Libidinous Neandertal men and the women who loved them

I keep seeing people, who really ought to know better, saying that the new Neandertal genome results show that the gene flow must have been Neandertal men mating with modern human women, and not the other way around.

You see, they're fixated on the idea that the mtDNA showed no signs that the Neandertal clade survived into the present-day population. That result really convinced some people that interbreeding was impossible. They're flummoxed that some of the rest of the genome has significant signs of intermixture. It's like their world is spinning out of control. I'm not naming any names, but if you've followed much of the press around the Neandertal genome, you've probably seen this suggestion.

I don't know why it hasn't occurred to them that the Neandertal mtDNA type was probably lost because of natural selection.

To avoid raising the awful specter of Darwin, they've been talking about weird mating restrictions. Well, I suppose that if you really have to find a way to get Neandertal nuclear genes into us, without bringing mtDNA along, a total lack of Neandertal women contributing genes is formally one way to get that.

I'd just like to see these people explain how exactly we managed not to get any Neandertal Y chromosomes, either.

Is it safe to talk about selection, now?

UPDATE (2010-05-11): A reader writes:

With regard to your latest blog post on lack of neanderthal mitochondrial and Y chromosome DNA in humans: yes, it's possible natural selection had a part. However, given that only a small proportion of our ancestors seem to have been neanderthals at the appropriate time, it strikes me that this is a case where drift could be the correct explanation - despite the fact that I'm usually not a big fan of drift as an explanation.

Much depends on the size of the ancestral population and the pace of population growth in the generations surrounding the pickup of Neandertal genes. Drift is less likely to eliminate alleles in a growing population, but it depends how many copies there were to begin with. The key questions -- where and when the population was growing -- are unlikely to be the same as assumed by the modeling that showed drift couldn't have eliminated the Neandertal mtDNA, as most assumed the location of contact would be Europe and the time would be late.

There were other deficiencies with the modeling, also. Here we've been working on a source-sink model as a possible demographic scenario for Pleistocene humans; that kind of metapopulation dynamic might easily explain allele losses without selection, and becomes more and more credible as we learn the variance of contribution of Neandertal-like alleles across the genome. It's a different world this week than last week.

These are all mathematically tricky answers, clever, but academic unless we have good matches to genome-wide variation. Meanwhile a very simple answer, easy to explain to anyone, lies fallow. Exceedingly curious.

I'd be happy to be proven wrong about the Y chromosome, by the way -- we don't really know that Neandertals didn't have a human-like type, although we do now that today's human population has an exceedingly recent coalescent time. Could be bad estimates of mutation rate. Maybe we'll have more surprises in store.