There are 20 million Hui people in China. These are traditionally Chinese-speaking Muslims. Though they are found in every region of China (and in the Chinese Diaspora), they are concentrated in the northwest, in Gansu and Ningxia in particular. Their origins have always been curious. They speak the local Chinese dialect, and look mostly Chinese, but they are traditional Muslims. Are they purely the descendants of local converts? The Hui do not believe so, and some of them physically do look more West Eurasian.
Thanks to genetics we know the answer, Genetic Origins and Sex-Biased Admixture of the Huis:
To investigate whether there was a sex-biased admixture in the history of NXH, we compared the admixture results obtained from autosomes, X chromosome, mtDNA, and Y chromosome. We estimated the admixture proportion assuming two major ancestral components, that is, western and eastern (fig. 3 and supplementary tables S2 and S3, Supplementary Material online). The estimated genetic contribution of the western ancestry into NXH was 8.6% for autosomes, 5.9% for X chromosome, 3.6% for mtDNA, and 39.3% for Y chromosome, respectively. The results of Y chromosome and mtDNA were consistent with the previous studies (Yao et al. 2004; Wang et al. 2019; Xie et al. 2019). Additionally, though the difference in genetic contribution was small, there was a significant difference in admixture proportions between autosomes and X chromosome (Student’s t-test, P < 107−10−7). This pattern was consistent across different regions in Ningxia (fig. 3C). These results indicated that the admixture of NXH was sex biased to the combination of Eastern females and Western males.
The 39% western Y chromosomes is key. This is probably a floor for the fraction of originally Muslim lineages. The Hui likely had Iranian and Turkic precursors, and the latter would have had eastern Y chromosomes. But the point is that cultural continuity was maintained in paternal lineage systems, and over the generations intermarriage with local Han women resulted in 90% of the genome being replaced over time.
This is a common pattern in large parts of the world. Paternal cultural transmission is a thing.
Reading about the Huis, I realized that China has a province called Shanxi and right next to it, there is a province called Shaanxi.
Harry, in Mandarin, Shanxi and Shaanxi are 2 different characters that are both pronounced “shan” (say “san” while rolling your tongue like you will pronounce a “r”) but with different tones + first tone “xi” (pronounced “she”) meaning “West”. But as it’s not typical to write pinyin with tones in the West, especially on maps, one of them got stuck with “Shaanxi” as the transliteration.
Judging by Fig. S13, the Tajik-style DNA underwent just one major admission ca. 20 generations ago. In contrast, East Asian components admixed in a continuous way with poorly defined dates but perhaps some of it as long as 60 generations ago. Some of these estimates predate the admixture of the “Western” component, suggesting that the oldest, shortest “Eastern” segments were already contained in the “Western” components before the migration. Which sort of makes a historical sense.
Historical context is as follows: Ningxia population has been thoroughly exterminated by the Mongols in the 1200s and the remnants moved to China in the 1370s to escape the Mongol threat. But by the 1600s, Ningxia Hui artisans were combining Persian technologies with the motives of their Mongol customers. This hints at a migration (and integration in the Mongol world) in the 1400s-1500s.
On the other hand the Timurid Empire went into its phase of decay with military defeats and sackings of its towns in the 1400s and completely disintegrated by the early 1500s. A move by the refugees from this chaos or captives of its defeats kind of makes sense…