As you probably know a new ancient genome paper was published last week in Nature, Terminal Pleistocene Alaskan genome reveals first founding population of Native Americans. There is at least one other involving Willerslev in the works for what it’s worth. Carl Zimmer has a good write-up in The New York Times, while Greg Cochran picked up the fact that the latest results show no evidence of “Australo-Melanesian” affinities that have been found in Amazonians.
The key issue here is that they found 11,500 year old remains from Alaska, one of which they sequenced at 17x coverage, which is rather good (not medical grade good, but really sufficient for a lot of population genomic work). It’s clear that the lineage represented by these remains is “basal” to that of other Native American peoples, whom David Reich’s group labeled “First Americans.” Later, the First Americans diverged into different populations, with the two in modern focus being a northern cluster, including the Aboriginal peoples of Canada and parts of the United States, and a southern one including everyone else. This does not mean that the Beringians were isolated outliers. There may have been many other peoples related to the Beringians who diversified, who went extinct as well. The settlement of Alaska by other peoples suggests to me that extreme conditions in the Arctic made it likely that there would be population turnover there. Also, the fact that these samples were located close to the source of settlement in the New World by modern humans makes their distant relation to all other New World populations unsurprising.
The big thing that the press is highlighting is the confirmation of the Beringian Standstill model, where modern humans percolated into the area between Siberia and Alaska, Beringia, and did not move east for thousands of years. Basically, the conditions were inclement toward human habitation on both sides of Beringia, while a relict modern human group likely occupied a pocket of more moderate climes for thousands of years, with minimal gene flow from the west, and blocked from migration to the east. Genetically the Beringia Standstill made sense for a long time…the divergence between Amerindian lineages and those of eastern Eurasia seemed too old to be accounted for by recent migration a bit more than 10,000 years ago (the old “Clovis first” hypothesis).
How old? This paper suggests that the portion of Native American ancestry which indicates an affinity to East Asians stopped exhibiting gene flow from that source around ~25,000 years ago, after diverging around ~36,000 years ago. This points to the fact that after modern humans came to dominate eastern Eurasia they began to diversify rapidly after 40,000 years ago, but gene flow between different populations did not always allow them to drift apart…at least initially. The ancestors of Native Americans and East Asians may have been in extremely separate locations by ~25,000 years ago, whether it be on the fringes of eastern Siberia, or somewhere in southern China (there is no reason that the modern Chinese have to have had ancestors resident on the North China plain before the Last Glacial Maximum).*
One aspect here I want to emphasize is that our image of a world thickly populated with humans may mislead us in our intuition about how patchy occupation was ~25,000 years ago. Yes, humans may have left artifacts all over the world, but that doesn’t mean that there weren’t centuries or millennia of no occupation, or, that meta-population dynamics were such that extinctions were common. For decades in population genetics there has been talk of “clines vs. clusters,” but if human population densities were far lower, or occupation patchier, then clines may have become much more important recently with high density than in the past.
Finally, back to the Australo-Melanesian issue. Either there is a lot of population structure in ancient Beringia to be explored, with diverse quasi-Asiatic groups, or there was an Australo-Melanesian group already in South America.
* Ancient North Eurasian ancestry came into Beringians ~20,000 years ago. Two groups which merged during the middle of the Last Glacial Maximum.
“… or there was an Australo-Melanesian group already in South America.”
Why already? Why could this group not have arrived after the descendants of the Beringians?
There is no archaeological evidence of early “Australo-Melanesians” in western Amazonia, where the Karitiana and Surui are located. This term itself is misleading, because it lumps together Melanesians and Australasians–not Australians. The modern population that stands in for the mysterious ancient “Y” population in Skoglund’s analyses are the Onge of the Andaman Islands, who are not particularly close geographically or genetically to Melanesians. The same ancestral component has been reported in the 40,000-yr-old Tianyuan genome from China. So I suspect your first alternative is correct–structure in the Siberian-West Beringian source population. I was a bit surprised that the new study does not make any use of the Afontova Gora 3 genome. Despite dating from 18,000 cal BP, this individual was more closely related to Native Americans than Mal’ta (24,000 cal BP) is. This suggests the input from ANE into Native Americans was somewhat later than the authors assume. For a good take on the implications of the USR genome see Ben Potter’s personal web page.
Why already? Why could this group not have arrived after the descendants of the Beringians?
Because a boatload of pseudo-Andamanese people landing in a South America which was reasonably full of Amerinds already wouldn’t have any detectable genetic impact. And given the huge distances involved, it is implausible (unless you want to consider ancient astronauts or something) that more than a handful could have crossed the Pacific Ocean.
The situation with the known second/third waves into North America is not analogous, because:
1. The geographic proximity of Siberia meant that hundreds of individuals could have taken part in crossing the Bering Strait over a period of several generations.
2. The colder weather in the arctic meant that they may have found virtually uninhabited areas in Alaska where they could initially establish a beachhead.
3. Particularly in the case of the Inuit, they developed new technologies which allowed them to exploit tundra landscapes far more effectively than the earlier migrations did.
i agree name is not optimal. you have alternatives?
One more general comment regarding the population tree from the paper – I find it interesting that the Ket and Athabaskans have linked ancestry only via a “East Siberian like” ancestor. That is to say, the Athabaskans source population in Siberia was not proto-Ket, since it isn’t enriched with Ancient North Eurasian. Also, the Ket cannot be descended from a back-migration from North America, or an Amerind genetic link would have been found. So if the Dené–Yeniseian language family is correct, the original proto-language must have been spoken by a “pure” East Eurasian/Siberian group. Which honestly makes me wonder if the linkage that linguists like Sergei Starostin made between these languages and Sino-Tibetan had something to it.
On the other hand, modern day Inuit and Koryaks have some ANE and East Siberian ancestry, but don’t appear to speak languages related to Dené–Yeniseian. These families have been linked together by Merritt Ruhlen along with some others in Eurasiatic. I’m doubtful this could ever be proven, but if such a proto-language ever existed, the source population would almost certainly be Ancient North Eurasian.
How about “ancient south-central Asian” or “ancient Onge-like”, to emphasize that this is a “ghost” population?
The predominant Onge Y-chromosome lineage is a clade of Haplogroup D, distantly related to Himalayan and Japanese(!) clades. There’s also a D branch in the Philippines–maybe the few Surui/Karitiana-related autosomal genes in Melanesia and NE Australia are the result of Austronesian expansion out of Taiwan and the Philippines.
“i agree name is not optimal. you have alternatives?”
I usually call this a “Paleo-Asian” component, which is relatively agnostic about its geographic source.
“The predominant Onge Y-chromosome lineage is a clade of Haplogroup D, distantly related to Himalayan and Japanese(!) clades.”
The Onge Y-DNA D clade is much closer to Himalayan and Siberian Y-DNA D in phylogeny, than to the Japanese which has a deep time depth split from other Y-DNA D. Realistically, a northern route Y-DNA D expansion that migrates from Tibet to the Andamans, rather than the other way around, and then is largely wiped out in the north except in select refugia, obscuring its route, during the Last Glacial Maximum, is probably the best explanation.
But, I would also not put good money on the autosomal Onge-like Paleo-Asian component having a very long term association with Y-DNA D. More likely, most of the ancestors of the Onge were a SE-most edge of an Ancestral South Asian cline which had other Y-DNA types (probably typical South Asian private Y-DNA haplogroups like H), but the folks who migrated to the Andamans ca. the LGM to form a relict population there had seen Y-DNA D replacement in this founding population. (Ironically, parts of Tibet were refugia during the LGM with more habitable weather than the Siberian flat lands at lower elevations to the north.)
More concretely, a band of Y-DNA D brothers from Tibet probably made their way down the mountains into SE Asia (possibly pushed by deteriorating conditions during the LGM to the north), conquered a tribe of ASI folks at their SE Asian extreme, and migrated with the conquered tribe to the Andamans with a very small founding population, while the most of the other populations of that part of the cline who didn’t migrate were (mostly) replaced by other folk.
The Andaman founding population was very small.
The Onge may have emigrated from the Asian mainland in the Holocene. The Andamanese have pigs and pottery, and there is no archaeological record dating older than 3000 cal BP. They seem to be most closely related to Malaysian Negritos. See Chaubey and Endicott in Human Biology 85(1)(2013).
I’ve just re-read Yang et al. (2017) on Tianyuan man’s genome. Take a look at their Figure 3. They model Surui as deriving 96% from the same Paleo-American ancestral population represented by Mixe (a simple mixture of 70% pre-Tianyuan Asian and 30% ANE) and the remaining 4% from an unnamed node (population Y?) that is a mixture of 44% related/ancestral to Tianyuan and 56% derived from the lineage leading to Papuans (which must have split from the pre-Tianyuan line long before 40,000 cal BP). They don’t depict Onge separately in the figure, so I don’t know if they are lumping them here with Papuans as a single statistical entity. In their words: “The relationship between the Tianyuan individual and the Amazonians is similar to that reported between the Papuan and the Amazonians and Onge and the Amazonians.”
One other point worth noting is that a new native group has been put into play by Yang et al. They observe that a people living in the Andean foothills in northern Argentina and Bolivia called the Chane have the same Tianyuan/Papuan/Onge component seen in the Surui and Karitiana. A cursory Google search revealed that the Chane originally were Arawakan speakers which would imply a relatively recent migration from the lower Orinoco.
Another fact that may be relevant is that Wang et al. (2007) reported that the Surui have extremely low genetic heterozygosity. I wonder if this is why the Tianyuan-related alleles are abnormally common here–a small and inbred population with strange drift effects (analogous to the very high frequencies of mtDNA haplogroup X in the Druze).
The Andaman Islanders didn’t really need to “migrate” to the islands, because the entire island chain was hilltops on a narrow peninsula intermittently connected to modern-day Burma during the Ice Ages.