Over the years I have written and talked about the Hui people of China a fair amount. Who are the Hui? In short, they are a group of Muslims who live in China proper and have assimilated to the predominant cultural forms of the nation. The Hui of Yunnan speak in the dialect of Yunnan. The Hui of the north speak the dialect of their locality. And so forth.
But, due to their Muslim religion, they are obviously quite distinct culturally from the Chinese majority. They do not eat pork, and they dress somewhat differently. That being said, the cultural distance between the Hui and the Han is far smaller than that between the Uyghur and the Han. There are about as many Hui and Uyghur in China, so it is a misnomer when you see headlines of the form “Chinese Muslims” when referring to the Uyghur, when in fact the majority of Chinese Muslims are not Uyghur (there are other Muslim ethnic groups so that Uyghur overall are less than 50%).
In most of China proper the largest ethnic minority, and the Hui are viewed as ethnicity, are the Hui. In Central Asia the Hui are known as Dungans, and are viewed as Chinese first, not Muslims first (in China proper the Hui see themselves as Muslims, but the Dungan communities in post-Soviet Central Asia see themselves as Chinese). A disproportionate number of Hui, Dungans, were involved in the Manchu driven conquest and control of Xinjiang. The Hui are not physically distinct from the Han on the whole if they do not dress distinctively, and they speak and write in Chinese. But, their religion allows them to interface with Central Asian Turkic Muslims in a way that is more difficult than the Han. In other words, they are a “middle-man minority” in places like Xinjiang.
This brings us to the question of the genetic origins of the Hui. Are they simply Islamicized Han Chinese? The short answer is they are not simply Islamicized Han Chinese, but they are quite similar to the Han genetically. Over the years various methods give a proportion of about 10% “West Eurasian” and 90% “East Eurasian” for the Hui. This aligns with what you see physically in their faces. They look Chinese.
A new preprint adds just a little bit on the margin to this with more samples from Hui from Sichuan. Significant East Asian affinity of Chinese Hui genomic structure suggesting their predominant cultural diffusion model in the genetic formation process:
…Analyses of over 700K SNPs in 109 western Chinese individuals (49 Sichuan Hui and 60 geographically close Nanchong Han) together with the available ancient and modern Eurasians allowed us to fully explore the genomic makeup and origin of Huis and neighboring Hans. The results of the traditional and formal admixture-statistics (PCA, ADMIXTURE, and allele-sharing-based f-statistics) illuminated a strong genomic affinity between Sichuan Hui and Neolithic-to-modern Northern East Asians, which suggested massive gene influx from East Asian into Sichuan Hui people. Three-way admixture models in the qpWave/qpAdm analyses further revealed a small stream of gene influx from western Eurasian related to French or Andronovo into these Hui people, which was further directly confirmed via the admixture event from the temporally different western sources to Hui people in the qpGraph-based phylogenetic model, suggesting the key role of cultural diffusion model in the genetic formation of the modern East Asian Hui. ALDER-based admixture date estimation showed that this observed western Eurasian admixture signal was introduced into East Asian Hui during the historic periods, concordant with the extensive western-eastern communication in the Silk Road and historically documented Huis migration history. Summarily, although significant cultural differentiation among Hui and their neighbors existed, our genomic analysis showed their strong affinity with modern and ancient Northern East Asians. Our results supported that modern Chinese Hui arose from the mixture of minor western Eurasian ancestry and predominantly East Asian ancestry.
In the preprint you’ll see an AdmixtureGraph which shows that the Hui of Sichuan can be modeled as 93% Sichuan Han and 7% “Scythian”. The model fits, but the history does not. Take a look at the admixture plot and it seems quite clear to me the most likely “donor” population are Central Asian Turks, not ancient Iranians. The issue is that the Iranians themselves are predominant contributors to the ancestry of many Turkic groups. These groups are about 50% West Eurasian, so the better model probably is that about 15% of the ancestry of the Hui of Sichuan derives from Muslims of Turkic origin.
In the preprint the authors used ALDER to date the admixture to 500 or 1000 years ago depending on the group. They note that these dates will pick up the last admixture, not earlier ones. I don’t really trust the specificity of the dates because I don’t trust the model of admixture. The most plausible scenario is the one that is presented as the most likely by historians. Large numbers of Muslims arrived during the Mongol Yuan dynasty to help them rule and exploit the local Han Chinese. There were always Muslim communities before, but they were periodically suppressed, exterminated, or assimilated (see what happened in Guangzhou during the Tang dynasty). In contrast, after the fall of the Yuan the Muslims of China rooted in these Turkic communities remained and assimilated.
The great Ming naval commander Zheng He was from the Muslim community of Yunnan. He was already quite assimilated to Chinese culture, practicing worship of native gods as well as Islam without perceiving a contradiction. He was a great-great-great-grandson of an emigre from Bukhara, albeit someone of Iranian background (“Tajik”). Many Hui clearly assimilated totally into being Han (some South Chinese lineages descend from Muslims, but are culturally Han in all ways).
But the flip side is that there was the assimilation of Han Chinese into Muslim identity as well. Though some men converted (a branch of the Kong clan, the descendants of Confucius, converted to Islam with the marriage of one of their members to a Muslim woman), most of the Han who became Hui were likely women. In the attached preprint every single Hui mtDNA lineage is East Asian. The vast majority of the Y chromosomes seem East Asian as well, but I have seen other papers where many Hui carry haplogroups R1a and J, which suggest descent in part from Iranian peoples. If the Hui are today 85% Han, and they have been in China 30 generations, then a 5% outmarriage rate per generation would suffice to allow for this outcome.
In sum, these preprints and papers give genetic evidence that the primary exogenous ancestry into the modern Hui is from Turkic Central Asians. This aligns with the scholars who argue that the Mongol conquest was the primary accelerant of the establishment of the Hui in China.
Some anecdotes. During my college and early work in China (80s), a lot of my colleagues with official ID cards indicating “Hui” ethnicity are eating pork, never attend Mosques, know nothing about Islam. These are Chinese version highway robbery of affirmative action for minorities. Only one friend is true Hui, who is seriously practicing Islam and has physical feature of an uyghur. But he refused to be recognized as uyghur and had some contempt for uyghur. However he believe his ancestor from asia minor during Mongol empire. His love interests were always pure ethnic Han girls. The relationships always failed due to girls refusal of conversion to islam. Finally he succumbed to the family pressure and married a Hui woman arranged by family. Now he is one of my super rich friends in Beijing. Though he graduated with medical degree, he never practiced medicine. He started a medical equipment company and was the first multimillionaire from my class. And nobody is more loyal to current China than him among my classmates. He is also bit of anti-american.
Funny, I only have few lifelong friends with one Han, one mongol, one one hui. We just have very similar non-dramatic personalities and mental compatibility. People with strong emotional display turn us off.
The “Hungary_Scythian” samples are basically European, with a wide variety within Europe (and might represent assimilated locals from different cultures in eastern Europe I assume), and not like the Sarmatian or more eastern Scythian steppe populations so your kind of doubling makes sense in the first place because a source related to them would be fully contemporary West Eurasian related, which their graph shows anyway I guess? Their particular population pick then doesn’t really make too much sense because despite the cultural label, it doesn’t really represent the mixed steppe populations of the post-BA period anyway and is not really different from picking e.g. “French” as they do elsewhere in the main text. They probably didn’t notice that.
In the samples contained in the Eurogenes Global25 the position of the Hui seems somewhat more shifted towards Central Asians, though a more direct comparison would be after the addition of the ones in the paper. Due to that greater shift, their closest populations there end up being some of the more East Eurasian derived Mongolic and Turkic speaking ones, unlike in the paper where they’re closer to some Sino-Tibetan speaking ones (including Han) from a quick look in the supplements. I’m not sure which country or province the Hui sample there is from but since it’s named “Dungan”, it might not be from China but rather the Central Asian Republics. Might be some variation depending on geography, with more western ones having some more western ancestry then.
Hui (referred to as Dungan on yfull) are also the carriers of a number of subclades of rare R1b-V1636, which is posited by the just-published PLOS paper on the Gjerrild burial cist to have been one of the original PIE markers, first seen in the Eneolithic North Caucasus Piedmont, later spreading to both Corded Ware in Europe and Kura-Araxes in the Near East.