There's a write-up in The New York Times on a new paper to come out in PNAS soon on the relationship of disease variants to human demographic history over the past few hundred thousand years. I'll probably review the paper when it comes out, but since it is the holidays & all here in the states it might be delayed a bit more than typical. Roots of Disease Found to Vary by Continent is the article. Here's the most important bit:
The Stanford study also sheds light on major aspects of human population history, like the time at which the first modern humans emigrated from Africa. Archaeologists believe it was about 50,000 years ago, since no modern human remains older than this have yet been found outside of Africa, but geneticists have long favored much earlier dates. Dr. Gravel and Dr. Bustamante now calculate that 51,000 years ago, give or take several thousand years, is the date best supported by genetic data, bringing the geneticists’ date into alignment with the archaeologists’ favored time for the exit from Africa. The common variations in the human genome were mostly present in the ancestral human population in Africa and have been inherited by all the descendant populations around the world. The rare variants occurred more recently. “Most of the common variants hark back to pre-Out of Africa,” Dr. Bustamante said. “Most of the rare variants come after the Neolithic revolution.” This is the event that marked the beginning of agriculture about 10,000 years ago and led to significant increases in the size of human populations.
On the first part, I'll have to wait on the paper. I've kind of become more and more skeptical at taking these estimates at face value. Though the point about the lack of archaeological finds is a major alternative confirmation of the result though which should make us amenable to its plausibility. When it comes to rare variants being different across populations, the logic of this should be pretty obvious. Imagine for example if natural selection results in the rise to higher frequency of a range of rare variants from novel mutations which happen to be favored in the Neolithic environment. This is going to differentiate groups as far as the Neolithic events occurred to different populations, since they'll draw from a different random set of mutations. One can make the same sort of case for demographic events such as population bottlenecks, which could reshape the pattern of population substructure in a historically contingent manner and groups are subject to fission over time.