December 12, 2011

The womb of nations: how West Eurasians came to be

One of the most interesting observations about the genetic structure of the Old World, is the fact that distantly located populations are often more similar to each other than to their more immediate geographic neighbors.

For example, in my recent K12a admixture experiment, the six components. I named Mediterranean, North_European, Caucasus, Gedrosia, Southwest_Asian, and Northwest_African have a maximum Fst between any two of them of 0.073 (between Gedrosia and Northwest_African), and a mimimum Fst between any of them and any of the others of 0.075 (between Gedrosia and South_Asian). Henceforth, I will call these six components simply "the Six."

It is remarkable that "Gedrosia", the component peaking in present-day Balochistan is about equidistant to the component peaking in Mozabite Berbers ("Northwest_African") and that peaking in South Indian Dravidian speakers ("South_Asian"). Going by geography alone, and even if we calculated distances "as the crow flies" and ignored all natural obstacles between Balochistan and the Sahara, we would have expected "Gedrosia" to be about 3 times more distant to "Northwest_African" than to "South_Asian".

Divergence between populations across the entire genome builds up mainly by two processes:
  • Genetic drift
  • Admixture
Over time, two populations stemming from the same root will shift in their overall allele frequencies because of random factors (drift). These differences may become even more pronounced by new mutations arising in each population, and by natural selection operating in different environments.

Moreover, if these populations migrate from their original homeland and absorb the indigenous inhabitants wherever they go, then they will diverge even more, depending on how much admixture they undergo, and how distantly related the aboriginal inhabitants are.

It is clear that admixture has played a role in the overall divergence of the Six. For example, the Northwest_African component is shifted (relative to the remaining five) towards the other African components; the Southwest_Asian is also thus shifted, but less noticeably. The Gedrosia component is shifted towards the South_Asian one; and the easternmost components, the North_European, and Gedrosia ones, are shifted towards the Asian components.

All these shifts are quite salient in the MDS plot, and there is ample evidence for the aboriginal populations being shifted in the expected direction in each region.

We can calculate the median Fst within the Six: it is 0.053, as well as the median Fst between members of the Six and all the rest: it is 0.13. The ratio of the two is ~40%. Of course, the various components have diverged from each other at different times: West and East Eurasians, for example, began diverging shortly after both diverged from Africans. Nonetheless, we can (conservatively) place the divergence time of the ancestors of the Six from East Asians, Ancestral South Indians, and Sub-Saharan Africans, at around 40,000 years ago, the time when the Upper Paleolithic (and modern man) makes its appearance all over the Old World.

Actual divergence times may be lower, and this is why 40k is a conservative estimate. The point of this back-of-a-napkin calculation is to give a rough estimate, rather than a precise date.

Assuming that Fst builds up roughly linearly with time due to drift (again a simplification), we can estimate that divergence within the Six dates to less than 16 thousand years ago. This may be an overestimate for two reasons:
  • Divergence between Proto-West Eurasians and the rest of mankind may be less than 40k years
  • Each of the Six have partially absorbed aboriginal inhabitants (e.g., Palaeo-Europeans, Ancestral South Indians, Pre-Berber North Africans, etc.) that spent most of the Paleolithic diverging from the common ancestors of the Six.
It is therefore clear, that the common ancestors of the Six may have been a people living in some small area of Eurasia until well after 16 thousand years ago. Much like a bubble of space in cosmological inflation, their living space multiplied by orders of magnitude, coming to encompass a huge region of space within most of West Eurasia and North Africa. Within that region a fairly homogeneous population was established, the people traditionally called "Caucasoids."

The womb of nations

The Neolithic of West Eurasia started, by most accounts, c. 12 thousand years ago. Its origin was in the area framed by the Armenian Plateau in the north, the Anatolian Plateau in the west, the Zagros Range in the east, and the lowlands of southern Mesopotamia and the Levant in the south. Intriguingly, the prehistoric site of Göbekli Tepe sits right at the center of this important area, in eastern Anatolia/northern Mesopotamia.

If there is a candidate for where the ur-population that became the modern Six lived, the early Neolithic of the Near East is surely it. This hypothesis makes the most sense chronologically, archaeologically, genetically, and geographically.

Migrants out of the core area would have spread their genes in all directions, becoming differentiated by a combination of drift, admixture, and the selection pressures they faced in different natural and cultural environments; some of them would acquire lighter pigmentation, others lactase persistence, malaria resistence, the ability to process the dry desert air or to survive the long winter nights of the arctic. These spreads were sometimes gradual, sometimes dramatic: they took place over thousands of years and from a multitude of secondary and tertiary staging points.

In Arabia, the migrants would have met aboriginal Arabians, similar to their next door-neighbors in East Africa, undergoing a subtle African shift (Southwest_Asians). In North Africa, they would have encountered denser populations during the favorable conditions of MIS 1, and by absorbing them they would became the Berbers (Northwest_Africans). Their migrations to the southeast brought them into the realm of Indian-leaning people, in the rich agricultural fields of the Mehrgarh and the now deserted oases of Bactria and Margiana. Across the Mediterranean and along the Atlantic facade of Europe, they would have encountered the Mesolithic populations of Europe, and through their blending became the early Neolithic inhabitants of the Mediterranean and Atlantic coasts of Europe (Mediterraneans). And, to the north, from either the Balkans, the Caucasus, or the trans-Caspian region, they would have met the last remaining Proto-Europeoid hunters of the continental zone, becoming the Northern Europeoids who once stretched all the way to the interior of Asia.

It is, perhaps, in the ancient land of the Colchi, protected by the Black and Caspian seas, and by tall mountains on the remaining sides, that something resembling the ur-population survived. The great linguistic diversity of the Caucasian peoples, the central position of the Caucasus component among the Six (see MDS plot), and the fact that theirs was a remote region, ignored by the empires of the Near East to the south, and the frequent travellers of the Eurasiatic steppe to the north, may all indicate the plausibility of this assertion. Through a peculiar coincidence, old Blumenbach may have been onto something, although for reasons he could scarcely have imagined.

We don't have to suppose that a single process drove the dispersal of genes from the core area of the Near East. Agriculture was definitely an early facilitator of dispersal, but a variety of technological developments may have been instrumental in further spurts of dispersal: the invention of pottery, metalworking, pastoralism, sea navigation, the engines of war, all the way to the recent past, when ocean-going vessels from Western Europe sailed to the New World, Arab merchants and warriors spread their new religion to the continent of Africa, and Russian hunters and farmers began the conquest of Siberia.

There are hints that scientists are beginning to realize that this story may describe what happened during prehistory. The next few years will help resolve this grand puzzle of "how West Eurasians came to be."

63 comments:

Pascvaks said...

And here I thought we were born and grew up on the shores of the Caspian Sea.

Grey said...

"the prehistoric site of Göbekli Tepe sits right at the center of this important area"

I find that a very intriguing possibility if you combine the following list:

1) Mountain
2) Wild wheat (or equivalent) growing on the mountain
3) Religious festival involving lots of people camping out on the mountain
4) Festival goers gathering wild wheat (or whatever) to eat
5) Latrines being dug all over the mountain and filled with wheat seeds

After a few years you get a mountain covered in wheat.

Then if you had some priests or monks taking up residence at the site they might enclose some of it in fields to keep animals from eating it.

Accidental discovery of farming.

Lank said...

"In Arabia, the migrants would have met aboriginal Arabians, similar to their next door-neighbors in East Africa, undergoing a subtle African shift (Southwest_Asians)."

On what are you basing the bolded statement?

Dienekes said...

>> On what are you basing the bolded statement?

Prior to the recent dispersals inaugurated by the Neolithic (and not only in West Eurasia, the same process happened in East Asia and Sub-Saharan Africa), the genetic landscape coincided with the geographic landscape: people were more similar to their geographic neighbors. In the millennia since the Neolithic there were major migrations of people, first as random movements, more recently organized movements of immigrants and armies: these have displaced gene pools far from where they originally evolved.

Lank said...

"Prior to the recent dispersals inaugurated by the Neolithic (and not only in West Eurasia, the same process happened in East Asia and Sub-Saharan Africa), the genetic landscape coincided with the geographic landscape: people were more similar to their geographic neighbors. In the millennia since the Neolithic there were major migrations of people, first as random movements, more recently organized movements of immigrants and armies: these have displaced gene pools far from where they originally evolved."
Sure, I am willing to agree with this.

I just think you are underestimating the barrier of the Red Sea. Prior to the Neolithic, I see no reason why the populations of Arabia should have been particularly close to East Africans.

I would guess that East Africa was inhabited mainly by people with mtDNA L3 (L3a, L3elkx, L3f, L3h) and L4, whereas Arabia was inhabited by mtDNA M/N people.

Dienekes said...

They don't have to be very close. Both Arabia and East Africa are not the place they were 12,000 years ago. First, they exchanged genes with each other; second, they exchanged genes from the rest of Eurasia/Africa respectively. But, a reasonable assumption is that given that Arabia is closer to Africa than the rest of Eurasia, its aboriginal population would have been closer to it genetically as well, and hence the absorption of this aborigines would have shifted population somewhat in an African direction; and, indeed, this is what we see with respect to the "Southwest_Asian" component.

Roy said...

"the prehistoric site of Göbekli Tepe sits right at the center of this important area"
..more or less the Kurds (unrecognized) homeland.

Dmytro said...

Are you saying that current populations in the areas of the "Six" actually descend from a fusion between the new arrivals (in various spurts) and the older migrants? If so, is there a way of genetically calculating the contribution of each group to the subsequent mix? Presumably we can do a little bit of this with theories about the expansion of Y-DNA and mt-DNA. Can we link the former to the autosomals to get a more accurate picture of things? or is that yet a task for future analysis?

Dienekes said...

>> Are you saying that current populations in the areas of the "Six" actually descend from a fusion between the new arrivals (in various spurts) and the older migrants?

Correct.

>> If so, is there a way of genetically calculating the contribution of each group to the subsequent mix?

Not an easy one. Reich et al. (2009) attempted to do this in the case of Indians, by exploiting the fact that the Onge Islanders appeared to be a sister group of the Ancestral South Indians.

It may be possible to extract "aboriginal" segments even if there are no extant unadmixed populations. The Taino Project is doing this, but I'm not sure whether they use information about other Amerindians in the process.

Ancient DNA will doubtlessly help, when it arrives for pre-Neolithic inhabitants of various parts of the world.

Matt said...

One of the most interesting observations about the genetic structure of the Old World, is the fact that distantly located populations are often more similar to each other than to their more immediate geographic neighbors.

Looking at actual population FSTs and not component FSTs from the Metspalu et al paper, that doesn't seem like the case anywhere between populations in West Eurasia, although true of http://download.cell.com/AJHG/mmcs/journals/0002-9297/PIIS0002929711004885.mmc1.pdf

It is true in India and between Uyghurs and Hazaras, but South Asia and steppe cases which are the result of recent historical time.

For you components it seems true only for the South Asian and Gedrosian

Each of the Six have partially absorbed aboriginal inhabitants (e.g., Palaeo-Europeans, Ancestral South Indians, Pre-Berber North Africans, etc.) that spent most of the Paleolithic diverging from the common ancestors of the Six.

I think the question here would be:

These aboriginal inhabitants are effectively hidden extra variables. Which historical facts constrain us from simply altering our model so that they show the required FSTs and required levels of admixture for a either "primarily replacement" or "primarily continuous" model, depending on which suits our bias?

Simply put, how do we use this to test models of replacement or continuity, without proponents of either model simply altering their assumptions as to the levels of differentiation and admixture between aboriginal inhabitants and the expanding population to suit whatever hypothesis they prefer?

I would guess the historical facts that constrain us would be : a) Population differentiation and time depth and mixture as measured by skeletal remains, but then we have only a very incomplete fossil record, b) technological and cultural horizons as evidenced by archaeology, c) mutation rates we can be very certain about.

Genomics on Mesolithic European adna would be really helpful here.

Dienekes said...

Looking at actual population FSTs and not component FSTs from the Metspalu et al paper

The actual population Fst's have been affected in both the sense of being brought together (e.g., West Asians moving to Europe and Europeans to West Asia throughout history), and in the sense of being torn aparat (in the sense that extant South Asians are part Gedrosia+South Asian, or k5+k6 in Metspalu parlance).

That is why I am considering components modal in particular regions, rather than extant populations.

These aboriginal inhabitants are effectively hidden extra variables. Which historical facts constrain us from simply altering our model so that they show the required FSTs and required levels of admixture for a either "primarily replacement" or "primarily continuous" model, depending on which suits our bias?

The continuous model doesn't work, because there is no way in hell it can explain why people throughout such a huge territory would be so similar to each other.

Gene flow across such a huge area could, indeed, homogenize these populations so that they would end up being quite similar to each other (low Fst). But, gene flow between bordering populations and their neighbors outside the West Eurasian area would mean that they would be more genetically similar to them, rather to people living halfway around the world.

Dispersal of some sort, from a relatively small area is the only reasonable solution for the observed homogeneity.

Matt said...

But, gene flow between bordering populations and their neighbors outside the West Eurasian area would mean that they would be more genetically similar to them, rather to people living halfway around the world.

Well, there is the case for intra-West Eurasian geneflow outside of India (including the North African fringe), on a similar basis that you have made for differential survival of aboriginal South Asians - Caucasoids don't thrive in the tropics.

That would be one other mechanism as to why Caucasoid populations within West Eurasia have more similarity to one another relative to neighbouring populations across a climatic threshold (varying with the steepness of that climatic threshold).

On the other hand, the ancient mtdna evidence you've mentioned before is against extensive homogenising geneflow between pre-neolithic West Eurasian populations.

CarlGrote said...

Do you think that gene flow from Africa into this West Asian urhemait (as evidenced by the finding of L2a1 mtdna in adna from Tell Halula) shifted the pre-expansion proto-West Asians away from other Eurasians?

Eze said...

"But, a reasonable assumption is that given that Arabia is closer to Africa than the rest of Eurasia, its aboriginal population would have been closer to it genetically as well, and hence the absorption of this aborigines would have shifted population somewhat in an African direction; and, indeed, this is what we see with respect to the "Southwest_Asian" component."

According to the Fst table the Caucasus component (which presumably isn't African admixed) is as close to the two Sub-Saharan African components as Southwest Asian is. Would this mean that the Caucasus component is also African admixed?

Dienekes said...

I hadn't noticed that the Caucasus and Southwest_Asian components are actually equidistant to Africa. That they are both closer to Africa than the rest makes sense, as they are both closer to it than all the other of the Six, so they were closer to it genetically in the Paleolithic landscape to begin with. This may also signify that the Southwest_Asian component either did not blend substantially with aboriginal Arabians, or these were not very African-like to begin with.

Extant populations of the area are African-shifted, but it turns out that this is due to their possession of small African components, rather than of the SW_Asian component itself being shifted relative to its immediate neighbor (Caucasus). In retrospect this is reasonable, if Semites (on which the SW_Asian component is centered) lived just to the south of the area indicated in the figure.

sykes.1 said...

Isn't this also the area that Colin Renfrew proposed as the homeland of the Indo-Europeans?

And, why didn't you identify these groups with Indo-Europeans?

Dienekes said...

I think Colin Renfrew placed the PPIE homeland in Central Anatolia, and also never claimed that the entire Neolithic was IE.

12,000 years ago appears to be beyond the capacity of linguistics to infer secure language relationships, although many have discerned such relationships in entities like "Nostratic", "Eurasiatic", etc.

I doubt that a single language was spoken by the ancestors of the Six. This is partly because hunter-gatherers tend to form smaller units than agriculturalists, and partly because of the fragmented landscape of the highlands of West Asia. So, I can envision a number of languages being spoken by them, which further diverged after geographical expansion and under the influence of different substrata.

I do think, of course, that Indo-European originated in the Near East, and in pretty much the same general area, being sandwiched between Proto-Kartvelian and Proto-Semitic.

http://dienekes.blogspot.com/2011/05/dolgopolsky-on-two-homelands-of-pie.html

But, its earliest spread dates to about 8,000 years ago, not 12,000 years ago.

Onur said...
This comment has been removed by the author.
Onur said...

According to the Fst table the Caucasus component (which presumably isn't African admixed) is as close to the two Sub-Saharan African components as Southwest Asian is.

Also, the Fst idstance between the "Caucasus" and "South Asian" components (0,080) is not that different from the Fst distance between the "Gedrosia" and "South Asian" components (0.075). Among the Six, next comes the "North European" component in Fst proximity to the "South Asian" component (0.087), and the remaining three components of the Six ("Mediterranean", "Southwest Asian" and "Northwest African") are equidistant to the "South Asian" component (0.098-0.099), obviously a coincidence in the case of the "Northwest African" component given the undeniable Sub-Saharan African ties of the "Northwest African" component, which would otherwise be closer to the "South Asian" component than it is now. So how can we explain this situation?

Regarding the Metspalu et al. equivalent of the "Gedrosia" component (k5) Metspalu et al. write thus:

"For alleles associated with k5, haplotype diversity is comparable among all studied populations across West Eurasia and the Indus basin (Figure S8 ). However, we found that haplotypic diversity of this ancestry component is much greater than that of those dominating in Europe (k4, depicted in dark blue) and the Near East (k3, depicted in light blue), thus pointing to an older age of the component and/or long-term higher effective population size (Figure S8 )."

"We found no regional diversity differences associated with k5 at K = 8. Thus, regardless of where this component was from (the Caucasus, Near East, Indus Valley, or Central Asia), its spread to other regions must have occurred well before our detection limits at 12,500 years."

So it seems that the "Gedrosia" component rather than the "Caucasus" component might be closest to the original Caucasoid genetic variation. The relative proximity of the "Gedrosia" component to the "South Asian" component might be due to tha fact that the Caucasoid race originated in northern South Asia or somewhere nearby to its north during the Upper Paleolithic rather than absorbtion of an "ASI" substrate. So as Caucasoids spread from northern South Asia to West Eurasia and North Africa during the Upper Paleolithic they differentiated from the original Caucasoid due to drift rather than absorbtion of another substrate (only Neanderthals lived in West Eurasia during the relevant era) and also this alone may explain why the rest of the Six are more distant to the "South Asian" component than the "Gedrosia" component is (this also explains their lower haplotypic diversity), and the "Gedrosia" component might have no "ASI" genetic influence.

Dmytro said...

I don't know to what extent your theory differs from that of Eupedia's Maciamo. He would like to associate the spread of IE northward with the expansion of an R1b people as main carrier. Leaving aside the linguistic difficulties (linguists have basically not accepted the Gamkrelidze-Ivanov thesis and are not likely to change their minds on purely inguistic grounds), the Maciamo version cannot be co-ordinated with archaeological data. He relies on the Maikop culture as "main spreader" of IE northward, but this is impossible, since Yamna does not derive from Maikop, and the cultures which do (Lower Mykhajlivka, Kemi-Oba,Usatovo) are partly dead ends and partly absorbed by Yamna. But really is it that important to insist on a Near Eastern origin for IE? Its later spread from a Black Sea-Caspian homeland only repeats the process of fusions and admixtures you've postulated quite correctly for human gene flows.

Dienekes said...

I don't know to what extent your theory differs from that of Eupedia's Maciamo.

I have no idea who that is or what his theory is. R1b may have been involved in spreading some Indo-European languages, but this was a secondary thrust.

He relies on the Maikop culture as "main spreader" of IE northward, but this is impossible, since Yamna does not derive from Maikop, and the cultures which do (Lower Mykhajlivka, Kemi-Oba,Usatovo) are partly dead ends and partly absorbed by Yamna.

Stanislav Grigoriev has created and archaeological model that aims to be the couterpart to Gamkrelidze and Ivanov's linguistic theories. So, I wouldn't discount the possibility of Near Eastern influence in the Pontic-Caspian steppe out of hand.

Personally, I don't see any strong reason to associate the Yamna culture with any type of Indo-European. The Indo-Europeanization of the steppe may have began from either (1) the Balkans, especially as part of the abandonment of Tripolye sites, all the excess population must have gone somewhere, or (2) the trans-Caspian area, with part of the population of the BMAC heading north after desertification commenced in the 2nd millennium BC.

We'll probably never know, since the first Indo-Europeans on the steppe are attested only in the first half of the millennium BC, too late to make any inferences about events 2,000 years earlier.

But really is it that important to insist on a Near Eastern origin for IE?

Nothing is "important", at present it is the best available model.

bmdriver said...

can anyone show me any articles that prove ANI to be closer to west asia and not ASI!

Onur said...

To put my above post in racial terms, I think the "Gedrosia", "Caucasus", "Mediterranean", "North European" and "Southwest Asian" components are all purely Caucasoid components while the "Northwest African" component is mainly Caucasoid with some Negroid admixture. "Gedrosia" is the one closest to the original Caucasoid genetic variation and the rest of the Six are more derived. I think northern South Asia and the immediate regions to its north and northwest is the homeland, i.e., urheimat of the Caucasoid race. As for the "South Asian" component, I think it is mainly "ASI" with some Caucasoid admixture.

Dienekes said...

We've known that ANI is closer to West Eurasia than to ASI ever since the publication of Reich et al. (2009). He estimates the ANI-ASI divergence in Fst units as 0.092 (Note S3 Figure 2)

Dienekes said...

might be due to tha fact that the Caucasoid race originated in northern South Asia

That is nonsense. It is archaeologically clear that Balochistan was colonized by agriculturalists from the Near East. There is absolutely no evidence that Caucasoids originated in northern South Asia. If they did, then they would not have been as divergent as they are from Ancestral South Indians, and they would have been much closer to them genetically than to Mongoloids. Also, this would require a veritable march to the West on their part, as if they had a teleological attraction to the setting sun. In short, there is no evidence at all for what you are proposing.

Onur said...

Caucasoids originated during the onset of the Upper Paleolithic at the latest. Cro-Magnons were Caucasoid. As for the situation in South Asia and environs, deserts were much larger and insurmountable there, as elsewhere, during the last glacial period than during the Holocene, that is probably what hindered Caucasoid interaction with the other races, including the ASI, for so long.

Dienekes said...

The Gedrosia component, or any other, is irrelevant to where Caucasoids ultimately originated, because, as we have seen, the Six became differentiated long after the first appearance of Caucasoids in the anthropological record. So, if you want to argue that Caucasoids originated in north South Asia, that's fine, but the genetics provide no evidence for your theory.

n/a said...

Anyone wanting a discussion of the actual "best available model" can read David Anthony's book. A "Near Eastern origin for IE" is not in serious contention.

As for the broader question of the peopling of western Eurasia: it's clear recent migrations have been important and most populations are probably composites; it's not clear that ADMIXTURE at this particular K with this particular data set is usefully decomposing western Eurasian populations into their constituent elements. Based on the available ancient DNA evidence, I expect some of the true constituents will turn out to be much more divergent. It's also clear the Caucasus has experienced gene flow from both directions and certainly doesn't harbor some untouched ur-population.

I'd again suggest looking at the behavior of the intra-western-Eurasian components as K is increased in the supplementary data from the Metspalu India paper. Possibly, none of the intra-WEurasian components will be very meaningful at any K (in the sense of representing ancient populations), or possibly more interesting results will appear if K is increased beyond 18.

The clearest answers will involve examining the phylogeography of individual autosomal segments the way we currently do for the Y chromosome and mtDNA. This will probably happen within the next few years.

Onur said...

The Gedrosia component, or any other, is irrelevant to where Caucasoids ultimately originated

The Metspalu et al. equivalent of the "Gedrosia" component (k5) has the highest haplotype diversity among the Caucasoid components wherever it is found. This argues in favor of its ancientness compared to all the other Caucasoid components.

the Six became differentiated long after the first appearance of Caucasoids in the anthropological record.

When they differentiated is far from clear.

Onur said...

Also, I do not argue that the "Gedrosia" component or Caucasoids in general necessarily originated in northern South Asia. Somewhere nearby in West Asia or Central Asia is equally probable.

Dienekes said...

David Anthony's model makes a big deal about the inclusiveness of the Indo-Europeans (the guest-host relationship) and proposes that the genetic signal of the PIE will dissipate as different groups adopt the PIE "package". As such, it is a poor fit to what you seem to favor, namely that there has been substantial genetic inflow into Europe from the direction of the steppe.

Dienekes said...

The Metspalu et al. equivalent of the "Gedrosia" component (k5) has the highest haplotype diversity among the Caucasoid components wherever it is found.

k5 is not equivalent to Gedrosia, it is rather equivalent to the Caucasus-Gedrosia component that appears at K lower than the K=12 used here. As such, it provides no support for the origin of Caucasoids in northern South Asia as you propose.

Dienekes said...

Also, the fact that West Asia is ultimately involved in the Indo-European phenomenon is strongly suggested by an examination of the two non-Indo-European populations of Europe, Basques and Finns, that differ from their neighbors in terms of the Caucasus component:

French 8.4%
French_Basque 0.1%
Spaniards 9.6%
IBS 8.9%

FIN25 1.6%
Finnish_D 1.5%
Lithuanians 6.7%
Lithuanian_D 6.6%
Russian 8.2%
Russian_D 9.6%

Note that Swedish_D and Norwegian_D have similar low % of the Caucasus component as Finns, but 6.1/6.4% of the related Gedrosia component which is 0.3-0.6% in Finns.

These contrasts are indicative that non-Indo-European populations in Europe have been less affected by the whatever Caucasus-Gedrosia represents than Indo-European ones.

Onur said...

As I said, South Asia is just one of the probable places of origin for the Caucasoid race, so there is no point in focusing on that part of my proposal so much.

The relationship between the "Caucasus" and "Gedrosia" components are pretty close, that is why they usually clump together in lower Ks in ADMIXTURE analyses.

princenuadha said...

"West Asians moving to Europe and Europeans to West Asia throughout history"

This is why we don't know if these K's were "real" populations.

As for the Caucasus dip in the basque and finns, can't that be explained by the neolithic? (Not to mention that IE could have caused the differentiation of caucasus in those places but that doesn't mean the IE were caucasus dominant)

More importantly is that the overall pattern of Caucasus/Ged mirrors the neolithic spread in Europe even to the point it runs counter to the IE spread. Such as there being more Cuacasus in s Italy than n Italy and the fact that Caucasus is about equal in Spain, France, and Germany even though IE probably came to Spain from France.

On the other hand NE drops from North Italy to South Italy (like the spread of italic), drops from France to Spain (like the spread of Celtic), and drops from the French to the Basque which is consistent with the spread of language. (Note the finns were already high in NE)

If you look at the larger context of the spread of IE, not just the baswue and finns, it seems to be more correlated with NE.

South Central Haplo said...

Why does the KLMNOP radiation is much more to east?. This theory may not support R1a dispersal. There may be an intermediate concentration/dispersal for R1a to the east.

Dmytro said...

Stanislav Grigoriev has created and archaeological model that aims to be the couterpart to Gamkrelidze and Ivanov's linguistic theories. So, I wouldn't discount the possibility of Near Eastern influence in the Pontic-Caspian steppe out of hand.

Grigoriev's views (published in Russian in 1999 and translated into English in 2002) have remained quite marginal. Anthony consulted him BTW but did not accept any of his contentions (some actually arguable but most rather bizarre). The case for Near Eastern influence is much stronger in connection with your next comment.

The Indo-Europeanization of the steppe may have began from either (1) the Balkans, especially as part of the abandonment of Tripolye sites, all the excess population must have gone somewhere

Substitute "Lower Dnipro" for "Balkans" and a case can be opened. It still needs a lot of work, but is not dismissable out of hand. The discovery of continuous exogamous relationships between steppe and Trypilia, plus the thousand year steady peaceful frontier (while Old Europe was collapsing further south) is surely meningful. You are one of the few who see the importance of explaining the mysterious "disappearance" of the large classical Trypilian population. Perhaps Dr. Nikitin's pending aDNA studies will help to solve the mystery. Especially since it is the incapacity of the traditional Gimbutas (amended post-Gimbutas really) position to explain the IE push westward which remains its major weakness.

Michael Russell said...

I'm wondering if a more precise title would be 'womb of races', rather than 'womb of nations'. Does 'nations' imply a political element that is not the focus of your inquiry?

Dienekes said...

I am using nation in the antique sense (with its etymology related to birth), rather than in a political sense.

bmdriver said...

Reich et al never said that there was any flow of west Eurasian gene into South Asia. To prevent such misleading readings of their article, they explicitly mentioned “These results do not mean that the Indian groups descend from mixtures of European and Austro-Asiatic speakers, but only that they derivefrom at least two different groups that are (distantly) related toCEU and Santhal.

Reich et al “The most remarkable feature of the PCA is a gradient of proximity to western eurasians, an analogous PCA in Europeans did not produce a gradient of
proximity to India”

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2842210/figure/F4/

-the chart gives no timeframe.

Reich et al “the fact that different Indian groups have inherited different proportions of
ancestry from the ‘Ancestral North Indians’ (ANI) who are relatedto western Eurasians, and the ‘Ancestral South Indians’ (ASI).”
It does not say that different Indian groups have inherited different proportions of ancestry from the CEU.

Sahoo et al had actually written the following words:“The perennial concept of people, language, and
agriculture arriving to India together through thenorthwest corridor does not hold up to close scrutiny.Recent claims for a linkage of haplogroups J2, L, R1a,and R2 with a contemporaneous origin for the majority of
the Indian castes’ paternal lineages from outside the
subcontinent are rejected, although our findings do support a local origin of haplogroups F* and H.” (p. 847).They also rule out arrivals from Southwest Asia becauseWest Asian haplogroups (like Y-Hg G) are not found in India

''The researchers found that Indian populations had more genetic diversity - which gives an indication of the age of a population - than either Europeans or East Asians.'' 2011. A six year project mapping human genome.The results from the Genographic Project are published in the journal Molecular Biology and Evolution.

Dienekes said...

Don't call it "West Eurasian" if it makes you feel better, but that does not change the fact that ANI is genetically much closer to Europeans and West Asians (i.e., West Eurasians) than it is to ASI.

South Central Haplo said...

So It makes you feel better calling "West Eurasian" Dienekes?.

Dienekes said...

So It makes you feel better calling "West Eurasian" Dienekes?.

Get a map and you'll agree with me.

Kurti said...

ANI has completely a West Eurasian character. it is somehow even closer to West Asian as Southwest Asian is to it. And you cant change this fact just cause of some .....centric agenda.

Vasishta said...

Hi Dienekes. A few questions;
- Is the Gedrosia component a by-product of ancient genetic drift? We know that the same South-Asian populations that previously exhibited West-Asian admixture now display Gedrosia, and in many groups, a non-trace amount of Caucasus admixture. Are both these components somehow descendants, of, say, a proto-West Asian component?
- I also notice that there is slight variation as far as the Gedrosia vs Caucasus component are concerned among individuals belonging to the same ethnic group in Indian_D. For instance, one individual will have more Caucasus admixture than the other, but that other individual also happens to have more Gedrosia than the individual who has more Caucasus than him/her. What can explain this phenomenon? Is it possible that the closeness of the components might have resulted in some sort of recombination that has made the Caucasus admixture seemingly less? Is there some selective inheritance or "luck" involved?
- You mention that the Gedrosia component is shifted towards the South Indian-modal component. Is this because the Gedrosia is a composite between a West-Asian and South Indian element, or is it because a Gedrosia-like component contributed to the West-Eurasian fraction of the South Asian (which peaks in the Metspalu et al. Tamil Pulliyar) component? While the "South Asian" component is certainly more ASI-like upon the appearance of the Gedrosia component, it still seems to have a West-Eurasian element subsumed under it.
- Finally, in light of the above, is Gedorsia a purely West-Eurasian component?

Kurti said...

I believe Gedrosia is one of the main Aryan elements and is made up of 25% North European, 60% Caucasus and 15% ANI from Dodecad v3.

6a2dab90-27b3-11e1-adc0-000bcdcb471e said...

Where are you getting that composition of Gedrosia? And is the North Eurasian component Western or Eastern? Didn't dodecad v3 show it to be western(which made me think it is neolithic or from the Caucaus or maybe it is from tocharians who were part of the kushans)?

Vasishta what type of phenotype would the West Eurasian components in pakistan/afghanistan produce assuming no asi(and later west eurasian/non indo-iranian/scythian intrusions) was present? Basically what would the population look like if it was composed of Indo-Iranian/BMAC/IVC/Tocharian/Scythian ancestry only?

6a2dab90-27b3-11e1-adc0-000bcdcb471e said...

Also i read somewhere the North Eurasian component in some Haryanvi jatts was 29%(about 22% western, 7 eastenrn). is this true? This is ridiculous. This imprlies it might have been the mjority component at some point which also seems ridiculous. Is it actually an indo iranian component or indo iranis assimilating north eurasians or both?

eurologist said...

Onur:

So it seems that the "Gedrosia" component rather than the "Caucasus" component might be closest to the original Caucasoid genetic variation. The relative proximity of the "Gedrosia" component to the "South Asian" component might be due to tha fact that the Caucasoid race originated in northern South Asia or somewhere nearby to its north during the Upper Paleolithic rather than absorbtion of an "ASI" substrate. So as Caucasoids spread from northern South Asia to West Eurasia and North Africa during the Upper Paleolithic they differentiated from the original Caucasoid due to drift rather than absorbtion of another substrate.

Yeah, something like that. I have been saying for years that the subcontinent was effectively divided into two parts due to extremely arid conditions after ~70,000 ya til ~55,000 ya, and then again ~40,000 ya til ~15,000 ya. Thus, the progenitors of ANI and Caucasoids likely formed in the NW (of the subcontinent), where they (i) developed a colder temperature cultural package and (ii) re-developed big-game hunting skills when the wetter conditions after ~55,000 ya formed extensive grasslands/ savannas. Anything west of this is primarily and firstly closely related because it originated here, ~50,000 ya. Yes, there was additional gene flow post LGM / agricultural that affected most everyone, but the foundations were put down long before.

This is not inconsistent with the Fst data if we are a bit more generous with the timings:

Say, ooA had two waves, one ~115,000 and another 85,000 ya (must be before 70,000 due to climate and SE Asian diversity). The average is 100,000 ya. If we assign this to Siberian or SE Asian vs. an average of West and East African (both slightly tainted), we get ~100,000y/0.2Fst or ~500,000y/Fst. This would date the ~0.1 Mediterranean, SW Asian, or NW African Fst distance from S Asia to 50,000 ya - consistent with the above model. Or 43,500 for Northern Europe - excellent fit to archaeology. 21,000 years between the Caucasus and Northern Europe: agricultural exchange shows itself and appears to shorten the true time distance of part of the background. Mediterranean and Northern Europe: 23,000 years - sSeparated during LGM - check. Etc.

Dienekes:

Also, the fact that West Asia is ultimately involved in the Indo-European phenomenon is strongly suggested by an examination of the two non-Indo-European populations of Europe, Basques and Finns, that differ from their neighbors in terms of the Caucasus component.

But that could be a side effect of admixture during initial agricultural expansion. It doesn't exclude the Balkans, or more generally the SW shores and inland areas of the Black Sea being the IE urheimat.

Kurti said...

@6a2dab90...
The answer is simple. Just look at the v3 admixture of Brits. They had 7% West Asian. And now they have 3% Caucasus and 8.5% Gedriosia. So 4% of West Asian moved into Gedrosia but still there are 4% whichj can only be explained with some West European and the minor 0.5%(South Asian) they had. 0.4% of it disappeared probably cause it was more Ani like and has fallen into Gedrosia.

Take the Kurdish samples. They had on v3 10-15% West and East European but now 7% what can be explained with that some of it (probably West European) moved into Gedrosia. While a good chunk of former West Asian moved into Gedrosia, some of the former Mediterranian found in Near East moved into Caucasus component.

So it is obvious for me that Gedrosia is made of mainly by West Asian of the former ADMIXTURE v3 with strong (25%?) input of West European, and some paleolitic Central-South Asian ANI(15%).

i believe the best place of origin for this component is somewhere between Kazakhstan and Pakistan (BMAC?):

Between you cant compare this component to any generel phenotype of a country. Even though Baluch, Burusho got the highest percentage, they still have a significant real South Asian and Southwest Asian input.

Just imagine 1000 People, 800 of them Gedrosia and 200 Southwest Asian and South Asian. obviously the 200 will have a strong influence on the phenotype.

Ar-Man said...

Dear Dienekes, firstly I want to thank you for your priceless work,
I know that you’re a supporter of Anatolian Hypothesis, do you have any clue about the autosomal frequencies(Caucasian, Mediterranean … etc ) that the Proto-IE speaking farmers had?
Thanks.

6a2dab90-27b3-11e1-adc0-000bcdcb471e said...

I know there is actual South/Southwest Asian influence.

I was speculating that the original Indo-iranians (who admixed with the BMAC and maybe IVC people) were Gedrosia, Caucaus and Atlantic-Balitc by ancestry.

I was curious on the percentages they would hav e had of those 3 components.

Also what phenotype do you THINK it would produce? We can't be sure but we can speculate.

Kurti said...

<<I was speculating that the original Indo-iranians (who admixed with the BMAC and maybe IVC people) were Gedrosia, Caucaus and Atlantic-Balitc by ancestry./

Yeah I know what you mean. The Tajik_Y are a good example. If you take out the far eastern and Sibirian and South Asian components which can be explained with historical waves of immigrations,

you get three main components Caucasus (20%) 2/7, Gedrosia (31%) 3/7 and North European (19%) 2/7. And this was probably the closest to the original Iranian admixture.

It might sound a bit boastful but the closest to this original Iranian admixture are the Kurd_D/Y (which are mainly from Iraq/Iran) and North Ossetian_Y samples in my opinion. Just that Ossetians have too much Caucasus and too little Gedorisa and the Kurds on the other hand just a bit too much Caucasus and a bit too less North European. Imagine a Kurd_Y samples with more (+12%) North European and the North_Ossetian samples with less Caucasus and more Gedrosia and you will have a picture of how the original Iranians looked like.

Now INTERESTINGLY enough outside the Iranian speakers. The best match for this admixture are LEZGIANS in NORTHEAST CAUCASUS. They fit perfectly how I imagine the genetic make up of the early Iranians. It seems to me that Dienekes theory, (at least based on Iranian speakers) the homeland of Indo-Europeans being from Northeast Caucasus fits very well.

6a2dab90-27b3-11e1-adc0-000bcdcb471e said...

That sounds ridiculous.

The Indo-European homeland is probably somewhere in Ukraine.

The Indo-Iranian homeland is most likely Central Asia (Andronovo). I see no reason to go with a West Asian hypothesis for either besides political reasons certain groups might have.

Also I don't see why Kurds and Ossetians would be the best representations of the original population giving their location. Ossetians have less R1a than their non Indo-Iranian neighbors. They also possess R1b, J1 and I at significant frequencies and lack R1a (especially Ossetians) when you compare them to Tajiks and Pashtuns. And R1b, J1, I were certainly not part of the Indo-Iranian gene pool given their complete absence in Afghanistan/Tajikistan/Pakistan/NW India. Plus Kurds have significant SW Asian/Med which are lacking in Indic speakers as well as Pashtuns and Tajiks for the most part. Given the fact that Tajiks and Pashtuns are closer to the Indo-Iranian homeland and have had contact with Indo-Iranians past the original migrations(Kushans, Scythians etc) they are probably closer to them. Their R1a frequency, and high Caucasus/Gedrosia/Atlantic-Baltic components only adds to the argument.

Dienekes said...

The Indo-Iranian homeland is most likely Central Asia (Andronovo).

Andronovo is not in Central Asia.

Kurti said...

[I]The Indo-European homeland is probably somewhere in Ukraine.[/I]

Where was here the talk about the Indo-European homeland but Iranian one.

[I]The Indo-Iranian homeland is most likely Central Asia (Andronovo). I see no reason to go with a West Asian hypothesis for either besides political reasons certain groups might have.[I]

Good you might be right but how does this change the fact that autosomaly spoken the Kurds, Ossetians and Lezgians come closer to the original Iranian genetic make up than other Indo-Iranian groups living in Central Asia today. The closest today known Iranian group to the original homeland are the Tajiks. If you take out the South Asian and East Asian genes you get similar Admixture to those Groups mentioned above.

[i]Also I don't see why Kurds and Ossetians would be the best representations of the original population giving their location. Ossetians have less R1a than their non Indo-Iranian neighbors[/I]

Just forgot about the Haplogroup distribution which can shift its frequency very simple. Even though Tajiks got 50-60% R1a1a why do they have almost the same genetic make up as Lezgians, Kurds, Ossetians etc?

And just for your information. J2a was found in all Iranian speaking Groups and is the second frequent Haplogroup.

<J2a:

Haplogroup J2 especially the subcadle J2a is frequently found among almost all groups of Iranic people. In comparison with the Haplogroup R1a1, J2 is not only restricted to geographically eastern and western Iranic populations, but also found among north-western and south-western Iranic populations such as the Bakhtiaris and Mazanderani,[68][69] as well as geographically north-western Iranic Ossetians.[70] Despite its supposed origin in the fertile crescent, J2a is also found among Iranic populations in the east such as the Yagnobi which are of Soghdian origin[71] as well as the Parsis of India.[72] Beside the relatively high percentage among the Yagnobis in Central Asia, other Iranic populations tend to have a higher frequency of J2a when compared to neighboring Turkic populations. The relatively strong presence of J2a among Ossetians as well as Yagnobis proves distant from the supposed Mesopotamian origin region of J2, are carriers of this Haplogroup.

In the Indo-Iranian context, the occurrence of J2a in South Asia is limited to caste populations, with the highest frequencies found among northern areas of South Asia.[73][74] Compared with R1a1, J2a shows a more conservative distribution, stronger limited to Indo-Iranian origin groups.[73]

http://en.wikipedia.org/wiki/Iranian_peoples

[i]Plus Kurds have significant SW Asian/Med[I]

Yes your right. I forgot to mention this. Southwest Asian can be explained with ancient and more recent historic events like the semitic waves of Akkadians and the Islamic expansion. The same for the Med component. Some of it is probably native some other might have been brought with Greek expansion. There are still huge Greek monuments in Kurdistan. Nisibis, the Region where I come from was once hugely settled by Mygdon Greeks.

[I] Given the fact that Tajiks and Pashtuns are closer to the Indo-Iranian homeland and have had contact with Indo-Iranians past the original migrations(Kushans, Scythians etc) they are probably closer to them. Their R1a frequency, and high Caucasus/Gedrosia/Atlantic-Baltic components only adds to the argument.[/I]

No one is denying the Iranicnes of Pashtuns or Tajiks but you must admit that those both are more mixed with not so much Caucasian components. I am simply saying that Lezgians are genetically closer to what the original Iranians (be it from Central Asia) were than Tajiks today are.

Scythian, Cimmerian and Alan presence can be found in all of the Caucasus, North Iran and East Anatolia. And not to forget the Mitannis.

Ar-Man said...

The Indo-European homeland is probably somewhere in Ukraine.

The Indo-Iranian homeland is most likely Central Asia (Andronovo). I see no reason to go with a West Asian hypothesis for either besides political reasons certain groups might have.


When Ukraine was transported to Central Asia?

Waggg said...

"Andronovo is not in Central Asia."

The Andronovo horizon, is located in central Asia. And its ancestral cultures are just south of the Urals (starting around 2300 BCE).

Kurti : "how does this change the fact that autosomaly spoken the Kurds, Ossetians and Lezgians come closer to the original Iranian genetic make up than other Indo-Iranian groups living in Central Asia today"

Irrelevant. Language shifting can occurs via the impulse of a dominant minority as seen several time in history. Besides they are kind of marginal given their history and location. Weird to take them as decisive examples (Besides, Kurds have more R1a1a than their neighbors).
Anyway, are modern Turks of Anatolia autosomally close to the Yakuts, the Kyrgyz or the Uyghurs?

Archeology and linguistics (*) favor an origin of Indo-iranian (proto-indo-iranian) close to the Urals (Russia/Central Asia) and the haplogroup to consider is thus R1a1a (J2a is very rare in this region). Indo-Aryan origin is to be found there too, despite early occurences in west Asia (presence of Indo-aryan words linked to horses and chariotry (indo-aryan deities as well) in the bronze age Mitanni kingdom, governor suardata at Qiltu near Jerusalem in the same time frame (apparently from indo-aryan "svardata"), and possible indo-iranian loanwords in the Kassite Dynasty in Babylonia), not only Iranic (in UGRIC languages their ancient indo-iranian input have matches mostly or only in indo-aryan (e.g. Hungarian tehén (cow) is way closer to sanskrit dhenu than to Avestan dainu)).

(*) example 1

(*) example 2 (go to page 385 in the book (197 to type in the search field))

This coupled with the fact that scythian-like populations were clearly iranic speaking, it's a BIG smoking gun.

princenuadha said...

" And, to the north, from either the Balkans, the Caucasus, or the trans-Caspian region, they would have met the last remaining Proto-Europeoid hunters of the continental zone, becoming the Northern Europeoids who once stretched all the way to the interior of Asia."

"North European" peaks in Northeastern Europe. A recent craniological study showed that the neolithic transition in this area happened by indigenous hunter gathers adopting neolithic practices, not middle eastern migrants.

Unless there was a more recent and significant migration, "North European" probably represents Mesolithic European heritage.

And about the indo Europeans, should we be assuming that the basques would have much less IE heritage? We know they have a lot of r1b and that r1b likely got there post neolithic.

Dienekes said...

"North European" peaks in Northeastern Europe. A recent craniological study showed that the neolithic transition in this area happened by indigenous hunter gathers adopting neolithic practices, not middle eastern migrants.

That is quite irrelevant, since it cannot be assumed that present-day Northeastern Europeans are descended primarily from any particularl group of people. It does seem, on the basis of mtDNA, that they are more descended from the Mesolithic mtDNA-U bearing stratum, but not exclusively by any means.

Unless there was a more recent and significant migration, "North European" probably represents Mesolithic European heritage.

That is inaccurate, both because there is no European population that is almost entiirely made up of mtDNA U as the Mesolithic populations that have been sampled, and also because the North European component has not diverged (in terms of Fst) to any significant degree from the other Caucasoid components, and hence cannot have been derived to a substantial degree from populations that went their own way much earlier than the Neolithic.

And about the indo Europeans, should we be assuming that the basques would have much less IE heritage? We know they have a lot of r1b and that r1b likely got there post neolithic.

We are not assuming anything, we are observing that the Basques do not possess some components found in their Indo-European neighbors.

There is no reason to assume that R1b has anything to do with the Proto-Indo-Europeans (its near absence in peninsular India, and paucity in Pakistan precludes that), although some of it may have been carried by secondary movements of Indo-Europeanized groups.

princenuadha said...

>because there is no European population that is almost entiirely made up of mtDNA U as the Mesolithic populations that have been sampled

A significant shift in mtdna frequencies does not imply a significant shift in admixture. Also, meso Europe did have non U.

I guessed that the northeast Europeans were mostly meso European because I havent seen (much) evidence for a shift since they were established as still being "meso" after the neolithic revolution. If there were a admixture shift I would have expected good evidence for a non meso E migration... But I don't.

> the North European component has not diverged (in terms of Fst) to any significant degree from the other Caucasoid components

I don't know how to estimate times but the close distance can come from recent divergence and/or gene flow (in both directions). Your argument is that there is a parent population that gas been largely preserved, ie the caucasus component. This implies a largely one way gene flow and i don't see how you can argue that for Northeast Europe without the neolithic impact.

> We are not assuming anything, we are observing that the Basques do not possess some components found in their Indo-European neighbors.

You are when you argue that info Europeans were mostly of that component.

> There is no reason to assume that R1b has anything to do with the Proto-Indo-Europeans

What i was tying to say is that I think the indo Europeans are a good candidate as to who brought r1b to the basques. If that's what happened then we could expect the basque to show IE admixture.

Dr Rob said...

Dienkes; were these calculations based on autosomal data or Y haplogroups ?

Dienekes said...

autosomal