Monday, 4 April 2011

A New World

Assuming the model and run are valid, Europe is where the Northwestern Nile Core wave finally runs out of steam. Perhaps coming out of Turkey through the Balkans, and through the Mediterranean village to village links established by their predecessors, it finally dilutes itself so much into the established Mesopotamian Core element that it ceases to be detectable, at least by this method. I suspect "Nile-Core acculturated" Mesopotamians from Southern Europe continued to carry some of the innovations well into Northern Europe, but this is harder to detect in autossomes due to higher similarity between the incomers and the settled. Dominant groups tend to spread Y-DNA far further than autossomes though.
Radiating from the Southeastern Corner of the Mediterranean, this people may have an even greater influence in Y-DNA explaining a large part of the current distribution haplogroups E1b1b and J1c3 and perhaps others, some of which may now be more frequent out of Egypt due to founding effects. Also maybe the tiny percentage of L mt-DNA in Europe.
Speculating, the Spaniard with the high Nile Core element may come from the coast, maybe the Mediterranean Coast of Spain. The Italian with low Nile element maybe from an Alpine valley?
I don't know what the Pygmy pole is picking up in Europe but it's definitely not Pygmy. I think it's probably nothing much, but I named it "Forager B?" rather than naming it "noise".
Hunter-gatherer tribes in Europe were likely much closer to each other than in Africa, even though much more different over the same distances than a Neolithic Core component. Siberians don't easily get their own ethnic component in ADMIXTURE runs as San and Pygmies do. Still don't know if "Pygmy" (Forager B?) and Hadza represent anything at all, most likely not.
Basques have no or only a small "Nile Core" element. They lack many of the more colourful elements found in residual amounts in both Spaniards and the French. In general a good question is are groups with different languages/cultures/religions from their related neighbours more isolated because of said language/culture/religion, or do they have less subsequent outside influence and different cultural paths because they were isolated for other reasons (like geography) from the start? This could also explain why the Druze and others have relatively less of a "second wave" element.
I suspect not many Southern, Mediterranean French were genotyped into this set, maybe there's one?
Cypriots are quite homogeneous, lacking extra elements to the Neolithic ones, as expected in an island population close to the source-areas.
Askhenazi Jews have residual "Siberian" elements, bringing Khazars to mind, but probably mostly due to their Eastern European origins.
Romanians have much "Evenki+Yakut" maybe due to their proximity to the steppe. Some of these individuals appear to have significant levels of Roma admixture.
Sardinians are a very homogeneous population, with no extra-Neolithic elements, and can serve as a "template" for other Southern Europeans. No wonder ADMIXTURE often considers them an "unadmixed" population.
Hungarian tribes were more Nganasan-like than Evenki-Yakut-like, coming from Western Siberia, and you can see this here.
Don't know why Orcadians have visible Nganasan here.
Utah whites likely consist of New England-derived settlers of mostly British but with also significant German, Dutch, Irish ancestry. The larger Nganasan segments (larger than in Orcadians, and I suspect, even larger than mainland Britons, you see in several may indicate residual Native American ancestry? I don't know why ADMIXTURE would choose Nganasan to pick up Amerindian segments. Drift making them randomly slightly more like them than Evenki? Chukchi was already taken as the Mesopotamian pole, since it doesn't exist in ME/Europeans (I think something similar may have happened to "San" and others. Maybe Evenki+Yakut was "stretched" too much towards Mongolian/East Asian groups.
"Siberian" elements in Finns, Balts and Russians are I think most likely aboriginal.
These I think are Northern Russians. Central Russians might look more like Byelorussians.
Fertile Crescenters and their crops and ways didn't find Siberia very inviting at first. Then Rye, a Middle Eastern grain that often grew there as a pest in the middle of more productive ones, started to be planted widely, probably in the Corded Ware zone, and the East Slavs perhaps radiated from there into the East. It's possible that armed with this new productive food tech and maybe some other innovations for colder climates, they were finally able to go all the way to Vladivostok.


  1. Many of the Siberians from Rasmussen et al. have Russian admixture.

    From my experience, only some of the Nganassan, for example, are useful in supervised runs - the ones that don't deviate towards Europe across the main dimensions on MDS plots.

  2. Yes, "Nganasan" may be overestimated in neighbouring populations due to this. But I doubt it has much effect in more distant ones. I'll redo the run removing all obviously recently admixed individuals later on. The results are still reasonable, confirm theoretical predictions and are quite homogeneous within populations even without removing them.