A
new study identified a small percentage (1 to 5% in average) of West Eurasian ancestry in the Khoisans, peaking at 14% in the
Nama people of southern Africa.
According to the
article in the New Scientist, this West Eurasian admixture resembles sequences from southern Europeans, including Sardinians, Italians and people from the Basque region. The study mentions that it could be southern European or Levantine in origin. They estimate that the introgression of these Mediterranean genes occurred some time between 900 and 1,800 years ago - a surprisingly recent, yet undocumented event. This would date from the late Roman, Byzantine or Moorish period. There is no mention of admixture from North Africa or the Arabian peninsula though.
I find it odd that an international team of geneticists from such institutions as Harvard Medical School, MIT, Max Planck Institute for Evolutionary Anthropology and CNRS do not have the means to distinguish between Iberian, Basque, Sardinian, North or South Italian, Greco-Anatolian or Levantine admixture. This is usually easily feasible with amateur programmes such as the Dodecad or Eurogenes. How is it possible that geneticists at high-tech institutions like that lag so much behind ?
The admixture into the Khoi San which is dated to 900 to 1800 years ago was not a movement directly into them from a Mediterranean like population. Rather, these alleles were transmitted to them indirectly via East Africans.
The initial signal of West Eurasian admixture which they are picking up was into East African populations, and is dated from 2700 to 3300 years ago.
From the paper...
We reconstruct the allele frequencies of the putative west Eurasian population in eastern Africa and show that this
population is a good proxy for the west Eurasian ancestry in southern Africa. The most parsimonious explanation for these findings is that west Eurasian ancestry entered southern Africa indirectly through eastern Africa...and that west Eurasian ancestry entered eastern Africa on average 2,700 to 3,300 years ago.
So they are not talking about a back to Africa migration during the Roman era. This is a migration around 1000 B.C. which was large enough that some East African populations are almost 50% West Eurasian. This explains the 23andme results for East Africans that showed essentially the same thing.
As for the source, the extensive modelling can be found in the supplement. Depending on the African population that was being targeted, one or another of the southern European populations provided the best fit, along with the Druze, but the differences were small. From the paper...the f3 statistics in nearly all eastern African populations involve a southern European or Levantine population as one reference.
As I mentioned above, the authors also reconstructed the allele frequencies of the west Eurasian population involved in the admixture in eastern Africa, and discovered that this hypothetical population is a better proxy for the West Eurasian admixture than any modern southern European or Levantine population.
From that, I think we can extrapolate that the population responsible for the west Eurasian admixture in Africa no longer exists. After taking into account, as the authors do, the archaeological evidence for migration from Arabia and the southern Levant into Africa, and the linguistic evidence as well, it would appear, as the authors posit, that the source of the migration was in those areas, but that the population that then existed in those areas is significantly different from the current, modern populations.
The prior work from the Reich lab has already shown that modern Near Eastern populations are not a good proxy for the EEF. Now, we find that they are not a good proxy for the source of the west Eurasian admixture in Africa. That suggests to me that the modern Levant and even Arabia have significantly changed even from 1000 B.C. to the present. Although some input from south Asia and perhaps ANE input as well may be contributing factors, it seems to me that the biggest change may have been the intrusion of SSA lineages in more modern times, although that is speculation at this point.