Monday, 9 May 2011

Western, Eastern, Northern, Southern: Motherland?

In this post I'll tackle the last unsupervised component: the West Asian-modal component, which peaks in Georgians and exists in high amounts in all West Asians. This component appears much throughout Europe, and strangely, in unsupervised runs it is a bit closer by Fst distances to the Lithuanian-modal component than it is to the Sardinian or Basque-modal ones. What is going on here?

To recapitulate my previous results/interpretations:
1. Basque and Sardinian-modal. In MDS and PCA plots, Sardinians and Basques plot not far from each other, but in distinct clusters out of the European "mainstream". Sardinians further towards modern Near Eastern populations, Basques more towards modern North European populations. In supervised admixture runs, both populations seem to be dominated by the same element, with Sardinians having a smaller element with North African/Near Eastern affinities, and Basques mostly without this element but with another element predominant in Northeastern Europe. I interpreted these results as indicating "2nd wave" influence on Sardinians, and "Eastern Wave" influence on Basques, exclusively superimposed on a common "Western Wave" element. This element, also found in large percentages in Southern, Northwestern and Central Europeans, but lacking in Northeastern ones, would correspond to the first Wheat-planting Neolithic expansion into Europe. Less fertile/colder areas not congenial to Wheat would have been moslty left to remaining forager populations providing meat and fish in exchange for wheat. Indeed such an arrangement is documented in several places in the World, and is still found in some Southeast Asian regions, where demographically dominant agriculturalists in more fertile/rice adapted soils exchange agricultural products with protein from foragers in marginal lands.
2. Lithuanian-modal, also present mixed with Siberian elements in the Chuvash, seems to be prevalent in cold/poorer soil areas of Europe. I think it may correspond to an Eastern Neolithic Wave from the Northern Near East through the Caucasus and Steppe into Eastern Europe. After adoption of cold-adapted agricultural techs, such as winter-rye, it expanded into the vast niches left mostly unoccupied by the earlier Western Mediterranean-Atlantic and Danubian waves. It replaced the forager populations still present in those marginal lands. An analogy is to the settlement of Japan by the Yayoi people. Only after developing cold-adapted rice varieties and techniques were they able to perhaps migrate and completely replace the Ainu-like foragers in rich coastal foraging environments less suited to earlier Neolithic lifestyles.
3. Mozabite/Tunisian and Bedouin/Saudi/Egyptian components are mostly about peaks of NW African and NE African ("NileCore") components. The Green Sahara may have had semi or full pastoralist developments explaining the presence of such components, admixed with "Mesopotamian Core" ones in such a high degree in some nomadic populations.
4. "Siberian" and "Amerindian"-like small elements in Finns, Russians, Scandinavians, Balts and Irish, British are I think not derived from any undocumented migration of ancient Siberian peoples but genetic traces of Native European populations. I've tentatively interpreted "Amerindian"-like segments in North Europeans and people from the Caucasus Mountains (which also appear in unsupervised runs) as survivals of an old "Amerindian"-like aboriginal population, making it's way to the Americas through the Ice Age ice cap with a lifestyle similar to that of Inuit/Eskimos today.

So the last component still eluding explanation is the "West Asian" one. Why is this closer to North European than to any other unsupervised mode component?
I've stated before, I think this component is a Mesopotamian Inner Core element, very similar to the European Western and Eastern Wave ones. Indeed these last may be "frontier" elements with less diversity, subsets of the Inner Core one. This would explain why Near Easterner and Southeast European populations tend to have large balanced "Basque5" and "Chuvash5" elements in previous runs. So the Inner Core element may be the "Mother" of these less diverse frontier elements that expanded into Europe. Later, the Inner Core populations, perhaps in Mesopotamia, would achieve more evolved Neolithic capabilities and expand into the Western and Eastern Wave settled regions.

So "West Asian" peaks in Northern Near East populations, such as Armenians, Iranians, Georgians, however using these populations as the pole is problematic since these same populations have signatures of "Nile Core" influence. It however also appear in high levels in Northern Caucasus populations such as Lezgins, Adyghei, and in the samples from the Mountain Dagestan towns of Urkarah and Stalskoe.
Using my "restricted pole" supervised run trick, I used first Georgians as the pole. As expected this erased much "2nd Wave" influence from European populations, since the "Georgian5" pole attracted many such segments together with the North African one. Using "Urkarah5" I got higher percentages in Northern Europe than seemed reasonable and since in previous unsupervised runs, Lezgins, Urkarah and Stalskoe seem to have some "Northern European" in addition to predominant "West Asian" I thought it would be best to combine some Georgian, Iranian and Urkarah individuals to allow the pole better to "focus" on the element I was searching for.
Naturally this is discussible, but results using any of the poles are not very different.
I also decided to ditch Mozabites and Egyptians and replace them by Tunisians and Palestinians as NW African and NE African (Nile Core) poles respectively. Mozabites and Egyptians are more southern populations and any influence from North Africa in European populations is already represented in Tunisians and Palestinians. I excluded all other North African populations in order not to complicate things further, except for predominantly agricultural Northern Moroccans, which I wanted to use to pull the "Tunisian5" pole into the region (in order to find influences in South Western Europe).

I'm dividing this run into three parts: Near East; Europe; and Siberia+Central Asia. They are all part of the same run. First I'll present results for the Near East. I included some populations in Near East and European posts to link them, since it's all from the same analysis.

Poles used:
1. Siberian1: 5 random Yakut individuals "Yakut5" restricted pole. This captured most of the Siberian Turkic and Mongolic elements. It peaks in Buryats, with Yakuts and Mongols not far behind.
2. WMPC: "Basque5". This pole captured an element which peaks in Basques and Sardinians, but is also present in Southern and Northwestern Europe in important amounts. Somewhat surprisingly, it is also important in the Levant (unlike the other element present in Northwest Europe, "Chuvash5"). This is the Neolithic first Western Wave, perhaps associated with a Mediterranean-Atlantic migration route and its Megalithic monuments, depending on river-valley wheat agriculture. It also presumably represents here the probably very closely related Danubian wave.
3. NMPC: "Chuvash5". I renamed it Northern MPC since it seems to be Northern relative to the Fertile Crescent area. It peaks in Europe in the North East, and I've called it "Eastern Wave" before. Maybe a cold and poor soil-adapted Neolithic Wave, and rye agriculture (also perhaps occupying much of the "wheat niche" in some parts of Eastern Europe).
4. WMPC+NC. West MPC with some Nile Core, also referred before as "2nd Wave", or out-of-Egypt. Perhaps associated with proto-Semitic (and Exodus tales). A late expansion from perhaps 6000-5000 years ago. It was based on "Palestinian5" however it came to be dominated by the Druze, perhaps due to multiple family connections in this somewhat isolated population. It doesn't matter though, since the Druze are still adequate representatives even as a modal population. The Palestinians, Jordanians and others had their higher Nile Core component taken over by the Tunisian pole. I could remove the Druze, but things wouldn't change much and I prefer to keep all populations at this point. French Basques also seemed "inbred" before, but I'm now convinced their "isolated population pole-pulling" tendency is mostly due to their unique Neolithic wave mix.
The second wave element is closer to the "Basque5" element than to any other MPC element. I think the native element of the Levant may be closely related to the "Basque5" element, and this would be the subset expanding into less advanced incipiently Neolithic, Nile Core dominated Egypt, synthesize with it, and reexpand into its homeland and beyond as the "2nd Wave". This would be a more advanced Wheat-planting expansion, and the probably much larger food surpluses making it possible also allowed higher organization levels, specialization, and elite forming in Egypt itself- a process leading to ancient Egyptian Civilization.
I seem to recall a R1b Y-haplogroup found in ancient Egyptian remains-perhaps it was native after all since ancient Egyptians might have been a synthesis of WMPC and NC?
5. EMPC: based on two Urkarah individuals+1 Iranian+2 Georgians. If based on "Urkarah5" or "Gorgian5" it has a greater tendency to draw too many non-pole individuals of said population and any 2nd Wave or NMPC elements also present there. By combining the restricted pole, this problem can be reduced. I think EMPC is an inner-core MPC element perhaps present in Mesopotamia itself. It seems to have expanded against the WMPC and NMPC within the Near East after these lasts' expansion outside of it, but likely well before the second wave. Also seems to be the subset corresponding to Ancestral North Indians.
6. WMPC+NWAf: based on "Tunisia5". This component likely consists of mostly WMPCA+some Green Saharan derived NWAfrican I found before. Tunisians were drawn at almost 100% into the pole, unlike Northern Moroccans, probably due to multiple distant family relations within the Tunisian sample. Still matters little, since it is representative of Western North Africans. Since I believe Tunisians are still representative of ancient Berber-speaking Neolithic populations I removed Mozabites. Using Tunisians may allow more sensitivity, since Mozabites are more distant.
7. Siberian1. "Nganasan5" using only unadmixed Nganasan. Representing a Western Siberian element.
8. Siberian3. "Mongolian5". Strangely this pole didn't pick up much in other Mongolians, but instead focused on Chukchis and Koryaks from the Far East. Mongolians turned out mostly "Siberian2". This happened presumably due to closeness of "Yakut5" and "Mongolian5" (just as before I used "Dogon5" to capture NWAfrican with the help of another West African pole). This happens frequently in "restricted pole" supervised runs, for instance in Northern Europe runs even a "Orcadian5" versus "Hungarian5" pole analysis reproduces imperfectly the "Basque5" vs "Chuvash5" scenario, since "Orcadian5" and "Hungarian5" shift away from their individuals' populations and peak in Basques/Sardinians and Lithuanians/Chuvash instead. So restricted pole runs in my opinion tend to shift towards real patterns in the data.

This is still an imperfect run, don't read too much from smaller components. Still results are similar using various different set-ups (and not too different from unsupervised ones).

Some speculative considerations:
-WMPC is present in the Levant and North Africa, whereas NMPC is not. It is also dominant over it in Turks, Armenians, different Jewish groups. This suggests that WMPC ("Basque5") has its ultimate origin here, in the Levant and Anatolia. It seems to have expanded not only into the Mediterranean and Balkans, but also into North Africa. I think the 2nd Wave likely is just WMPC admixed with some North Eastern African ("Nile Core").
-NMPC is dominant over WMPC in Iranians, Kurds, North Caucasus and Eastern Caucasus. I tend to think it derives from an ancient population living in the Northern Mesopotamian/Caucasus range/Western Iranian plateau region.
-EMPC may correspond from an expansion into both the WMPC homeland (Levant/Anatolia) and NMPC homeland (Northern Mesopotamia?/Eastern Anatolia?/Western Iran?) from an inner area, perhaps Southern Mesopotamia.

I'll present European results later.

No comments:

Post a Comment