to describe it. Unless you want to claim that selection doesn't happen, I would say that any scenario we both agree involves selection is handily described by "selfish gene".

      Delete
    6. Metaphors are a way of getting across to people the fundamentals. We might talk of air molecules 'trying' to move from high pressure to low, when all they are really doing is bouncing about, with an excess of trajectories in one direction. When you do statistical mechanics, you drop the metaphor. The fact that some molecules actually go the wrong way, against expectation, is a detail that cautions against over-extension. (They weren't trying hard enough!).

      Dawkins isn't just talking to the general public, of course. He's also advocating an understanding of the evolutionary allele by the general biologist. This can be over-emphasised, indeed. Not every stretch of DNA is there because it is 'selfish' - because a positive differential got it there. Once fixed, it is not an easy matter to untangle the reason why.

      The same would apply if we talked in the metaphor of 'selection pressure', or 'hill-climbing', although the latter does allow for vast swathes of flatland, if that's what you want to emphasise.

      But the dynamics of interaction of genes within and between genomes (in the same nucleus, cell, organism or ecosphere) has had a massive influence on the forms of sexual eukaryotes, and seeing them as 'selfish' helps to understand that dynamic. 'Selfish' is not really a mere, or even very precise, synonym for Natural Selection at the organism level. It's about doing what you can, if you can, to get copied.

      Delete
  13. Textbooks do without the metaphor of the selfish gene - indicating the metaphor might be as confusing as illuminating. Futuyma Evolutionary Biology 1998, still the standard to measure textbooks by, has a discussion of the selfish gene on pp 353-354. has an in introductory discussion, at the least indicating the selfish gene point of view has never been generally espouse.
    For any quantitative geneticist thinking in selection differential, selection gradient and selection response, selfish genes are an unnecessary abstraction.

    ReplyDelete
  14. If anyone would want an example of the sort of disaster that can come about if 'The Selfish Gene' is superficially read, see:
    http://dare.ubvu.vu.nl/bitstream/handle/1871/39335/dissertation.pdf?sequence=1

    ReplyDelete
  15. In his review of the Aeon article, Coyne does find things to approve in the offerings of the research scientists. He cites Robert Sapolsky's success in narrowing in on a "huge flaw" in Dobbs' argument.

    His approving quotation includes "When there’s a mutational change in the DNA sequence coding for a gene, and that new variant gets selected for, evolution happens. But critically, when there’s a mutational change in the DNA sequence coding for a regulatory element, and that new variant is selected for, evolution also happens."

    (Let's skip over whether it selection for "regulatory" genes is uncomfortably analogous to careless arguments for group selection of organisms, except for "genes." And whether it's also vague on whether the selection is for advantageous or potentially advantageous traits.)

    The thing is, I thought evolution occurred even when there was neutral mutations, or nearly-neutral mutation. This seems to be a gross error in evolutionary theory committed by a professional scientist even while taking to task a science journalist for committing gross errors in evolutionary theory. Coyne adds ".. all adaptive changes in the genome, be they in regulatory or coding elements, involve natural selection. And natural selection is precisely what the idea of the “selfish gene” is meant to encapsulate." The great contribution of the selfish gene metaphor in understanding evolution is to emphasize natural selection.

    Coyne does a good job demolishing Dobb, but what he says to demolish Dobb directly refutes the point being made in the post?

    Suppose we say: Every organism has genes, which are selfish and want to win, metaphorically speaking. These selfish strivings for victory in the race for survival of the fittest gene leaves only winners, necessarily the fittest, i.e., most adaptive genes. How is this a distortion of the selfish gene metaphor?

    Coyne's main assertion really is that all critics of the selfish gene are sensationalizing evidence against genetic determinism and (I think it's sometimes called) panselectionism as evidence against the selfish gene, which is natural selection solely for personal gain. His only evidence against this is a bald assertion that Dawkins denied this. Since people can be diplomatic in presenting their views, I think the only effective rebuttal would be to rehearse the well-known ways that the selfish gene metaphor is entirely independent of genetic determinism. Perhaps the problem is that they're not so well-known? Maybe, maybe not. But epistasis, methylation, gene expression, genetic assimilation, epigenesis, the whole rogues' gallery of failed assassins who aimed at the selfish gene, are all evidence against genetic determinism. (Neutral and nearly-neutral evolution are irrelevant apparently.) In this context, I disagree Coyne's vituperation of motive is effective. I can't even feel that it's a real contribution to discussion.



    ReplyDelete