The remarkable landscape of polymorphism sharing in modern humans

Once again, I used the Complete Genomics data to study the problem of polymorphism sharing in different human populations. There are different sample sizes for the various populations, so I took random samples of 4 individuals/population. I extracted individuals from CEU, TSI, CHB, JPT, and YRI for chromosome 1, removing all polymorphisms if there were any missing alleles and with the --remove-indels flag of vcftools.

Here is the Venn diagram, showing how many sites were polymorphic in all populations, in only one of them, or in any combination thereof:

The total number of polymorphic sites at different populations are:


 CEU 346,417
 TSI 352,039
 CHB 326,398
 JPT 325,071
 YRI 498,536

This confirms the oft-mentioned observations that Africans are more genetically diverse than other populations. This is also confirmed by the existence of 222,255 polymorphisms that are unique to the YRI population.

A few years ago, the pattern of polymorphism sharing would be interpreted in some way similar to this:

In the above tree model, the story of Homo sapiens proceeds in isolation, from our earliest ancestors, through a series of splits (the first of which is "Out-of-Africa"), all the way to modern populations. Such a story is pretty much necessary to explain the super-homogeneity of Eurasians vs. the super-diversity of Africans. If the human story is self-contained, then the only plausible explanation is the Great Out-of-Africa Bottleneck.

But, the human story is not self-contained. Multiple lines of evidence suggest that several demes within the species Homo heidelbergensis, and perhaps, even H. erectus, co-existed during the Middle Paleolithic in the Old World with people who were genetically and anatomically most similar to living humans. A new possibility emerges: that the heterogeneity of modern humans was created in part by admixture between such divergent populations.

It is a worthwhile exercise to study the behavior of SNP subsets vis a vis the two  archaic hominins (Neandertal and Vindija) whose genomes we possess. In a recent experiment, I did just that. The finding that population-specific polymorphism (e.g., variable sites that are specific to Europeans, East Asians, or Africans) show an increased evidence of archaic admixture seems to suggest that the roots of regional populations of Homo sapiens may lie in very divergent regional Middle Paleolithic Homo populations interacting with a successful deme within anatomical Homo sapiens that emerged prior to 100 thousand years ago and concluded its conquest of the Old World about 50-60 thousand years later.

If this idea is correct, then it may turn out that the remarkable landscape of polymorphism sharing (and non-sharing) between modern human populations was not only shaped by what Homo sapiens did as he followed his own story. Thanks to technologies invented by some of his descendants, we are beginning to piece together the remarkable panorama of the prehistory of our species. But the cadre will not be complete until we find a place in it for those mysterious Others, fragments of whose DNA persist in our cells.