Pages

Arlin Stoltzfus on Sandwalk!

Arlin Stoltzfus is guest blogging on Larry Moran's Sandwalk on neutral evolution and the disconnect between mathematical/ molecular evolution and the verbal stories commonly told about it. He's also the Stoltzfus 1999 JME guy I keep citing every once in a while, sometimes to annoy people. Anyway, you should check it out, his first post is here: Introduction to "The Curious Disconnect"

3 comments:

  1. I shall watch with interest, but I remain skeptical about these revolutions. I'm no especial fan of 20th C pop genetics, with their "in an infinite population ... " and "evolution is change in allele frequency". Recombination load, for example, is a dubious piece of reasoning. But he does overstate the case. The central logic of Darwinian selection is sound - the alternative to "daubing mud on a baroque castle" would appear to be dispensing with something that is logically inescapable. Sure, the extent and interaction of subdivisions of the genome that behave in a Darwinian way is up for debate. So is the extent to which neutral segments can combine in novel and interesting ways. But ...

    I idly wonder how much of the perceived problem resides in peoples' separate conceptions of what other people think. I often find when I am talking about, say, selection, I am talking about something different to my opponent - I see it as DNA sequences having an impact (or not) on their own survival and copying, according to how the 'agents' of selection distinguish (or not) their phenotype, at whatever level(s) they have a phenotype. I don't expressly mean Darwinian gradual NS, or s<>0-mediated population dynamics, or adaptation per se. Now, maybe I'm too dumb, or too adaptationist, to get it, but I'm not sure what refinement is needed to that viewpoint on the basis of recent discovery. A new variant starts with a mutation? You're kidding! Mutations are not all equally probable? Say it ain't so! Mutations are discontinuous? I knew that biochemistry degree would come in handy!

    Good paper by Eugene Koonin here. http://nar.oxfordjournals.org/cgi/content/full/37/4/1011

    If course, I'd disagree with his thesis that Darwinism (TOL + gradualism + dominant NS) is holed below the waterline by genomics. Transposons, for example, are Darwinian entities, in their own little world.

    ReplyDelete
  2. It's not a revolution per se, but just a time to update the overall model of evolution, ESPECIALLY the way it's taught to outsiders (general public and undergrads). There are two problems that irk me the greatest:
    - focus on vertebrates/animals, which are a negligible subset of all the things the evolve (the vast overwhelming majority of which are prokaryotes, and the remainder almost entirely protistan, to an approximation). This vastly skews what you know about evolution, since each subgroup has its own special quirks that do not carry over to the overall generalisation, and the general theory should try its best to satisfy the 99%, not the 1%.

    - The weird idea that adaptive stories explain anything. Selection doesn't -drive- anything, it meerly enables things through varying strengths of constraint against features. Eg. complexity happens through neutral evolution and ratcheting, but is enabled and maintained by selection. Telling stories like "this is how trait X would be selected for" are not only wrong (selection only works against; 'positive' selection is an epiphenomenon of negative selection wherein the changing environment, incl competitors, renders the 'weaker' unsatisfactory) but utterly pointless: it only explains how a trait could be possible. We already know it's possible, because it's THERE. So to me, adaptationist stories are just meaningless.

    They could yield some clues to actual mechanisms, but only of used properly and with great caution. Seems like this has yet to catch on in some circles, especially among those far from studying mechanisms of evolution, instead focusing on large scale traits that are poorly defined and hardly possible to work with at the moment, like behaviour.


    Btw, Lynch and Stoltzfus seem to do proper popgen. I'm also skeptical of much of it (for one thing, there's an obsession with exclusively vertebrate features of it, which are completely irrelevant to me), but they make sense. Precisely BECAUSE they realise populations aren't infinite, nor is time infinite, which is kind of required for neutral approaches...

    I do like that Koonin paper; a few disagreements here and there, but overall a nice summary. Haven't torn into it in depth yet though ^^

    ReplyDelete
  3. Yes, I agree about the 1% bias (I call it diplocentricity). The bulk of population genetics is predicated on a linked set of pre-existing recombinant genomes. Then we notice that - hey! - there is a bunch of life for whom that model isn't a very good fit, so ... we say that what we really meant by "allele" includes a caveat on the extent to which DNA is sliced. Prokaryotes were just waiting for diploid recombination to come along so we could stop "allele" and "genome" from being synoyms at the level of DNA.

    Since, if you place bacteria in a chemostat, you can get competitive curves that are a mirror of allele spread in a random-mating recombinant genome, then you think you've got a unified model. Then you try and use that model to explain the evolution of eukaryotes or sex (which came first?)... many years of head-scratching ensue.

    That's my main interest, and I'm trying to get my thoughts down on paper. I find myself flummoxed by the apparent conviction of the field that the only valid answer will be a pop-genetic (and adaptive) one. "Look! Twofold Cost!" "We're f***ed!".

    I think people do forget that 'selection pressure', or 'selection for' are just as much metaphors as 'selfish gene'. And until you have at least two recombinant organisms in the world, there is not even a metaphorical vacuum creating an environment for 'selfish' DNA fragments to flow into, by whatever means are within their grasp.

    So yes, I think adaptative thinking can be misapplied, particularly when mechanism is routinely ignored. Genes can only do what they can do, and our bird's-eye-view of their best interests is not visible to them.

    ReplyDelete

Markup Key:
- <b>bold</b> = bold
- <i>italic</i> = italic
- <a href="http://www.fieldofscience.com/">FoS</a> = FoS