October 23, 2009

140,000-year divergence time between Eurasians and West Africans

From the paper:
Perhaps our most interesting demographic results are the inferred divergence times. Other studies [11],[12] have estimated divergence times between Europeans and East Asians similar to the ≈23 kya we infer. Interestingly, archeological evidence places humans in Europe much earlier (≈40 kya) [1]. Our inferred divergence time of ≈22 kya between East Asians and Mexican-Americans is somewhat older than the oldest well-accepted New World archeological evidence [2]. The divergence we infer may reflect the settlement of Beringia, rather than the expansion into the New World proper [14]. Finally, the divergence time of ≈140 kya we infer between African and Eurasian populations is consistent with archeological evidence for modern humans in the Middle East ≈100 kya [1], but it is much older than other inferences of ≈50 kya divergence from mitochondrial DNA [1]. This discrepancy may be explained by our inclusion of migration in the model. Migration preserves correlation between population allele frequencies, so an observed correlation across the genome can be explained by either recent divergence without migration or ancient divergence with migration. In fact, the African-Eurasian migration rate we infer of ≈25×10−5 per generation is comparable to the ≈100×10−5 inferred from census records between modern continental Europe and Britain [55].

One difficulty in interpreting our divergence times is that the sampled populations may not best represent those in which historically important divergences occurred. For example, the Yoruba are a West African population, so the divergence time we infer between Yoruba and Eurasian ancestral populations may correspond to divergence within Africa itself. Future studies of more populations [56]–[58] will help alleviate this difficulty.
Some comments:
  • The 140ky divergence time between Eurasians and West Africans is consistent with my idea of Africans being divided into Palaeoafricans and Afrasians. The 140ky gap includes both divergence since Eurasians left Africa, as well as divergence between ancestral Eurasians ("Afrasians") and earlier African populations ("Palaeoafricans'). It remains to be seen whether the bulk of the 140ky occurred in Africa itself (in which case conventional Out of Africa c. 40kya is affirmed), or outside Africa (in which case Out of Africa c. 100kya, evidenced by early modern skulls acquires a new significance).
  • East Asian - Native American divergence of 22ky predates the known settlement of the Americas. But, this is not a problem, since we must split the 22ky into 15ky (since the settlement), and 8ky, which is reasonable time for divergence between the ancestors of the Chinese and the populations that headed north, and perhaps stayed in Beringia before crossing into the Americas.
  • East Asian - European divergence of 23ky suggests, in agreement with recent mtDNA evidence, that Europeans are not primarily descended from the earliest (or even the latest) Upper Paleolithic population of the continent. This seems to also be in agreement with Y-chromosome evidence that suggests interesting East-West connections in Eurasia that are not easily explained by a simple early divergence model with no subsequent link between east and west (e.g., western E vs. eastern D, western R vs. eastern Q, northern N vs. eastern O, etc.). It is also in agreement with the mtDNA picture (West Eurasian N dominance, East Eurasian mixed N/M).
Naturally, both the increased age of divergence of 140ky (compared to other studies), as well as the wide confidence intervals suggest that we must always treat genetic age estimates with the utmost caution.

PLoS Genet 5(10): e1000695. doi:10.1371/journal.pgen.1000695

Inferring the Joint Demographic History of Multiple Populations from Multidimensional SNP Frequency Data

Ryan N. Gutenkunst, Ryan D. Hernandez, Scott H. Williamson, Carlos D. Bustamante

Abstract

Demographic models built from genetic data play important roles in illuminating prehistorical events and serving as null models in genome scans for selection. We introduce an inference method based on the joint frequency spectrum of genetic variants within and between populations. For candidate models we numerically compute the expected spectrum using a diffusion approximation to the one-locus, two-allele Wright-Fisher process, involving up to three simultaneous populations. Our approach is a composite likelihood scheme, since linkage between neutral loci alters the variance but not the expectation of the frequency spectrum. We thus use bootstraps incorporating linkage to estimate uncertainties for parameters and significance values for hypothesis tests. Our method can also incorporate selection on single sites, predicting the joint distribution of selected alleles among populations experiencing a bevy of evolutionary forces, including expansions, contractions, migrations, and admixture. We model human expansion out of Africa and the settlement of the New World, using 5 Mb of noncoding DNA resequenced in 68 individuals from 4 populations (YRI, CHB, CEU, and MXL) by the Environmental Genome Project. We infer divergence between West African and Eurasian populations 140 thousand years ago (95% confidence interval: 40–270 kya). This is earlier than other genetic studies, in part because we incorporate migration. We estimate the European (CEU) and East Asian (CHB) divergence time to be 23 kya (95% c.i.: 17–43 kya), long after archeological evidence places modern humans in Europe. Finally, we estimate divergence between East Asians (CHB) and Mexican-Americans (MXL) of 22 kya (95% c.i.: 16.3–26.9 kya), and our analysis yields no evidence for subsequent migration. Furthermore, combining our demographic model with a previously estimated distribution of selective effects among newly arising amino acid mutations accurately predicts the frequency spectrum of nonsynonymous variants across three continental populations (YRI, CHB, CEU).

Link

35 comments:

Gioiello said...

Anyway 23kya between Europeans and East Asians falsified completely the theory of Vizachero of a recent immigration of hg. R to Europe. But we must see which are Europeans and which are East Asians. Like Amerindians sejourned in Beringia, "Europeans" can have sejourned somewhere before reaching Europe. And which is the mix of Y and MtDNA that determined those results?

Aaron said...

I'm still puzzled at the lack of N descending mtDNA among Amerindians. Would this imply a pre-N mtDNA expansion by yDNA Q from the subcontinent of India?

Jean said...

"Anyway 23kya between Europeans and East Asians falsified completely the theory of Vizachero of a recent immigration of hg. R to Europe."

Why? Movement of people within Western Eurasia, such as Y hg R, has nothing to do with the divergence between what they call "Europeans" (i.e. Western Eurasians) and East Asians e.g. Chinese.

The modern political border between Europe and Asia is not a deeply significant genetic dividing line.

As it happens I suggest a Copper/Bronze Age movement of hg R from SE Europe into Central and Western Europe, while other people think that part of that movement (R1b) could have come from Anatolia. No great distance geographically or genetically.

Aaron said...

"We estimate the European (CEU) and East Asian (CHB) divergence time to be 23 kya (95% c.i.: 17–43 kya)"

-This is pretty reasonable I think. Though I think so-called 'European' can and should be superimposed with pre-Neolithic West/Central Asian.

Vincent said...

Why? Movement of people within Western Eurasia, such as Y hg R, has nothing to do with the divergence between what they call "Europeans" (i.e. Western Eurasians) and East Asians e.g. Chinese.

I agree.

For one thing, this study was not limited to only the Y and we know that different parts of the genome can (and often do) have different phylogeographic histories.

For another, as far as the Y goes, there do indeed exist some very distinct differences as you move across Eurasia. There is very little R1 of any sort in East Asia, for example, and all the evidence points to an R1/R2 split that post-dates this 23 kya estimate of East Asia/European divergence.

That the split observed in this paper may be more accurately labeled a split between W. Asia and E. Asia is an interesting twist which this paper does not seem designed to investigate.

VV

Gioiello said...

Jean, you forgave to quote the rest: “But we must see which are Europeans and which are East Asians. Like Amerindians lived in Beringia, "Europeans" can have lived somewhere before reaching Europe. And which is the mix of Y and MtDNA that determined those results?”

I did mean that, as Hg.R (a and b) is massive in Europe, if its arrive is so recent, the distance between European and East Asians should have been lower and I put the problem of mtDNA. Anyway it is very strange that a Y haplogroup have come from East without women. We should hypothesize a colonization like that of America from Europeans.

Aaron said...

"And which is the mix of Y and MtDNA that determined those results"

--At the very least Y-HG R and mtDNA U, IMHO. For all intents and purposes HV and JT seem to arrived into the typically Eurasian profile a bit later. Whether this ties in exactly with the spread of Neolithic is hard to say. There is no reason why it has to be cut-and-dry. The Eurasian "isolates" usually have very high numbers of U and R1.

Maju said...

I find it all kind of inconsistent with the archaeological data. The limited array of population samples may be distorting the results a lot.

Regardless of the extremely recent 23ky age for NW Europeans (CEU) in relation to Northern Han (CHB), and its oddly similar age to Amerindians, the most striking oddity would be 140 ky age for the Yoruba/Eurasian divergence, which would only fit (and rather forced) with the earliest possible OOA dates. This would be highly incompatible with the coastal migration model via South Arabia.

I think the research needs a reality check.

Also the Euro-EA divergence age seems to be taken from "other studies" and not be part of this one specifically.

I'm still puzzled at the lack of N descending mtDNA among Amerindians.

Amerindians have about half of their mtDNA derived from N: A, X2 and B. So no idea where you get that concept.

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

Argh. I mean R. Can't seem to get it right.

Aaron said...

I see B is also under that umbrella, but the lack of U, HV, et al. That was my point. Would that imply that these haplogroups were geographically distinct or perhaps not around at the time of the movement Q over the Bering Strait. It's an interesting divide either way.

Maju said...

Aaron: mtDNA R0 is a West Eurasian lineage and Y-DNA Q is not. While I suspect that Amerindian X2 is related to the spread of Q, this one seems to have gone through its own mostly independent dynamics in Siberia. The mtDNA lineages were surely "picked" at some point in this process, maybe already in Beringia or not too far away in any case. The population logically went through its own episodes of founder effects and drift, so I don't find the Amerindian haploid genetics really diffiult to understand: it just points us to the relatively early phases of the colonization of NE Asia and its complexities in a context of very small population units, as is normal in such extreme environments.

Lamprecht said...

Full study PDF:
http://arxiv.org/PS_cache/arxiv/pdf/0909/0909.0925v1.pdf

Ebizur said...

Note that the authors of this study have not used any sample of Native Americans, but rather a sample of Mexicans from Los Angeles (n=22), who have been determined by the authors to be of approximately 50% European biogeographical ancestry on average.

Jean said...

Gioiello said...

"Jean....I did mean that, as Hg.R (a and b) is massive in Europe, if its arrive is so recent, the distance between European and East Asians should have been lower..."

No. Y-Chromosome haplogroup R is Western Eurasian. It has always been Western Eurasian, as far as we can tell. It has nothing at all to do with the genetic differences between Western Eurasians and East Asians, such as Chinese, Japanese, Cambodians and Tibetans.

Don't be fooled by the way that this paper uses the label "Europeans". They mean Europe and Western Asia.

eurologist said...

At a reasonable confidence level, the results are consistent with Europeans largely isolated and derived from before LGM.

terryt said...

"It remains to be seen whether the bulk of the 140ky occurred in Africa itself (in which case conventional Out of Africa c. 40kya is affirmed), or outside Africa (in which case Out of Africa c. 100kya, evidenced by early modern skulls acquires a new significance)".

I've always assumed the presence of 100k skulls outside Africa was significant. In fact it has always amazed me that they are dismissed as being part of an OoA movement that lies at the base of modern human ancestry. Can anyone offer a suggestion as to why they are so readily dismissed?

"This would be highly incompatible with the coastal migration model via South Arabia".

Here we go again. Aren't you making a huge assumption here?

"Would that imply that these haplogroups were geographically distinct or perhaps not around at the time of the movement Q over the Bering Strait".

Geographically distinct would be my bet. U, HV, et al. probably emerged from Northwest India and B moved up the East Asian coast. That's part of my reasoning for placing their ultimate common origin in island SE Asia rather than in S Asia.

"Anyway it is very strange that a Y haplogroup have come from East without women".

Why is that strange? It has happened often. Maju mentions the phenomenon in regard to America: 'The mtDNA lineages were surely "picked" at some point in this process, maybe already in Beringia or not too far away in any case'.

Maju said...

Something called my attention suddenly:

... in which case conventional Out of Africa c. 40kya is affirmed...

I was always puzzled when some people claimed to see such 40kya date for OOA and wondered where could they get such idea, when European colonization is already of that date (and Europeans are highly derived within Eurasians). Now I know: Dienekes is the one spreading such misconception! And worse: describing it as "conventional".

Unnacceptable, sorry.

Maju said...

Note that the authors of this study have not used any sample of Native Americans, but rather a sample of Mexicans from Los Angeles (n=22), who have been determined by the authors to be of approximately 50% European biogeographical ancestry on average.

Unbelievable! What is wrong with scientists today?!

Why is that strange? It has happened often. Maju mentions the phenomenon in regard to America: 'The mtDNA lineages were surely "picked" at some point in this process, maybe already in Beringia or not too far away in any case'.

But I don't percieve the male and female lineages as being two different populations just because one has some affinities with the west and the other not. Y-DNA Q is in any case NE lineage, at least by frequency and not any West Eurasian lineage, regardless that it's a cousin of R.

Probably all that lineage shaking happened between Altai and East Siberia/Beringia, among tiny bands of pioneering subartic hunters and fishermen, what would cause lineages to end up in nearly any combo.

It's futile to attempt to assign ethnic entities to lineages like R or Q that must have been around since 40 or more milennia ago.

mathilda said...

"I was always puzzled when some people claimed to see such 40kya date for OOA and wondered where could they get such idea, when European colonization is already of that date (and Europeans are highly derived within Eurasians)."

Aargh! Don't get me started :(

Theres a population movement into North Africa and the near east about 125/130k ago, I guess the seperation of West Africans would have been a little before that time.

Just why DO people believe that the remains in NA and the NE were dead ends?

"This would be highly incompatible with the coastal migration model via South Arabia".

Yep- I've always believed up the Nile was much more likely. None of the migrations that came in that area until the late neolithic had anything to do with the gate of Tears, I'm constantly bemused as to why everyone obsesses on it as the OOA route anyway.

terryt said...

"Y-DNA Q is in any case NE lineage, at least by frequency and not any West Eurasian lineage, regardless that it's a cousin of R".

True. But before the mutations that separated them they obviously lived in the same place. So is R an NE lineage?

"I've always believed up the Nile was much more likely".

Don't you mean down the Nile? Great to see you active again, by the way. Hope all is well.

Maju said...

But before the mutations that separated them they obviously lived in the same place. So is R an NE lineage?-

No. P branched out surely in some small group of Central or South Asia. That was very long ago, IMO. For me the distribution pattern of these lineages suggest a quite old age depth, so I just see their original carriers as indistinct early Eurasians, who were in the right place (near Central Asia) at the right moment (when the NE and the West were being colonized early on). So P should be like 40 or more ky old. A serious possibility is this group had something to do with Central Asian "proto-Aurignacian".

...

Also glad to see around again, Mathilda.

I'm constantly bemused as to why everyone obsesses on it as the OOA route anyway.

The main reason seems to be that there are not too many mutations between the clearly African L3 node and the clearly Eurasian M and N ones in mtDNA. So it seems to demand a quick migration to South/SE Asia, without many twists.

pconroy said...

@Mathilda

Welcome back!

German said...

Luis: "the clearly African L3 node and the clearly Eurasian M and N ones in mtDNA."

Not true. L3 is found in Africa - true, but it lacks the defining RFLP characteristic of the other two (or more depending on how you count) African lineages: "For African mtDNAs, one of the oldest and most significant polymorphisms is the HpaI site at np 3592. This site is African-specific with the oldest 2/3 of the tree having the HpaI 3592 site (+3592 HpaI) and being defined as macro-haplogroup L, with haplogroup subdivisions L1 and L2. The absence of this site is defined as haplogroup L3." (Wallace 1999, Mitochondrial DNA variation in human evolution and disease, p. 215). The same works for CR: "African L3 haplotypes are characterized by the absence of L1/L2-specific polymorphisms at nucleotide positions 769, 1018, 3594, 4104, 7256, 7521, and 13650″ (Herrnstadt et al. 2002. Reduced-Median-Network Analysis of Complete Mitochondrial DNA Coding-Region Sequences for the Major African, Asian, and European Haplogroups. Am J Hum Gen 2002 70 (5)).

M, N and L3 lineages are closer to each other than either of them to the other lineages found in Africa. This makes it likely that African L3 lineages are derived from some sublineages of M and N. These facts is further reinforced by phylogeography: L3 lineages aren't found outside of Africa, while some sublineages of M and N are found in North and East Africa. More specifically, M1 is found in North, East Africa and in the Caucasus, while L3 lineages aren't found in the Caucasus. There's no evidence for earlier M or N lineages in Africa. M and N have non-African origin. The best way to think about L3 lineages is, therefore, as an umbrella term for some derived sublineages of M and N.

Gioiello said...

German, the same anomaly is in YDNA. Why Africa has E without D? I have always supported that E is Eurasian. Probably the original Africans were YDNA A-B and mtDNA L (x L3).
Also from a linguistic point of view the original African languages are the San languages, and my great compatriot Alfredo Trombetti demonstrated a century ago that they were widespread to Central Africa ( Wa-Sandawe), then submerged not only from the recent Bantu migration but probably from a previous arrival of YDNA E (but also R1b1*).
And why we don’t find ancient skulls of Homo sapiens sapiens. We were always told by the soil nature. Probably because they weren’t.

terryt said...

"P branched out surely in some small group of Central or South Asia".

I'm sorry Maju. Perhaps I've become so used to disagreeing with you that I didn't actually carefully read your explanation before I sent my comment. You'd written, 'Probably all that lineage shaking happened between Altai and East Siberia/Beringia'. We do seem to be agreeing on maore and more.

German said...

"German, the same anomaly is in YDNA. Why Africa has E without D? I have always supported that E is Eurasian. Probably the original Africans were YDNA A-B and mtDNA L (x L3)."

This is exactly right. Once we get L3 out of the way, we'll see a good fit between Y-DNA and mtDNA. Probably around 40,000 BP, there were two sets of widely dispersed populations in the world: M and N (CD and PQ in Y-DNA) outside of Africa and L1-L2 (using an older nomenclature) (A and B) in Africa. I tend to think that African L1 and L2 lineages are ultimately derived from "eastern" non-African lineages (C, D, A, S, X, etc.) after a bottleneck with re-expansion. At least, linguistic diversity is much greater outside of Africa, with Papua New Guinea and America as standouts, and such features of African languages as clicks aren't found outside of Africa, which means they are local African developments and not an ancestral state of the human phonological system.

Ponto said...

I don't believe the paradigm that Europeans are the descendants of those AMHs like Cro Magnons or even their predecessors in Europe the Neanderthal. Europeans are the descendants of humans who came from the Middle East in many immigration waves from the period just before the Holocene and afterwards when the Neolithic farming revolution took place. All that Aurignacian, Mousterian, Chatelperonian and so on belong to those predecessors who lived and died in Europe before the Holocene immigrants arrived. That is, nothing to do with any of our European ancestors.
I don't understand how the separation of R1 and R2 predate 23 kya. R1 is only 18,500 years old. Not so old as you think it to be. R1b, R1a entered Europe about the same time from Asia and took advantage of the opening of new lands for colonisation that became available after the LGM. Their dominance in Europe (frequency) shows a total founder effect expansion, a star like pattern of haplotypes. The refuge idea is over stated. Some mesolithic populations, a very small number of humans, expanded out of the habitable parts of Europe after the LGM but not R1b or R1a. As to the separation of non African based populations from Eurasian populations, it must be remembered that AMH existed in the Middle East more than 100 kya. They supposedly went extinct but maybe they didn't, they just moved on further north towards Central Asia.

Cheddar Man is not much help. He was murdered 9 kya, so his mtDNA can only be dated to that time or a little earlier, he had a mother!
The Pagliaci Cave result of the CRS in some Cro Magnoid human, not significant as many haplogroups have the CRS in HRV1. My opinion: the result of the CRS is a lot of Italian bullshit. Out and out fraud, or contamination from all the people who handled the bone before the geneticists. Of the two I go for fraud.

Gioiello said...

Ponto writes: “AMH existed in the Middle East more than 100 kya. They supposedly went extinct but maybe they didn't, they just moved on further north towards Central Asia”.

We were two to think so (German and me). Glad you are the third.
But why you are denying that those men had a mtDNA? If they had, it can have generated L3. Why not?

Maju said...

R1 is only 18,500 years old.

Not sure but Dienekes posted recently another study that made R1b1b2a alone somewhat older than that figure.

Whatever the case this only proves how unbelievable are the MC results: each one is different.

But you and others keep claiming that it "is" that way, no doubt. Please, show at least some decency and introduce an element of doubt in such claims (and provide a source, though guess it's Karafet's popular age hunches list).

They supposedly went extinct but maybe they didn't, they just moved on further north towards Central Asia.

Or to South Asia, where there is some fossil record that could support it. Central Asia became Neanderthal territory as well in "the 60s" (i.e. c. 60,000 BP).

Anyhow I find funny that you happily argue for the unproven continuity of people who went missing 100,000 BP and left a blank for 50,000 years, and instead you argue that people whose highly continuous fossil record is well known, just vanished from the face of earth.

That is double standards, and a clear case of arguing not on reason but on a private agenda.

terryt said...

"We were two to think so (German and me). Glad you are the third".

That makes Maju fourth:

"Or to South Asia, where there is some fossil record that could support it. Central Asia became Neanderthal territory as well in 'the 60s' (i.e. c. 60,000 BP)".

And I agree with that, so I'm number six. We find 'modern' humans along the Levant coast around 100k. At a guess I'd say they actually got much further out of Africa than just that far. They probably expanded through their prefered habitat (perhaps semi-open grassland) as far as they could. They would have come up against Neanderthals in Anatolia and Iran. Did the boundary between them develop into a hybrid zone? That would be the usual thing when two subspecies meet.

We then find Neanderthals along the Levant coast around 70k. As modern humans retreated from the cold any hybrid zone had presumably also moved south, in the west perhaps even into Africa. What about the hybrid zone in the east?

We find technological continuity in South Asia right through any time frame you might care to offer for the change there from archaic to modern humans.

We also find technological continuity in the region stretching from the Altai mountains to Japan, right through any time frame you might care to offer for the change there from archaic to modern humans.

Modern humans reappear in the Levant about 50k. what might we conclude from all this? Most obviously that humans evolved as a product of a complex pattern of genetic movement. In other words there's no sudden burst of 'modern' humans OoA.

Besides, do we really know that the woman who first carried mtDNA L0 was already a 'modern' human, or are we simply assuming that she was? The situation may actually be very complicated.

Gioiello said...

Also the rests found under the Toba eruption (74kyBP) aren't explicable by OOA theory (less than 60ky).

Maju said...

Also the rests found under the Toba eruption (74kyBP) aren't explicable by OOA theory (less than 60ky).

Sorry. This is not as you say: the "OOA theory" only states that humans began their history in Africa and only later moved elsewhere. It doesn't matter if 100 or 60 thousand years ago.

The "rapid coastal migration" theory does suggest that there was a "fast" migration from East Africa to South and SE Asia. The dates most used are those of 60-70 kya but I am sure you can consider other scenarios as well.

The Jawalpuram rests under (and above) the Toba ashes could be facture of H. sapiens (and I suspect they are) but they are not confirmed as such. In any case, they are typologically most similar to those of South African (and not East African) MSA.

There are still many issues to clarify but the OOA as such stands in all possible scenarios.

terryt said...

"The 'rapid coastal migration' theory does suggest that there was a 'fast' migration from East Africa to South and SE Asia".

As far as I'm aware the Bab el Mandeb 'rapid coastal migration theory' was invented by committed 50k OoA supporters to accomodate the consequent necessarily rapid Australian colonisation. Specifically by Spencer Wells. Incidently, we're still waiting for his research on tribal groups in the Philippines:

http://blogs.nationalgeographic.com/blogs/genographic/2008/12/meeting-the-ati-peoples-in-the.html

Perhaps his results don't fit his theory. With acceptance of a date nearer 100k for OoA the settling of Australia presents no problem. They have plenty of time to reach there no matter what route they took.

Maju said...

Maybe. I understand that it has to do with mtDNA phylogenetic structure, though if L3 and L2 expanded c. 100,000 BP, then M and N can't be as young as half that age.

Anyhow this is not the OOA theory as such but a version of it, which goes into further detail about the process.