I noticed that many Japanese people score between 0.5% and 2% of South Asian admixture in various calculators, including Dodecad K12 and K12b and HarappaWorld K16. I thought at first that it might be a misattribution of very ancient admixture linked to Y-haplogroups C and D, but I recently saw some Japanese results from Living DNA, and that individual scored about 1.8% of South Asian too, which according to the admixture map was closest to the Balochi of southern Pakistan. He also had a surprising 1.9% of northern Near East (Caucasus/Turkey). I ran the same sample in HarappaWorld, but he got 1.89% of South Indian and 0% of Baloch.
I checked the Dodecad K12 spreadsheet for other East Asian results. The Japanese samples from the HapMap and HGDP all scored 0% of South Asian, but the Dodecad participants scored 1.5%. Individual samples I ran also scored around 1.5%. So there seems to be a discrepancy between the full genomes from the HapMap and HGDP and the commercial (23andMe, FTDNA and Living DNA) genomes. I observed the same thing with the Chinese samples. The commercial ones got 0.5% of South Asian, but those from HapMap scored 0%. The Cambodians from the HapMap had 10.5% of South Asians though, while the Thai scored even higher at 15.7%. The Dai of SW China and Vietnam scored 1.4% and the ethnic Vietnamese 3.2%.
The K12b spreadsheet showed 0% for all Japanese samples, even the commercial ones. However that's probably a mistake considering that the 4 samples I ran got between 0.2% and 1.2%, with an average of 0.7%. Cambodians scored even higher, at 15.2%, while the Burmese have 18.7%.
Based on the commercial samples analysed by the Dodecad K12 and K12b, it seems that the Japanese could possess about 1% of South Asian admixture, and that it could have come through South East Asia, as it is completely absent in East Siberia, Korea or northern China.
In my experience, HarappaWorld is a more reliable calculator than Dodecad (or most others for that matter). So I checked the HarappaWorld's average scores by population. There are two Japanese values. I believe that the first (99% NE Asian) may be an ethnic Ainu, as I have never seen such scores in regular Japanese people. The other one gives 74% NE Asian, 11% Siberian, 11% SE Asian and 1% South Indian.
Both the North and South Han Chinese lack any trace of that South Indian admixture, but the Vietnamese have 1% like the Japanese. The Khmer-Cambodians score 13% (+1% of NE Euro), which is just between Dodecad K12 and K12b. The Thai score 14% pf South Indian + 2% of Baloch, 2% of Caucasian and 1% of NE Euro.
So it does seem that the Japanese have a bit of South Asian admixture and the most likely explanation is that it came from South East Asia, although no historical migration is recorded. Once again DNA uncovers traces of prehistoric migrations about which nobody knew anything.
I checked the Dodecad K12 spreadsheet for other East Asian results. The Japanese samples from the HapMap and HGDP all scored 0% of South Asian, but the Dodecad participants scored 1.5%. Individual samples I ran also scored around 1.5%. So there seems to be a discrepancy between the full genomes from the HapMap and HGDP and the commercial (23andMe, FTDNA and Living DNA) genomes. I observed the same thing with the Chinese samples. The commercial ones got 0.5% of South Asian, but those from HapMap scored 0%. The Cambodians from the HapMap had 10.5% of South Asians though, while the Thai scored even higher at 15.7%. The Dai of SW China and Vietnam scored 1.4% and the ethnic Vietnamese 3.2%.
The K12b spreadsheet showed 0% for all Japanese samples, even the commercial ones. However that's probably a mistake considering that the 4 samples I ran got between 0.2% and 1.2%, with an average of 0.7%. Cambodians scored even higher, at 15.2%, while the Burmese have 18.7%.
Based on the commercial samples analysed by the Dodecad K12 and K12b, it seems that the Japanese could possess about 1% of South Asian admixture, and that it could have come through South East Asia, as it is completely absent in East Siberia, Korea or northern China.
In my experience, HarappaWorld is a more reliable calculator than Dodecad (or most others for that matter). So I checked the HarappaWorld's average scores by population. There are two Japanese values. I believe that the first (99% NE Asian) may be an ethnic Ainu, as I have never seen such scores in regular Japanese people. The other one gives 74% NE Asian, 11% Siberian, 11% SE Asian and 1% South Indian.
Both the North and South Han Chinese lack any trace of that South Indian admixture, but the Vietnamese have 1% like the Japanese. The Khmer-Cambodians score 13% (+1% of NE Euro), which is just between Dodecad K12 and K12b. The Thai score 14% pf South Indian + 2% of Baloch, 2% of Caucasian and 1% of NE Euro.
So it does seem that the Japanese have a bit of South Asian admixture and the most likely explanation is that it came from South East Asia, although no historical migration is recorded. Once again DNA uncovers traces of prehistoric migrations about which nobody knew anything.
