Do the northern Chinese have Scythian ancestors?

There was some question regarding possible Scythian admixture into the early Zhou below. This is possible because of the Zhou dynasty, arguably the foundational one of Chinese imperial culture (the Shang would have been alien to Han dynasty Chinese, but the Zhou far less so), may have had interactions with Indo-European peoples to their north and west. This has historical precedent as the Tang dynasty emerged from the same milieu 1,500 years later, albeit the Tang were descended from a Turkic tribe, not Indo-Europeans.

I looked at some of my samples and divided the Han into a northern and southern cluster based on their position on a cline (removing the majority in between). I also added Lithuanians, Sardinians, Uyghurs, Mongols, and Yakut. As you can see on the PCA the Mongols are two clusters, so I divided them between Mongol and Mongol2.

Running ADMIXTURE after some outlier removal you see that the northern Han are distinct because they share ancestry from the Yakut modal cluster. In contrast, the Mongols and Uyghurs have ancestry from the Lithuanian modal cluster. Uygurs also have quite a bit of ancestry from the Druze modal cluster, which is West Asian. Also notice that the Mongol2 cluster, which shares more ancestry with the Yakuts also has more Lithuanian modal cluster ancestry. Two of the Mongol2 individuals are labeled as Khalkha.

Using some of the Sarmatian/Scythian samples from David Reich’s lab, I ran ADMIXTURE again. These ancient samples need to be interpreted with caution, as usual. But notice again that the northern Han obtain their minor ancestry from the Yakut. The Iron Age nomadic modal ancestry is found at low levels in the Mongols and Uygurs. I think this is a real effect. The presence of Alans with the hordes of the Mongol Empire is well attested, though the admixture is almost certainly earlier.

I ran some three population tests. This is what was notable.

  1. Han_N looks like it is mixed somewhat with Yakut
  2. Mongol has gene flow from Mongol2
  3. Mongol2 has gene flow from Lithuanians and Iron Age nomads

I literally spend an hour on this assembling the data. But I think the easiest conclusion to draw is that the “West Eurasian” shift in modern Chinese (north) is probably mediated through Turkic people.

5 thoughts on “Do the northern Chinese have Scythian ancestors?

  1. I wonder if there is any good way to search for introgressed autosomal haplotypes with potential selective fitness benefits in the sample. The search should be hard, but picking the right haplotypes could be.

    Culture and selective fitness enhancing genes can be transmitted to a large number of people from a very small number source individuals, resulting in a great societal impact, particularly if the admixture involved at least moderately high prestige local partners. A dozen or two Bronze Age West Eurasians could still leave a considerable cultural impact and selective fitness enhancing gene impact on the already quite large population of the Zhou that would be invisible in admixture measurements or ancestry informative uniparental markers.

  2. ” albeit the Tang were descended from a Turkic tribe, not Indo-Europeans”

    You are talking specifically about Emperor Gaozu’s maternal ancestry right? His paternal ancestry was ethnic Chinese (Han), related to the Western Liang (西凉). He did have ancestors that worked for the Western Wei (西魏), which were a clan of Xianbei. However, despite being given a Xianbei surname, it appears he was Han, but I’m not sure who he married. It could be Gaozu was more Turkic in ancestry than Han, despite his paternal ancestry.

  3. To my shame, my eyes glaze over a little when I see your charts and I just focus on the text.

    I used to have the same experience when I transitioned from being a lawyer (text focused) to a banker, when my mind would instinctively ignore the numbers on a page. If only out of self-preservation for my employment in finance, I have trained myself out of that habit.

    Is there an online guide that will give me more information on how to properly read your charts? If it’s just a case of “just study them closer, because it’s self evident”, please let me know and I’ll apply myself harder.


Comments are closed.