One of the more interesting things about reading a book like The Making of Modern Japan, which is a relatively deep dive into the political and social history of Japan from 1600 on, is that it gives you an interesting window onto your own country. For example, I know from American history that there was a period of Southern dominance from the early 1800s down to the decades before the Civil War, followed by marginalization of the South in national politics. I had not known that for 50 years after the Meiji Restoration Japanese politics was dominated by a group of men who emerged from the Choshu and Satsuma domains, which were at the far west of Honshu and in southwest Kyushu respectively. These men were arguably the founders of modern Japan, and they were geographically and socially (i.e., they came from the samurai class by and large) very narrow in their origins. I had assumed that Japanese history, and the loci of power, had shifted gradually east in a unidirectional sense, from Kansai to the Kanto plain and what became Tokyo (Edo).
The reason I’m fascinated by the resurgence of Choshu and Satsuma is that this is the region of Japan where the Japanese as we understand them began. That is, the Yayoi culture, which spread across Honshu between 500 BC and 500 AD, from the south and west, to the north and east. Long present before them on the Japanese islands were the Jomon people, a relatively advanced hunter-gatherer culture. They are presumed to be related to the Ainu of Hokkaido. More intriguingly, in the second half of the first millennium the Tokohu region of northern Honshu was inhabited by a non-Japanese people termed the Emishi. Some of these people apparently accompanied a Japanese embassy to the court of Tang China, and their hirsute appearance was commented upon.
A fair amount of DNA evidence seems to suggest that the Ainu and Jomon are connected and that some of the ancestry of modern Japanese descend from the Jomon. Additionally, the Yayoi were probably rice farmers from Korea. I have some Korean data, so I ran TreeMix a bunch of times. As you can see usually the Japanese and Koreans are rather close. Japan is placed closer to the Yakut though. This makes sense if the Jomon were Siberian.
In May there was a publication that confirmed using Bayesian methods that the Japanese are a dual origin population. Instead of simple inspection and confirmation of a particular pattern, they tested an explicit set of models; replacement of Jomon by Yayoi, Jomon to Yayoi continuity, and Yayoi absorption of the Jomon. They found that the hybridization model was 29 to 63 times more likely than the replacement and continuity models. It’s great precision on what we already knew. And, if you look at their dates, the Yayoi begin to admix with the Jomon thousands of years before they show up in Japan archaeologically! They have some clever ways around this, but it seems that they’re trying to square a circle.
So today I was talking to Greg Cochran and he said someone should do some D-statistics on Ainu. Well, after I got off the phone with him I found this paper, Unique characteristics of the Ainu population in Northern Japan. They didn’t use D-statistics, but they employ and f3 and f4 ratio tests (see this paper for what all this means).
On the PC plot to the left you see a cline running from Chinese, to Koreans, to Japanese, to Ryukyuans (Okinowa and other islands). Then, on the larger PC (explains 3 times more proportion of the variance), you see a cline out toward the Ainu. In the TreeMix plots above there were gene flow edges from Southeast Asians and Yakuts into the Southern and Northern Chinese populations respectively. One could hypothesize that this reflects demographic absorption of indigenous groups by an expanding group of Han agriculturalists. The Yayoi were almost certainly part of this broader set of groups on the move. To the east you have the Japanese, but also the Han in all directions, and the various Southeast Asian groups as well.
The second PC plot is very interesting. It has a huge data set from all over Japan and a much thicker marker set (that’s why the Koreans aren’t on the plot, they didn’t overlap with it). What you see is a cline toward the Chinese, and then an axis defining the Ainu and Ryukyuan. The Ryukyuans have been found to have more Jomon-like ancestry in previous research, so it’s entirely unsurprisingly that they’re somewhat Ainu-shifted. But, observe that they are somewhat parallel to the Ainu. In all likelihood the Japan developed some population structure. The Ryukyuan and Ainu would be the northern and southern ends of the Jomon distribution, respectively. Additionally, there’s a subtle pattern in the results from the Japanese from the main islands: the Tohoku region. The authors observe that these samples, highlighted in red, seem skewed. And not toward the Hokkaido Ainu. In all likelihood that means that the Emshi were another branch of the broader Jomon population, distinct from the Ainu, and those peoples to the south. Additionally, there’s probably a “Jomon-cline” within Honshu, with more the further you go from the far west of Japan.
For the f4 ratio tests to determine admixture across a specified topology, they used the one to the left. It seems reasonable enough to me. By all phylogenetic results that I’ve seen the highest probability seems to be that the Ainu are a branch of Northeast Asian people if any. Roughly they find that the Jomon ancestry for mainland Japanese is a bit less than 20%, and for Ryukyuans it is a bit less than 30%. Reference population matters, so these are not gospel estimates. Probably they are a slight underestimate if I had to guess, as the Ainu are not perfect proxies for the Jomon.
Finally, they used rolloff to estimate admixture dates. For the mainland Japanese it was about 1,500 years ago. For the Ryukyuans it’s closer to 1,000 years ago. This dovetails well with the order in which the Yayoi occupied these regions. I wouldn’t, and they don’t, take the dates as gospel. They probably are underestimates (see The Spatial Mixing of Genomes in Secondary Contact Zones for some good arguments and results which indicate these methods underestimate time since admixture). But, they’re far closer to the archaeology than what the other paper above found, and don’t require a lot of special conditions to make the genetic inferences align with the rest of reality.
We also identified SNP loci that are differentiated between the Ainu and Mainland Japanese by using pairwise Fst values. The pairwise Fst values ranged from 0 to 0.8903, with a mean of 0.0407. The majority of the SNPs (approximately 400 000) have Fst values of less than 0.02. We picked 6413 SNPs that were within the top 1% and had Fst values higher than 0.36. The Fst values and annotations for these top 1% SNP are listed in Supplementary Table 1. Within those top 1% of SNP, some of them were found in genes reported to be associated with facial structure in Europeans2…and hair and tooth morphology in East Asians…The distribution of Fst values for SNPs in those genes are shown in Figure 4. Two out of five genes for facial morphology (PAX3 and COL17A1) contain highly differentiated SNPs, as with the hair/tooth morphology gene (EDAR). The results of gene annotation analysis on those top 1% SNPs showed enrichment for biological processes and cellular components involving collagen (Supplementary Figure 8).
The derived variant of EDAR is found in high frequencies in both East Asians and Amerindians, so it has to be somewhat old. But in places like Korea it’s very close to fixation. To me the high pairwise Fst on this gene is a tell that the Ainu did look very different when they were numerous and not admixed, even if they shared Pleistocene ancestry with the incoming Yayoi. That being said, when I looked at the list of genes that back with high Fst hits in their supplemental table a lot of them looked very familiar. So we should be cautious about over-imputing form these results. But perhaps with a better exploration of the genomics of these phenotypes we can figure out the details of the epicanthic fold…though I suppose phenotyping adults would be a little strange in Korea.
Finally, the very last sentence kind of annoys me: “If we consider a very unique genetic status of Jomon people, then it is understandable that Ainu people, who inherited the highest proportion of Jomon DNA, are quite unique among all extant East Eurasians.” I am curious how the Ainu and the Jomon relate to other Asian, an non-Asian, people. Normally I don’t do anything like this, but I sent an email asking for the genotypes to test myself.