May 08, 2009

Archaic admixture in modern humans? (Wall et al. 2009)

John Hawks points me towards a new paper which aims to estimate the presence of archaic admixture (from Neandertals, Indonesian "hobbits", and assorted candidates) in the modern human genome. I highly recommend reading that post for an analysis of the paper; while Dr. Hawks is probably on the "admixture happened" side of the debate, he offers both reasons to like and dislike the new study.

There is, however, what I think is a more important weakness of the paper. To substantiate the case for admixture, the authors compare West African Yorubans separately with Europeans and with East Asians.

Let's assume that they do in fact detect what is genuine archaic admixture, i.e., the introgression of chunks of DNA into regional humans from a species other than modern Homo sapiens. They can achieve this by detecting regions where Yoruban DNA sequence differs from European DNA sequence in a manner suggestive of a very ancient time depth.

But, is there a reason to ascribe these differences to Neandertal introgression?

Actually, there is not, as Europeans are derived from a specific East African population: they are not derived from Sub-Saharan Africans in general, or West Africans in particular. Indeed, this very paper finds evidence of archaic admixture in Africa itself!

Thus, what appears as a piece of Neandertal DNA in Europeans, could in fact be a piece of ancestral East African DNA which differs from Yoruba DNA because of population structure in Africa itself, for which there is more than enough evidence.

So, what this paper does, is tell us that a chunk of DNA in Europeans, and the corresponding chunk in Yorubans don't share ancestry within a conventional Out-of-Africa time frame. It does not, however, tell us that this is because of archaic introgression in Europeans. The culprit could equally well be long-term population structure in Africa, i.e., the presence of "modern" and "archaic" populations in Africa itself.

The way forward is to compare Europeans with Middle Eastern Caucasoids and East Africans. If the "archaic" European DNA is found across many of these populations, then the case for Neandertal introgression will weaken, and the hypothesis presented in this post (ancient African population structure) will be supported.

UPDATE (May 9): John Hawks has updated his blog entry in response to this post. He makes two points:
The East Asian and European comparisons come up with different genes showing evidence of putative introgression.
What we don't know at this point is whether either (or both) of the European/East Asian introgression candidate genes are found in East Africa. If so, then the presence of different sets of genes in Europe and East Asia could be the result of random survivals of the diversity of the initial African population. Or, introgression happened in either Europe or East Asia, and one region preserves archaic DNA inherited from Africa and the other archaic DNA that introgressed in East Asia or Europe, which is why the two are different.

Furthermore, while Out-of-Africa is conventionally seen as one migration c. 40Kya which eventually spawned both Europeans and East Asians, there is Y-chromosomal evidence for a separate process linking Western Eurasia and East Africa (Y-chromosome haplogroup E). Thus, a later movement may have spread "archaic" East African genes into Western Eurasia but not in East Asia.

John's second point:
The entire point of the out-of-Africa replacement idea is to draw humans from an unstructured ancient population. Humans have to be inbred to explain the low genetic variation today. A long bottleneck in Africa is one explanation for this inbreeding
We do know that there was substantial variation in Africa at the time when modern Homo sapiens emerged. A long period of inbreeding in Africa would indeed create the fairly homogeneous species we discover when we look at most genes. This species would, however, during its population expansion -within Africa- come into contact with other previously isolated African populations, some of which would go extinct, while others might be absorbed, their genes persisting at low frequency in the expanding species.

This requires no great leap of faith. It parallels directly what happened in Africa in the last few thousand years, where previously isolated Pygmy and Khoi-San populations came into contact with expanding farmer-pastoralists, and contributing a little bit to the farmers' genomes.

Thus, the expanding African population that eventually spilled over into Eurasia, would indeed be quite inbred and homogeneous, but its gene pool would also contain traces of the smaller, less successful African populations it had absorbed. Because of their low frequency, these traces would be more susceptible to extinction in the series of bottlenecks that led to Europeans on one side and East Asians on the other, with different sets of archaic genes preserved in either region.

Molecular Biology and Evolution, doi:10.1093/molbev/msp096

Detecting ancient admixture and estimating demographic parameters in multiple human populations

Jeffrey D. Wall et al.


We analyze patterns of genetic variation in extant human polymorphism data from the NIEHS SNPs project to estimate human demographic parameters. We update our previous work by considering a larger data set (more genes and more populations), and by explicitly estimating the amount of putative admixture between modern humans and archaic human groups (e.g., Neandertals, Homo erectus, H. floresiensis). We find evidence for this ancient admixture in European, East Asian and West African samples, suggesting that admixture between diverged hominin groups may be a general feature of recent human evolution.



Maju said...

Hmmm... A very interesting hypothesis to debate around. If my memory is correct, I had the occasion or reading some draft of that paper a couple of years ago and I did not find it too convincing then (but admittedly the sheer complexity of the analysis overwhelms me).

Hawks himself finds potential weaknesses in the study, one of which is that it could just be differential selection.

But I am, quite intuitively, even more in line with Dienekes in this case: the Yoruba and Eurasians do not need to descend from exactly the same original African gene pool. There may be genes in East Africa that have never reached the Yoruba or Eurasians.

Hawks replies to Dienekes (in an update to his original post) with two arguments. In sythesis:

1. That the pattern isn't consistent with Europe and China being drawn randomly from the same ancient African population.

2. That there should be a single original "bottlenecked" (or inbred) population in Africa, what would not allow for any kind of differential structure.

I think both caveats are pretty much answered by considering that all that did not happen in a single moment or short simple epysode but in a very long time of many dozen milennia.

If the main branch leading to West Africans can be identified with, for example, the mtDNA L2 expansion, we have a point of divergence towards modern Yorubas (L2'3 split) before the divergence towards modern Eurasians (L3 split). Sure that modern Yorubans also have L3 but the sublineages are distinct from those in Eurasia and even from many of those in East Africa.

So the components from the different, dynamically evolving, subpopulations within that original African pool was dynamically affecting differently to ones and the others.

And of course, we are here ignoring a lot of other African peoples that may have and surely did interfere as well, like (say) Pygmies, who branched out even before, etc.

Moving into Eurasia, Chinese and Europeans are of course not only "randomly" drawn from Africa. There was an intermediate stage in Southern Asia, that for some odd reason remains being ignored, most probably. Also the expansion towards East Asia happened long before the espansion towards West Asia (and certainly to Europe) did. The sub-populations within Asia from where the two genetic pools were drawn were already differentiated (Chinese evolved most directly from SE Asians, Europeans from West Asians and South Asians before them).

So, well, the structure is not just random, and certainly is not static nor strictly parallel.

The bottleneck that Hawks mentions must have happened before all these expansions, either before the L0'1 split and/or (maybe more likely, as no Khoisan or East Africans were studied) between that point and the L2'3 split. Additionally there was surely some mild bottleneck(s) at the OOA process.

The influence of "minor" human populations (like the alredy mentioned Pygmies or maybe others long since vanished) would not significatively alter the overall structure of the main L2'3 population (the one studied here and the one dominant within modern Humankind) because their input would have been small in any case.

Maju said...

Or in other words, they could more easily prove their point if they studied (also or only) any populations that are not at the end of the process but it the middle (East Africans, South and SE Asians) or at the margins (Pygmies, Khoisan and even Papuans).

Also one can reasonably question how good are those 4 standarized samples (CEU, YRI, CHB and JPT) to study the sheer complexity of humankind at such levels of complexity. They may serve to get a prelimnary work hypothesis but to get a full-fledged theory at work you need much more complete and, crucially, centrical data.

magnessan said...

So that is where the ancient and current Greeks got their admixture from. I knew they had subsaharan African admixture.

You can just tell by looking at them.

Maju said...

No you cannot just tell by looking at them and there's no clear evidence of any "subsaharan" component among Europeans (the occasional erratic even as north as Norway, that's all). Greeks and other Europeans do have an African component clear in Y-DNA genetics (E1b1 - nothing in the rest of ancestry) but this component is North African, not Tropical African, by origin. E1b1 probably diverged, long ago, at Upper Egypt (or maybe lower Sudan), which is in the northern half of the Sahara and by no means "subsaharan" (damn, I really hate that word: south is not "below" = "sub", that's just a cartographic convention).

Maju said...

Both Upper Egyptians and Lower Sudanese are Negroes. This means they are subsaharans. In fact, many of them look exactly like Afro Americans.

No, by no means. And your tone sounds nazi/KKK, what says nothing good about you and will surely make me drop this conversation at the slightest further racist word.

Egyptians cannot be considered black people. Just make a tour through the country: they may have some admixture (logically). Lower Sudanese (Nubians) are more the mixed kind, though maybe they were darker in the past.

Whatever the case they do not look like Afroamericans because Afroamericans originated at the West half of the most diverse continent on Earth. The Tropical conncetion among Afroasiatics (and E1b lineages in general) would be with NE Africa, going as as far south as Ethiopia maybe.

Egypt is not south of the Sahara in any case.

Therefore by your own words you provide proof that Greeks and other mediterraneans are partial negroes.

And all other Europeans if you go by the "one drop rule". E1b is found in all Europe. We are all negroes if you go so extremist, even the Finns. E1b is not that rare in places like Britain or Denmark and is quite common in Central Europe anyhow (about 10% in Hungary or Slovakia). So yeah, maybe, the frizzy hair of Sarkozy is of African ascent, but his Dracula look is 100% European. ;)

Onur Dincer said...

maybe, the frizzy hair of Sarkozy is of African ascentIt might have come via Sarkozy's maternal grandfather, who was a Sephardic Jew.

magnessan said...
This comment has been removed by the author.
magnessan said...
This comment has been removed by the author.
Dienekes said...

Greeks do not look Swedes. You can see they are heavily admixed. Find another venue for your trolling.

magnessan said...
This comment has been removed by the author.
magnessan said...
This comment has been removed by a blog administrator.
Onur Dincer said...


Primary distinguishing factor between European ethnicities is by far the different distributions of Caucasoid haplogroups among various populations. A secondary distinguishing factor is the notable presence of some Y-DNA Mongoloid haplogroups among some North Eastern European populations and almost total lack elsewhere. Negroid haplogroups are very marginal in Europe, and they are not a distinguishing factor as they may seldom pop up - if ever - anywhere in Europe.

Onur Dincer said...

BTW, E1b1 is surely Caucasoid. Its presence in the Horn of Africa can only demonstrate the Caucasoid admixture in those populations, which is known for a long time.

Maju said...

Greeks do not look Swedes. You can see they are heavily admixed.

LOL. So Swedes are now the archetypical aboriginal European, the perfected Caucasoid archetype... or whatever.

Nordics are a rather recent evolution from more basic Caucasoid types much better represented by Greeks than by Swedes surely. They are extremophyles, so to say.

BTW, E1b1 is surely Caucasoid. Its presence in the Horn of Africa can only demonstrate the Caucasoid admixture in those populations, which is known for a long time.

Haploid lineages are not races. You can perfectly be black and have "white" both Y-DNA and mtDNA. And vice versa. It is not very likely but can happen.

Now, E is a lineage quite clearly of African (probably NE African) origin. By NE Africa here I mean the aera between Egypt and Somalia, which has some marked continuity, specially in regard to Y-DNA.

The area of origin of Y-DNA E and also E1b1 and "Caucasoid" E1b1b is precisely the "border" (never well defined) between "Caucasoid" North Africans and "Negroid" Tropical Africans. It is certainly the area of most intense contact between these two "races", because of the Nile and (secondarily) the Red Sea waterways. Arguing wether these lineges are black or white is like arguing about the sex of angels, so to say. It has no valid answer, as it will necesarily depend on all the other ancestors.

Now, if you go by the KKK one-drop rule, then we are all Blacks: everybody without exception has some ancestors in Tropical Africa.

Maju said...

Onur: all those clades make up a single haplogroup, E1b1b-M215; its "brother" is E1b1a-M180 is the most commom male haplogroup in Tropical Africa, together with other lineages of that same E super-haplogroup (E1a, E2).

E is African and, if you push me on where: Black African by origin, and if you push me in where in Black Africa, I'd say, roughly, Sudan-Ethiopia, the area that also produced Eurasians and that can to a large extent be considered the ancestral homeland of Humankind (maybe to the exception of Khoisans??).

You can say almost the same of E1b1b, as diversity is concentrated around the Nile. It is maybe less clearly "Black African" but it is not clearly "white" either. I normally think in colorless terms, as there are no true barriers but rather flows. But if you have a problem with some nordicist bonehead, you can always remind them that R is "Indian" and IJ "Jewish" - that will piss him off a lot. Alternatively, I guess you can go eugenic and make a small favor to humankind's mental and genetic health by supressing nazi idiocy... but I digress.

Now that we have clarified that haplogroups have no color, maybe we can get on topic again...

Onur Dincer said...

E1b1b may be ultimately black African in origin, but the long isolation of its Caucasoid subclades from the other black-originated haplogroups makes them very different from any black-originated haplogroup.

BTW, R most probably originated in Central Asia, not India.

magnessan said...
This comment has been removed by a blog administrator.
Maju said...
This comment has been removed by the author.
magnessan said...
This comment has been removed by the author.
magnessan said...
This comment has been removed by the author.
magnessan said...
This comment has been removed by a blog administrator.
Maju said...
This comment has been removed by the author.
terryt said...

"suggesting that admixture between diverged hominin groups may be a general feature of recent human evolution".

Apart from creationists, who believe in the total separateness of humans from all other species, I find it amazing that so many are so vehemently opposed to even considering the possibility. Sure the evidence in favour of admixture is circumstantial, just some hybrid-looking individuals and the odd gene that seems too early to have come 'out of Africa'. And mtDNA and Y-chromosomes argues against admixture, but we know that these haplogroups do not tell the full story. On the other hand many working in the field of evolutionary biology are quite prepared to accept the possibiliy for admixture, because that's how evolution works for many species. But a surprising number of others are extremely reluctant to entertain the thought, even for as long as a minute. I'd be very surprised if some concrete evidence for admixture does not emerge in the near future.

"40k out of Africa... someone tell people there were modern humans in Oz over 55k ago. An absolute minimum OOA date is 70k. And aren't there signs of habitation under the Toba ash layer in East Asia? So 90k min".

And modern humans appear in the Levant about 100k, but the current zeitgeist states that this is too old to fit the Y-chromosome evidence. However it is congruent with the mtDNA evidence, as long as we're prepared to accept the possibility that the mtDNA introgressed into a modern human/Neanderthal hybrid population somewhere round the Zagros Mountains, or further north on the Iranian Plateau, and later spread from there. There is then no need to assume the presence there of modern human Y-chromosomes at a date when the evidence suggests they were still confined to Africa, or hadn't even yet appeared.

mathilda said...

" Indeed, this very paper finds evidence of archaic admixture in Africa itself".

Didn't Hammer find evidence for that on the X chromosome? Something about one of the pygmy groups having an archaic X chromosome. Wouldn't surprise me if some sub species got isolated in West Africa and then got absorbed when the modern population expanded over their territory.

Terry, the mtDNA age for India aren't that far off fitting a 100k exit for the M there. One East Asian M has an est age of 87k, if I remember right.

The Viking said...

Well what could have caused the split between north europeans and south europeans? Its very obvious when you see the differences between north europeans and south europeans.

terryt said...

"One East Asian M has an est age of 87k, if I remember right".

I don't think that the most ancient reckoning for Y-haps goes back that far. Suggests the women in India were having children with someone else at that date. Possibly 'modern humans' but not with haplogroups belonging to C, D, E or F.

WingsMassageBdw said...

Wow! Amazing how racist you people are. "Greeks are part negroid" "dark and unattractive" "Egyptians do not look black" "Egyptians look like Afro-Americans"...blah blah blah. As if there is something not quite 'good' or 'kosher' about being black, African or whatever so-called title you folks want to put on the richly-hued humans who happen to be ALL of our ancestors and who happen to be the originators of the human race. Much of this dna blog sounds like a bunch of mumbo-jumbo, pseudo-science, Nazi eugenic bull-crap! ALL HUMANS ORIGINATED IN AFRICA PERIOD! The first people on earth were AFRICAN, i.e. BLACK. So just get over it!