Saturday, April 25, 2009

A Horse of a Different Color

 
John Hawks continues to post interesting articles on his blog and he continues his policy of not allowing comments. I want to ask a question about his latest posting [A horse of a different color] so I'm asking it here.

John is referring to a short paper on the evolution of coat colors in horses. Apparently, the analysis of DNA from ancient fossil horses reveals that most of them were bay in color. The chestnut coat color wasn't detected until about nine thousand years ago.

The authors of the paper claim there is no evidence for selection of coat color in horses prompting the following comment by John Hawks.
The pigment-altering mutations at these genes do not all show statistical signs of selection in contemporary samples of horses. But they aren't there in the ancient horses. That's the best evidence of selection you could possibly have. Message: tests of selection on contemporary samples are weak, particularly for loci with rare alleles or more than two alleles.
John, if I understand you correctly, you're saying that as long as an allele wasn't detectable in ancient populations but is detectable today then random genetic drift is ruled out.

That doesn't make sense to me. Perhaps you could explain? There must be more to your statement than that.


11 comments :

  1. The test of genetic drift used in the paper uses the time series of sampled individuals. Under selection that was constant and directional, the frequency of an allele over time is like a biased random walk, with the size of the random component depending on the effective population size. If a significant bias is detected, this test rejects the hypothesis of drift.

    But in a sample of a few individuals -- the rarer alleles in this case were found in only four or fewer archaeological specimens -- the sampling variance is much larger than either the stochastic effect of drift or the bias in frequencies caused by realistic levels of selection.

    In this case, we have three pieces of evidence that aren't used in the test. First, we know really well what the frequencies of these color variants are in living horses. Second, we know we failed to find them at all in pre-Bronze Age horses. That's much better evidence against drift, because the sampling variances are much lower. And third, the alleles all show the same pattern -- the non-black/bay allele always increases over time. Under drift, we would predict the opposite change some of the time.

    ReplyDelete
  2. Hmmm ... I see.

    So, it's much more complicated than just saying that because the alleles weren't present in the past then they must have been selected.

    Right?

    You could easily have a situation where an allele wasn't present in ancient populations but it is present today in, say, 10% of Koreans, and the explanation could be random genetic drift.

    Correct?

    ReplyDelete
  3. "the non-black/bay allele always increases over time. Under drift, we would predict the opposite change some of the time"

    Hahaha
    Basically, Hawks thinks that anything that is not being selected for will disappear. Hence, if we have a documented increase,it is in itself evidence of selection, since if it were drift, it would disappear. Clear now, Larry? hehehe

    ReplyDelete
  4. In any case, it is well-known that under conditions of domestication, color becomes more variable (you get all those patchy funny looking things instead of neat camouflaged-uniform coats)).
    I suspect relaxation of selection on the trait. There is nothing better to increase variability and evolution than to take selection away.

    ReplyDelete
  5. Under conditions of domestication, selection is enhanced. Humans craving novelty will selectively breed up any rare allele that can provide such novelty, including a funky new cool horse coat color.

    ReplyDelete
  6. "Under conditions of domestication, selection is enhanced"

    As a rule? Nope. Obviously some traits are artificially selected, but for many, selection is relaxed in comparison to wild conditions.

    Many artificially selected races have typical genetic diseases. Do you think this would be possible if all there is to domestication is increase of selection?

    Unfortunately some people are just bent on just seeing selection enhanced everywhere, anytime.

    ReplyDelete
  7. So, the diversity of dog breed phenotypes is the result of drift, right? All the human-meaningful phenotypes in domesticated species are under tremendous selective forces. Explain why horse coat color should be any different?

    ReplyDelete
  8. because i dont think it is under tremendous human selection. Other colors in horses are OK, too.Under wild conditions, much color coat variation can be selected against. Not so under domestication.
    This seems very hard for you to comprehend, but there IS relaxation of selective pressures alogn with domestication.

    ReplyDelete
  9. http://en.wikipedia.org/wiki/List_of_horse_breeds
    Domestication involves human-mediated selection that emphasizes, preserves and expands representation at the extremes of the phenotypic distribution. This means that when a new trait (eg coat color) shows up, humans cause this trait to expand in the population by selective breeding. A new phenotype that would take impossibly long to be fixed into a small percentage of the population by drift, if indeed it ever were to be fixed in the population, is rapidly preserved and expanded by selection. Drift is slow. Horse color coat change happened fast. It's selection.

    ReplyDelete
  10. anonymous says,

    Horse color coat change happened fast. It's selection.

    That's a reasonable argument but it's not the one we're discussing.

    ReplyDelete
  11. Oh I'm sorry. When you asked,

    John, if I understand you correctly, you're saying that as long as an allele wasn't detectable in ancient populations but is detectable today then random genetic drift is ruled out... That doesn't make sense to me. Perhaps you could explain? There must be more to your statement than that.From your question I thought you didn't understand that all rapid genotypic change in sizeable populations is the result of selection, and that the absence of the allele in ancient, yet sizeable horse populations is therefore prima face evidence for rapid change, and therefor for selection.

    In John's response to the question, I believe when he replies, "And third, the alleles all show the same pattern -- the non-black/bay allele always increases over time. Under drift, we would predict the opposite change some of the time." he is making the point that for random walk in allele frequencies caused by drift, the random aspect of the walk implies the frequency should not increase in a monotonically non-decreasing manner, as is (apparently) observed for the bay allele, but rather should show periods of allele frequency contraction from time to time.

    if that's not the question, can you clarify what the question actually is?

    ReplyDelete