The recent paper on three ancestral European populations has some truly interesting stuff buried deep in the supplements. This is not the first time that this has happened: if you read the supplements to their big Neanderthal paper, back in 2010, you could see a strong implication that Melanesians, (people in New Guinean, the Solomons, and Australian Aboriginals) had experienced an additional admixture with archaic humans ( as it turned out, ~5% from Denisovans). That implication was obscured by an incorrect caption, and in any event it was on page 136… I’m not sure anyone else noticed.
In this case, they talk about the third ingredient that moved into Europe. They conclude that in most parts of Europe, it looks as if modern populations inherited the three EEF/WHG/ANE groups ( Levantine farmers, West Hunters and Sibermen) via only two proximate ancestral populations. Europe at the time was almost entirely occupied by Sardinian-like farmers – then another population moved in, one that had about 3 times as much West Hunter as Sibermen. The hint is that the ratio of WHG to ANE is almost a constant. They put it this way: a large number of European populations are consistent with descent from identical “Farmer” and “Hunter (= WHG + ANE)” populations. ”
Here’s a thought experiment. Imagine that Mexicans are half Spanish and half Amerindian – then drop them in varying amounts in various Korean provinces. The total Mexican fraction may vary, but the amount of Amerindian ancestry in a given province will always be the same as the amount of Spanish ancestry.
This means that the WHG fraction you see came with along the ANE guys – it is not a comeback by the descendants of the original hunter-gatherers of central and Western Europe, who were somewhere between much diminished and extinct, but an invasion (from farther east, probably) by a group that is somewhat related to those original hunter-gatherers and also has a fair amount of ancestry from Sibermen- a population that also contributed to Amerindians. You could see hints of this in the haplotypes and uniparental lineages – I talked about it earlier.
This also means that the change in ancestry around 3000 BC was not small – it’s much bigger than the ANE fraction. The suddenly-appearing component accounts for half or more of the ancestry of all the populations of northern Europe, all the way from Belorussia to Scotland. That population might also have had some EEF, in which case the population replacement factor would be even bigger.
The WHG + ANE fraction is about half in France, lower than that in the South of France. It’s much lower in southern Europe. The Albanians, Ashkenazi Jews, Greeks, Maltese and Sicilians have essentially no WHG ancestry, while a decent fit is possible for the Basque, French_South, Bergamo, Pais-Vasco, and Sardinians using no Siberian component. Using another estimation approach you get a nonzero but lower Siberian component in Basques, and almost none in Sardinians.
This last wave can only have come from a region that had both WHG-like and ANE-like ancestry. That rules out the Middle East: they don’t have WHG. You see quite a lot of ANE in the northeast Caucasus and the northern Near East – the high is in Chechens and Lezgin.
All this is pointing to a big wave of genetic change around 3000 BC, a wave that did not originate in the Near East (including Anatolia). Perhaps the strongest hint is what they don’t say in this paper. Although they mentioned such possibilities in an earlier related paper, they make no mention of the Indo-Europeans. They don’t even use the word language anywhere in the supplement. All this while they’re clearly working out the origin and scope of the Indo-European invasion!
So they must have a paper in the works – one with strong conclusions, if they’re that worried about someone scooping them. They were too obvious – they need a course in maskirovka. They have got to be looking at ancient DNA from relevant populations: from Kurgan burials, from Russia north of the Caucasus, Tocharian mummies, etc. They’re going to need strong evidence, and a baseball bat, to get linguists to pay attention. Looks to me like we”re going to get at least half of the story (the European end) of the Indo-European expansion out of this, and probably we’ll learn something about Indo-Aryan end as well.
I will predict this: they’ll pussyfoot about the likely historical process, which undoubtedly was awesomely bloody. The Balkans looks to be where this started, and there a fairly sophisticated agricultural population (with very advanced metallurgy for the time) seems to have been utterly squished.