November 16, 2010

Demographic history of Oceania (Wollstein et al. 2010)

frappe analysis on the left: New Guinea Highlanders split at K=4 (light blue), with Polynesian-Fijians remaining aligned with East Asians; at K=5 the specificity of the East Eurasian component in Polynesians-Fijians (teal) is revealed; at K=5 the specificity of Borneo is apparent (red), but there are individuals of clearer East Asian ancestry remaining.

From the paper:
Among the three demographic models examined for the peopling of Near Oceania (Figure 4, models 2a–2c), the model receiving the highest support involves a split of New Guineans from a common European-East Asian (i.e., Eurasian) ancestor population. This finding does not support the southern dispersal hypothesis of separate human migrations from Africa to Near Oceania and to East Asia [33, 34]. The existence of a single ancestral population for all present-day non-Africans is supported, among other genetic evidence, by recent data from the Neandertal genome sequence, indicating that all present-day non-African genome sequences studied(including one from a Papua New Guinean) have equivalent amounts of Neandertal admixture [46].
However, the authors date the split of Near Oceanians from the common Eurasians at 27ky and of East Asians from Europeans at only 18ky. These dates are far too low, in my opinion, as there is evidence that Upper Paleolithic Europeans were already robust versions of modern Caucasoids.

Moreover, if Eurasian unity broke down at 27ky, then where were the Eurasians since the time they acquired "Neandertal admixture" until 27ky?

It is difficult to imagine Eurasians camping in the Near East or Europe (where Neandertals are attested) for tens of thousands of years before starting to split off at 27ky. And indeed, the fact that there are anatomically modern humans from South China and the Levant at around 100ky, make the idea that Eurasians got Neandertal admixture in one place before starting to disperse a few tens of thousands of years ago hard to believe.

I personally don't buy the idea that New Guineans have the same "Neandertal admixture" as Europeans. In fact, I doubt there is any substantial Neandertal admixture in Eurasians at all, and if there is, it is certainly not the 1-4% evenly distributed element across Eurasia that was discovered in the recent paper.

In any case, this issue is peripheral to this paper which offers important new data on the question of Oceanian origins.

Related:

Curr Biol. 2010 Nov 10. [Epub ahead of print]

Demographic History of Oceania Inferred from Genome-wide Data.

Wollstein A, Lao O, Becker C, Brauer S, Trent RJ, Nürnberg P, Stoneking M, Kayser M.

Abstract

BACKGROUND: The human history of Oceania comprises two extremes: the initial colonizations of Near Oceania, one of the oldest out-of-Africa migrations, and of Remote Oceania, the most recent expansion into unoccupied territories. Genetic studies, mostly using uniparentally inherited DNA, have shed some light on human origins in Oceania, particularly indicating that Polynesians are of mixed East Asian and Near Oceanian ancestry. Here, we use ∼1 million single nucleotide polymorphisms (SNPs) to investigate the demographic history of Oceania in a more detailed manner.

RESULTS: We developed a new approach to account for SNP ascertainment bias, used approximate Bayesian computation simulations to choose the best-fitting model of population history, and estimated demographic parameters. We find that the ancestors of Near Oceanians diverged from ancestral Eurasians ∼27 thousand years ago (kya), suggesting separate initial occupations of both territories. The genetic admixture in Polynesian history between East Asians (∼87%) and Near Oceanians (∼13%) occurred ∼3 kya, prior to the colonization of Polynesia. Fijians are of Polynesian (∼65%) and additional Near Oceanian (∼35%) ancestry not found in Polynesians, with this admixture occurring considerably after the initial settlement of Remote Oceania. Our data support a greater contribution of East Asian women than men in the admixture history of Remote Oceania and highlight population substructure in Polynesia and New Guinea.

CONCLUSIONS: Despite the inherent ascertainment bias, genome-wide SNP data provide new insights into the genetic history of Oceana. Our approach to correct for ascertainment bias and obtain reliable inferences concerning demographic history should prove useful in other such studies.

Link

6 comments:

Andrew Oh-Willeke said...

There are two analytically distinct questions.

First: Was there are common founder population for all Eurasians? There is a great deal of evidence that this was the case.

For example, both New Guinea and Australia have locally specific mtDNA haplogroups that descend from M, N and R. So, the phylogeny of mtDNA must have had considerable time to evolve before a group containing all three types went on to settle New Guinea and Australia respectively.

If Neanderthal admixture whose traces remain in existing populations happens with a common source population anytime 48kya or earlier, give or take a couple thousand years, then it is entirely plausible that Neanderthal admixture is more or less uniform among all Eurasians.

This isn't inconsistent with a higher level admixture in Upper Paleolithic Europeans. The Europeans who cohabited with Neanderthals would have been partially replaced in the retreat leading up to the LGM and subsequent repopulation of the contintent, they would have been partially replaced again at the Neolithic. They would have been partially replaced again during the IE expansion. Successive waves of partial replacement dilute those early Upper Paleolithic component of those that follow. And, even if a minority of Europeans do trace a significant share of their ancestry all the way to the early Upper Paleolithic, that share is probably small enough that it is unsurprising that the couple of West Eurasian genomes that were used for the initial comparison. You'd expect the highest Neanderthal admixture in people in people with the most recent ties to Northern European hunter-gatherers, probably Saami, Baltic populations, Mari Mountain populations, and Siberians who speak languages that don't have recent origins in the historic period (i.e. non-Turkish, non-Mongolian).

Second: "[T]he authors date the split of Near Oceanians from the common Eurasians at 27ky and of East Asians from Europeans at only 18ky." Is this an accurate date?

Surely, the answer to the second question is "no." Archeology shows New Guinea and Australia inhabited by 45kya. Both New Guinea and Australia have genetic markers that are highly distinct from everywhere else in the world and from each other. There is no solid evidence that population genetically significant contact was maintained with the rest of East Asia for 18,000 years after initial contact.

terryt said...

"the phylogeny of mtDNA must have had considerable time to evolve before a group containing all three types went on to settle New Guinea and Australia respectively".

We don't know that all three types went on to settle New Guinea and Australia at the same time. perhaps three separate settlements?

Andrew Oh-Willeke said...

"We don't know that all three types went on to settle New Guinea and Australia at the same time. perhaps three separate settlements?"

We do know that (1) the effective population size of the founding group of Australia was very small; (2) there are a great many mutations unique to Australian and New Guinea specific haplogroups that make them very old compared to other haplogroups by most mutation clock measures and that mutation clocks probably "ticked" slowly in those areas due to low population density; (3) most indigeneous genetic diversity in Australia is confined to the country's North Coast where contact can't be ruled out and has existed since at least the Austronesian era; (4) the New Guinea and Australian unique M and N and R types do not have any known source in Asia or Indonesia; (5) there is a significant pre-Austronesian genetic population strata in modern Indonesia which would imply that most mtDNA types present during a hypothetical subsequent settlement of New Guinea and Australia would still leave traces in Indonesia today; (6) we can estimate the secondary arrival date of the Dingo with some specificity and the extremely low genetic diversity of Dingos indicate that this was probably a one off event or very close to that; and (7) there isn't any solid archeological evidence that there were boats sufficient to make it to Australia on an ongoing basis pre-Austronesian and post-risen sea levels.

So, while one could hypothesize three separate settlements and that might even be possible, there is really no evidence of any kind of which I am aware that would prefer that conclusion.

There is also no evidence of significant population replacement in Australia prior to the arrival of Europeans, and evidence of population replacement in New Guinea is pretty much limited to the Northern and Western lowlands, and even there is not complete. The evidence seems to support population continuity in the New Guinea Highlands at least since plants were domesticated there.

terryt said...

"The evidence seems to support population continuity in the New Guinea Highlands at least since plants were domesticated there".

That seems to correlate with the spread of languages of the trans-New Guinea phylum. certainly pre-Austronesian, in spite of what German is arguing on another thread.

"So, while one could hypothesize three separate settlements and that might even be possible, there is really no evidence of any kind of which I am aware that would prefer that conclusion".

"(2) there are a great many mutations unique to Australian and New Guinea specific haplogroups that make them very old compared to other haplogroups"

But in general the two regions have different haplogroups. Y-hap C in Australia, but C seems only to have reached New Guinea with the Austronesian expansion. K-derived Y-hap M in New Guinea/Melanesia is unknown in Australia. Basal N mtDNA S in Australia but not in New Guinea. Basal M-derived mtDNA Q in New Guinea but not in Australia. Suggests two separate movements into the region, one to Oz the other to NG.

On the other hand several Y-hap Ks and mtDNA P are found in both places. Suggests a later wallacean crossing.

"(4) the New Guinea and Australian unique M and N and R types do not have any known source in Asia or Indonesia"

They certainly became isolated and developed on their own once they had crossed wallace's line. I agree that crossing it wasd a spasmodic affair, but seems to have happened more than just once or twice.

terryt said...

"the authors date the split of Near Oceanians from the common Eurasians at 27ky and of East Asians from Europeans at only 18ky. These dates are far too low, in my opinion"

A product of later admixture? Similar to Andrew's comment, 'Successive waves of partial replacement dilute those early Upper Paleolithic component of those that follow'.

"Fijians are of Polynesian (∼65%) and additional Near Oceanian (∼35%) ancestry not found in Polynesians, with this admixture occurring considerably after the initial settlement of Remote Oceania".

It has long been accepted that there was a lter movement of 'Melanesian' people along behing the Lapita (or Polynesian) expansion.

German Dziebel said...

Apologies, I haven't read this paper yet.

"That seems to correlate with the spread of languages of the trans-New Guinea phylum. certainly pre-Austronesian, in spite of what German is arguing on another thread."

Terry is referring to our discussion here http://dienekes.blogspot.com/2010/09/post-glacial-migrations-of-humans-into.html

What I'm pointing out is that frequencies of Y-DNA C in south Wallacea and in island Melanesia suggest that C*/C2 lineage (high frequencies in Polynesians) is of Austronesian origin, with non-Austronesian populations in the area receiving it through gene flow from expanding Austronesians. Taking the existing Y-DNA stratigraphy at face value, C*/C2 is a Pleistocene lineage that's related to the Australia-specific C4 lineage. Coupled with mtDNA Q found at low frequencies in Melanesians and Polynesians and the recent suggestion of kinship between Austronesian and Ongan languages (Y-DNA Dx, mtDNA M31, M32), this places the origin of Austronesians into a time frame earlier than Chinese Neolithic and likely well into early Holocene/late Pleistocene. The fact that 90% of male lineages east of the Wallace Line are Melanesian and not Asian in origin suggests that the interaction and gene flow between Austronesians and non-Austronesians in Melanesia is of considerable antiquity. Subsequent "Mongoloid" gene flow from Taiwan and East Asia brought more recent lineages into the Pacific. At some point in the Late Pleistocene/early Holocene, Austronesian languages existed in a SEAsian or Wallacean refugium as a dialect chain (and not a superlarge family) similar to Andamanese or Ainu. Trans-New Guinean is the closest analogy to an agriculture-driven language spread of early Holocene antiquity.