Ancient Mal’ta and Afontova Gora DNA Again

Now that Michael Balter’s news piece is out, I can retrieve my original post written in the wake of Eske Willerslev’s presentation at the Paleoamerican Odyssey conference in Santa Fe.

Eske Willerslev‘s paper was on ancient DNA from the Mal’ta and Afontova Gora sites in South Siberia (24,000 and 17,000 YBP, respectively). Apart from the fact that Eske was supposed to report on ancient DNA from the Anzick (Clovis) site in North America and not on Mal’ta in Siberia and that he’s somewhat of a controversial figure in genetic circles, his paper may be the major breakthrough in our understanding of the prehistory of Siberia in the past 20 years. Pending further testing, Willerslev assigned the Mal’ta sample to hg U (mtDNA) and hg R (Y-DNA). This is the earliest attestation of these haplogroups in Siberia suggesting that these West Eurasian lineages were much more widely distributed in Eurasia in pre-LGM times and common in Eastern Eurasia, too. The results are consistent with the finding of hg U2 in 30,000-year-old Kostenki remains in Central Russia and support the pattern whereby the earliest ancient DNA samples (Tianyuan Cave in China with hg B, Kostenki with hg U2 and now Mal’ta with hg U) have so far turned up members of mtDNA macrohaplogroup R. While the dates of all these samples are consistent with 50,000 YBP chronological frame proposed by geneticists for the divergence of mhg R, it’s still noteworthy that not only is the most downstream mtDNA macrohaplogroup also the most widely spread among human populations, but also that it seems to be more wide-spread in the past than now.

Another surprising finding reported by Willerslev is that the Mal’ta autosomal sample falls between West Eurasians and American Indians (and is not close to East Asians) in a PCA plot and that ADMIXTURE analysis decomposes the Mal’ta sample into a West Eurasian, South-Central Asian and American Indian components (again without the (North)East Asian component). This suggests that a) in pre-LGM times there were no East Asians but there were Amerindians and modern East Asians may be a product of post-LGM re-expansion and an overlay on top of the more ancient population strata with ties to the New World and West Eurasia. On Willerslev’s TreeMix diagram, there’s a gene flow vector connecting the Mal’ta sample with American Indians (which Willerslev, on totally spurious grounds, interpreted as suggesting a migration to the New World and not the other way around, as the finding of an Amerindian-like component in West European populations indicates) but, in mtDNA and Y-DNA terms, the connection is not direct as mtDNA hg U is a sister clade to hg B (frequent in the Americas) and Y-DNA hg R is a brother clade to hg Q (pervasive in the Americas).

Very importantly, the Mal’ta sample did not turn up any of the common American Indian mtDNA and Y-DNA markers tentatively suggesting that neither Y-DNA hg Q (found in such Siberian populations as Kets), nor mtDNA hgs A, B, C, D and X were present in South Siberia in pre-LGM times. From an out-of-America perspective, their current presence in Siberia may represent a back-migration from the Americas in post-LGM and early Holocene times. It’s also possible that some of the instances of Y-DNA hg R found in North America do not derive from post-Columbian admixture but in fact go back to pre-LGM times.

Update. A careful reader of Michael Balter’s piece would notice that there is an important difference between Willerslev’s and David Reich’s interpretations of the ancient DNA findings. While Willerslev’s believes that Amerindians are a mix of East Asians and West Eurasians, Reich is talking about a population that contributed genes to both Amerindians and West Eurasians. While Reich calls this population “Asian,” this population could just very well be called “American,” as there is nothing in the data that says that this ancestral population was in Asia and not further east in the Americas, because there is no relationship between Mal’ta and modern East Asians.