A few years ago there was a paper out on high altitude adaptation in Tibetans which posited a somewhat ludicrous historical demographic scenario: that the population of the Tibetan plateau diverged within the last ~3,000 years from the Han Chinese. Due to the recency of this event the authors argued that high altitude adaptations at these loci were the fastest sweeps on record so far. The problem is that this demographic model didn’t pass the smell test. China is not a region of the world where there is no documentation for that period (Chinese history starts ~1000 BC). To posit a Han migration into the Tibetan plateau, as the media at least was portraying it (though I think this was an easy step from the way the results were presented in terms of the semantics), is difficult to imagine because the ethnogenesis of the Han themselves had not truly occurred in a way that we would recognize today at that period. Today Jeong et al. in Nature Communications, Admixture facilitates genetic adaptations to high altitude in Tibet, may have gone some way to resolving some of the confusions that came out of that paper. The basic conclusion is that the high altitude adaptations have swept up in frequency among Tibetans after the admixture of a Sherpa-like population and a Han-like population on the order of 3,000 years ago.
It seems a plausible model (especially after you take in their qualifications which imply perhaps a somewhat older date for admixture), and might explain why you obtain a test statistic implying a recent divergence between the Tibetans and Han. I suspect that there isn’t a pure Sherpa-like population at this point, and the unsupervised admixture estimates are giving low fractions for Han-like admixture (the “high altitude” fraction showing up in populations as disparate as Gujaratis and Japanese is a tell that it’s an imperfect proxy). If you were a plant geneticist this sort of phenomenon wouldn’t be surprising at all, as one way that you can breed for better cultivars is mix together lineages and allow good phenotypes to introgress across genetic backgrounds. Also, the authors did not find any elevation of archaic admixture in the Sherpa-like element, though I wonder if they might miss a population “X” because they don’t have a reference genome (they looked for Neandertal and Denisovan). Intriguingly the divergence between the Sherpa-like element and the Han-like component is ancient, on the order of ~20-40,000 years ago. Though likely part of the broader family of East Eurasian populations, it seems that this Sherpa-like cluster was well diverged. Not very surprising due to the terrain, but perhaps it illustrates the power of agriculture to demographically and culturally transform societies, and reduce genetic distances over the last 10,000 years. The only reason that the Sherpa-like fraction persists in reasonable fractions is probably the difficulty of farmers and agro-pastoralists invading the terrain.
Citation: Nature Communications 5, Article number: 3281 doi:10.1038/ncomms4281
Note: Sherpa, and Tibetans, speak a language posited to have a distant relationship to Chinese. Ergo, the Sino-Tibetan language family. This implies that the original language of the highlanders was lost, though perhaps a substrate can be excavated.
Comments are closed.