These are the remaining populations from yesterday's unsupervised Euro/Caucasus/Siberia/Amerindian/East Asian run.
East Asians all cluster very close to both Siberians and Amerindians in World-wide PCA. I'm growing to believe that the Chinese Yellow River/Yangtze River Millet/Rice Neolithic Revolution is about an expansion of Siberian-like peoples independently adopting Agriculture. This would again happen elsewhere once or twice (according to whether Mesoamerican and Peruvian agriculture are truly independent of one another or not) in the Americas.
But the most successful "Siberian"-derived Neolithic is definitely the East Asian one. Expanding not only across East Asia, but also into the Pacific (from Taiwan, ultimately from China's Yangtze valley?) and marginally into India and it's Ocean up to uninhabited Madagascar (before the Bantu could get there with their better adapted Tropical crops, mix with them at around 50/50, while less admixed people remained in the Highlands).
These may correspond to two Core Populations, and I think a model similar to the Fertile Crescent one is likely, with an earlier maturation of Millet-based agriculture in a previously Siberian-like forager population living in the Yellow River area and it's expansion South to the Yangtze River but not much further due to mis-adaptation of Millet to warmer/wetter climates. Another wave perhaps from the Yangtze River valley with Rice, a much more productive cereal, into Southeast Asia and Taiwan (and from there onwards), and also expanding into the North and into Japan. As in the Fertile Crescent, this second wave would face higher drag, and absorb more and more First wave elements as it travelled, explaining perhaps low "Second wave" in the Japanese (which may have only successfully settled Japan from Korea after getting Rice). I think I can perhaps see a third reexpansion of the Northern Core, maybe related to the early adoption of Fertile Crescent-derived Wheat?
Another point is why didn't the Chinese occupy Siberian lands before the Russians did so? Again I think the answer may lie with Rye, a much better cereal for the Eastern Siberian climate and soil types. And probably also the development/adoption of Pastoralist lifestyles in marginal lands.
I'll review the model with a supervised analysis later, since these unsupervised modal populations might be artificial (though relationships should hold).
Some minor elements of note:
1. There is a Papuan-like "Coastal Migration" element in Japanese I think might correspond to the aboriginal Japanese (Ainu and related populations to the South), which may have also Siberian admixture?
2. You can see Amerindian-like very tiny elements in some Southeast Asians. I think since the Chinese Neolithic was probably a "Siberian" development, these may have been carried south by the Neolithic waves since I don't think they'd be aboriginal. Indeed this element together with more conventional Nganasan-modal green one, can be found to some extent in all East Asians, up to the Chukchi living in the Bering Straight.
3. There is a small Fertile Crescent element in Thais, Khmer and other Southeast Asians. I've seen it repeatedly under different conditions and set-ups. There is a large Fertile Crescent element in all Indians, and I've seen it eslewhere in an ADMIXTURE run including East Bengalis. Maybe Burmans have it as well and it extends here. It may have arrived before the Northeastern Waves, since it is a minority component. Interestingly, it is quite larger than the Papuan-derived "Coastal Migration" component. I suspect low adaptability to wet and hot conditions of Fertile Crescent crops meant it could not penetrate most ecosystems of the region thoroughly. Yet it seems it was latter easily swamped out by the Chinese probably Yangtze River-derived wave. Yet this did not happen in India, or in the Melanesians, Papuans, despite long contact with the East Asian Neolithic. Papuans have contemporary documented independent Agriculture in the highlands of New Guinea. South Indians don't, yet the genetic result of high "Papuan" containing Tongans and Samoans and South Indians appears similar...
Still, these results are preliminary ones, I'll do a supervised run later on...