Thursday, August 11, 2005

Western barbarian + Han women = Hui   posted by Razib @ 8/11/2005 02:19:00 AM

My previous post dealt with the Hui of China from a purely historical angle. Out of curiosity I decided to see if recent archeogenetic work had been done on them (my previous literature search from about a year ago came up with little of great interest or specificity). To my surprise I found a paper surprisingly apropos, Different matrilineal contributions to genetic structure of ethnic groups in the silk road region in china. Here is the relevant part:

The prevalent presence of the eastern Eurasian-specific mtDNA haplogroups (93.3%) in the Hui compared with other Muslim samples, if not of recent gene flow from the Han people, would be attributed to the historical Han women contribution. The high frequencies of the south-prevalent haplogroups...the presence of an M7b3 type (Hui98), and other matched types across China in the Hui samples added further tallies to this suggestion and mirrored the trace of the earlier inhabitants along the southeast coast of China...In addition, neither Hui nor Han samples harbored any H and U types that were highly present in other populations in Central Asia. At this point, the matrilineal gene pool of the Hui was shaped by their historically encouraged intermarriage with the Han people, and the religion did not have much influence on it.

You can see the principle component chart. It is important to note that mtDNA is maternally transmitted, so this isn't giving us a "full genome" picture, but, nevertheless, from an inspection of Hui phenotype one can infer that they are likely a predominantly East Asian people with a residual West Asian admixture (even if their Y lineages are purely West Asian, generation upon generation of admixture with the Han via conversion of local women to Islam would have shifted the autosomal genome, which determines the general character of their genetic sequence, ergo, phenotype). The genealogies of the Hui elite point to the migration of male ancestors from the West, but say little about women, so it seems likely that the introgression of Han genes that resulted in a East Asian phenotype occurred via females. The PC chart as well as the haplotype tables shows that the mtDNA of the Han and Hui are very close in their profiles, with only a minimal level of West Asian admixture on the part of the latter. Additionally, the evidence of "southeastern" mtDNA is a testament to the relatively cosmopolitan and peripatetic history of the Hui in China, not only were their male ancestors brought in as administrators, merchants and court astronomers, but they were shifted to various parts of the Empire as needs warranted (the Hui, unlike the Han, had no qualms about the martial vocation and so were often used as soldiers, and as I noted before, "enforcers" against regional minorities, including Muslim ones). There were large Hui communities in the northwest, in Yunnan and in the southeast. Southeastern mtDNA might simply be evidence of the Muslim marriage network, which spanned the expanse of China, and the early period of admixture with southern Chinese Han women (Yunnan, in the southwest, was a non-Han province when it was first settled by the Muslims, so in that case there is likely a great deal of Dai [Thai] ancestry in the Hui of Yunnan).

There is also one point in the text that isn't reported in the abstract that was of interest to me: "Intriguingly, the haplogroups that were mainly found in southern Pakistan, India, the Near East/Caucasus region, the Iranian plateau, and the Arabian Peninsula, such as HV2, R2, and U7...were only present in Uygur and Uzbek, which harbored an approximately equal amount of western Eurasian types." The authors note that other groups, like the Kazakh, do not harbor these markers. That is not surprising, the Kazakh were relatively late arrivals to the southern steppe and were nomads who had little truck with civilization. In contrast, the Uyghers/Uighers (they use both spellings, and sometimes distinguish between the two groups and sometimes do not1) are traditionally assumed to be a ancestrally composite population of predominantly Turkish origin (their form of Turkish is nearly intelligible with almost every other language of that family, which of course facilitates pan-Turkish nationalism) superimposed upon a substrate of Tocharians and Indo-Iranians (both Indo-Europeans groups of disparate origin who populated the northern and southern edges of the Tarim basin prior to 1000). West Asian mtDNA could have entered into the Uyghers through the long established trade networks along the Silk Road, or, I suspect more likely, the West Asian lineages were brought by the expanding Tocharian and Indo-Iranian peoples thousands of years ago.

In any case, genetic data helps us solidify what we already know, which isn't a bad thing. Historical conjecture is usually jelly-like in its firmness, but over time it can be bounded and scaffolded so that it is constrained to a particular shape.

1 - There is a peculiar backstory to terms like "Uigher" and "Uygher," and there are multiple groups who claim the appellation, with various degrees of historical legitimacy based on their relationship to the 8th century Uigher Empire.