Science 29 August 2014: Vol. 345 no. 6200 DOI: 10.1126/science.1255832
The genetic prehistory of the New World Arctic
Maanasa Raghavan et al.
The New World Arctic, the last region of the Americas to be populated by humans, has a relatively well-researched archaeology, but an understanding of its genetic history is lacking. We present genome-wide sequence data from ancient and present-day humans from Greenland, Arctic Canada, Alaska, Aleutian Islands, and Siberia. We show that Paleo-Eskimos (~3000 BCE to 1300 CE) represent a migration pulse into the Americas independent of both Native American and Inuit expansions. Furthermore, the genetic continuity characterizing the Paleo-Eskimo period was interrupted by the arrival of a new population, representing the ancestors of present-day Inuit, with evidence of past gene flow between these lineages. Despite periodic abandonment of major Arctic regions, a single Paleo-Eskimo metapopulation likely survived in near-isolation for more than 4000 years, only to vanish around 700 years ago.
Link
23 comments:
This is quite interesting.
One question: What do they mean by a single paleo-eskimo metapopulation surviving in near-isolation? I thought a metapopulation consisted of a number of closely related groups that 'cycled' through a given geographic area?
If anyone could explain this, I would be very grateful.
This paper definitively resolves many key issues regarding North American pre-history:
It establishes that the Saqqaq and Dorest populations were a single, closely genetically related migration wave, despite a division in archaelogical culture.
It establishes that the Na-Dene people do not have significant Paleo-Eskimo admixture and instead represent a separate wave of migration (the study does not address the timing of the Na-Dene wave), while affirming that they are a distinct wave for the early wave Native Americans. Thus, there were at least four pre-Columbian waves, early Native American, Na-Dene, Paleo-Eskimo and Inuit, plus around Y1K, a small Viking population in Vinland and a small Asian population in Peru, neither of which persisted or had much long term impact. There is also some possibility that early Native Americans may have derived from two or three subwaves all pre-Clovis and post-25kya.
It establishes that the 6th to 7th century CE Siberian Birnirk culture is the source of the proto-Inuit Thule people.
It notes that: "Although we cannot preclude later gene flow between the Dorset and the Thule (that is, subsequent to the more ancient gene flow that occurred at least 4000 years ago), the contrasting genetic and cultural affinities of the Sadlermiut individuals present a conundrum. This culture that went extinct in 1903 CE from European disease has long been considered Thule-acculturated Dorset people, likely due to intermarriage; however, genetic evidence from this study suggests that they were Thule people who had somehow acquired Dorset stone technology."
The only hypothesis that is not analyzed as directly as it might be (although the supplemental data may make it possible to discern) is the hypothesis that the Paleo-Eskimos may have been genetically related to proto-Uralic or proto-Altaic populations, rather than Paleo-Siberian linguistic populations.
Raghavan is again drowning in his own data.
Here's the data. Raghavan admits that Paleo-Eskimos, Neo-Eskimos, Na-Dene and Amerindians stem from the same parent population which is separate from East Asians or Europeans. He finds that Paleo-Eskimos form a clade with modern Greenland Eskimos and modern Aleutians and the latter form a bigger clade with Amerindians, Na-Dene and the rest of modern Eskimos. His ADMIXTURE analysis clearly shows that Saqqaq (Paleo-Eskimos) have the Amerindian component, which is at 100% in Karitiana. He observes that Saqqaq behaves in relationship to MA-1 in the same way as Amerindians do, namely Saqqaq is closer to MA-1 than Han or other East Asians and Siberians. We also know from Raghavan 2013 that Eskimos and Na-Dene are closer to East Asians than Amerindians, so Saqqaq (Paleo-Eskimos) is not different here. Raghavan 2014 shows that the direction of admixture went from Saqqaq to MA-1 and from modern Amerindians to modern Europeans. So all the data is most consistent with an out-of-America scenario whereby the Amerindian component migrated out through Beringia across Eurasia and the differentiation into post-Amerindian (paleo-Eskimo, neo-Eskimo and possibly Na-Dene) and pre-Eurasian populations happened in the northern North America and Beringia.
Instead, Raghavan postulates a wild scenario whereby Neo-Eskimos originated in NE Asia and contributed Karitiana-like genes to Paleo-Eskimos who, once a pure East Asian population, thus became more closely related to MA-1 and modern Europeans than East Asians and Siberians. Neo-Eskimos originated from the same NE Asian area as paleo-Eskimos 3000 years later and proceeded to colonize exactly the same territory as paleo-Eskimos, and the latter went extinct quickly in the aftermath of the neo-Eskimo colonization. It's like driving your car backwards.
Also, mtDNA A2 is a Native American haplogroup. NE Asian Birnirk A2a must be derived from New World A2, so even on this count neo-Eskimos went the other way, namely into Asia from America.
From the link:
"Paleo-Eskimos likely represent a single migration pulse into the Americas from Siberia, separate from the ones giving rise to the Inuit and other Native Americans, including Athabascan speakers".
I'll be interested to hear German's take on that.
@German:
Raghavan 2014 shows that the direction of admixture went from Saqqaq to MA-1 and from modern Amerindians to modern Europeans.
You know very well that's not true German. Raghavan unambiguously shows that admixture went from MA-1 into Amerindians. You are welcome to your own personal theories and interpretations, but please don't try to make out they are supported by genuine experts in the field.
@andrew, actually Raghavan doesn't postulate any Na-Dene clade. If I'm reading correctly, this paper says Na-Dene are part of the Karitiana clade.
A good question is what happened to Paleo Eskimos. Are the Haida or Tlingit related to them? They look "Asian" but not very "Eskimo."
@Terry. A successful "single migration pulse" requires both male and female components. The female, as indicated by mtDNA analysis was D2. The male is less certain but could have been NRY haplogroup C. Bosch et al. (2003) detected some haplogroup C Y chromosomes in Ittoqqortoormiit. The Inuit, who became the inhabitants of this eastern Greenlandic settlement were drawn from Ammassalik where elements of Dorset and Saqqaq cultures have been noted.
Andrew, you write that ”the Paleo-Eskimos may have been genetically related to proto-Uralic or proto-Altaic populations, rather than Paleo-Siberian linguistic populations.” I fail to see this connection with Altaic or Uralic populations. The autosomal analysis of MA-1 showed that he shared 10% of his genes with Arctic people, including eskimos, and 0% with modern Siberians. Modern Greenland Eskimos have 0% of Siberian ancestry and modern Siberians (Yakuts and Nganasans) have 0% of Beringian/Eskimo ancestry. (https://docs.google.com/file/d/0B9o3EYTdM8lQbEVvdkZiRk04aWM/edit)
According to Zegura et al. paper on Native American haplogroups, ”haplogroup C has a much more patchy distribution, with most of the C-P39 chromosomes in our sample concentrated in the three Na-Dene populations.” (http://mbe.oxfordjournals.org/content/21/1/164.full.pdf). On the basis of the existing evidence,if we try to distinguish a Na-Dene specific yDNA, it is C2c, if any, while Saqqaq represents the older migration of Q lineages.
Paleo-Eskimos were not pure asian as they had Q y-dna which is verified by Saqqaq. Unless Saqqaq is a mistake and they had C3?
A little nosing around reveals that the oldest Na-Dene archaeological sites in central Alaska where the most basal Na-Dene languages are found, date to about 1500 BC. This time depth is about right if there is a linguistic link to Yenesian of the strength that appears to be present as linguistic evidence suggests. This puts Na-Dene migration more than 2000 years before the Inuit migration, but about 1000 years after the Paleo-Eskimos.
@Tobus
"You know very well that's not true German. Raghavan unambiguously shows that admixture went from MA-1 into Amerindians. You are welcome to your own personal theories and interpretations, but please don't try to make out they are supported by genuine experts in the field."
Look at Fig 5C in Raghavan 2014 and see for yourself. The migration edge arrow points from Saqqaq to MA-1. There's more in Suppl Mat. Also see S15A for arrows from New World populations to modern Europeans in full accordance with the "Amerindian admixture" in Europe discovered last year. Remember, anthropologists are the only "genuine experts" in the field of human origins. Geneticists just contribute their input into a bigger synthesis.
"Paleo-Eskimos were not pure asian as they had Q y-dna"
Whenever I come here I am thoroughly entertained.
Correction! I meant to say that the North American-specific subclade is C-MP39 which is named C3b in Zhong et al paper!
You should read this this recent paper "Y-chromosome analysis reveals genetic divergence and new founding native lineages in Athapaskan- and Eskimoan-speaking populations".
They say that "The Athapaskan-speaking Gwich’in and Tłįchǫ had higher frequencies of C3b than the Inuvialuit, whereas the Inuvialuit had significantly more Q1a6 lineages. Additional haplogroups that seem to be indigenous in origin were found at low frequencies in the Athapaskan groups."
This Q1a6 (NWT01) lineage seems to be the Eskimo-specific lineage, and as its TMCRA is oldest and its diversity highest in Siberian Yupiks, it probably originated in Beringia. The TMRCAs for Q1a6 and C3b were comparable. For Q1a6, the TMRCA was between 4,000 and 7,000 y ago. The overall TMRCA estimate of the C3b lineages was 5,000 y ago. Saqqaq is 4,000 years old.
On the basis of this, it is possible that modern Eskimos have ancient Arctic and Native American mtDNA (A2) but a new yDNA Q-NWT01 was incorporated into their yDNA pool from the Asian side of the Beringian Strait. Koryaks also harbour this Q-NWT01 and perhaps also Chukchi who share a high amount of A2 with Eskimos and Athabascans, with 9/103 of their samples being assigned to haplogroup D2 which is Saqqaq man’s haplogroup. These new Asian elements in Paleo Eskimo’s autosomal results could be due to Q-NWT01’s contacts with northeastern Asian populations and to the introduction of D2a1a into Beringia. The age of STR variation of yDNA C2c-M48 which is frequent in Northeast Asia is c. 10 000 years and it is frequent in Koryaks (28%). It is probably also a vector of new autosomal elements in Beringia.
It is interesting that Wikipedia has a page on haplogroup Q-NWT01 where it is argued that the ancestor of Q-NWT01 is Q-MEH and its descendant is Q-M120.
@German: Look at Fig 5C in Raghavan 2014 and see for yourself.
You mean the Fig 5C that is captioned "A known migration edge is inferred from MA-1 to the root of Native Americans and Inuit, but this gene flow event excludes the high-coverage Saqqaq individual"
Separately to this MA-1 into Amerindian admixture, as you point out, "TreeMix infers gene flow from the high-coverage Saqqaq into MA-1 and vice versa, the latter admixture edge being consistent with (65), but with low support to substantiate this signal and the inferred direction".
They continue: "In contrast, gene flow from MA-1 into the root of the clade comprising Native American populations (Karitiana and Anzick-1) and including the Greenlandic Inuit is detected with high bootstrap support, in agreement with (53)"
You must know all this since you are referring to the paper and have obviously read it, so what do you hope to gain by misrepresenting it in this way?
In order to understand the relationship of Palaeo-Eskimos with the Asian Mainland, the paper of Malyarchuk et al, 2011, is very useful, as they say that ”it is noteworthy that according to results of SNP genotyping, the populations closest to the Saqqaq individuals are Koryaks and Chukchi. As that study suggested that the ancestral Saqqaq peoples separated from their Old World relatives about 5.5 Ka, it is possible that Q1a*-MEH2 lineage detected in Koryaks represents an ancient genetic component, which in the past united the peoples of Northeast Asia, North America and Greenland. In order to determine the range of Q1a*-MEH2 haplotypes in Koryaks we have performed a search for similar STR haplotypes in the YHRD 3 database. As a result, we did not find identical haplotype … but the only similar 9-marker haplotype was found in Yukaghirs from Northeast Siberia. According to Pakendorf et al., this haplotype was detected in 4 out of 13 Yukaghirs … Therefore, it is likely that the range of Q1a*-MEH2 may cover a distance of about 1000 km between the coasts of East Siberian Sea and Sea of Okhotsk. Coalescence age of the Koryak/Yukaghir Q1a*-MEH2 is about 3.5 Ka, that is within the bounds of the Saqqaq culture dating”.
As a further confirmation, Saqqaq mtDNA, D2a1, is found in Chuvantsi Yukaghirs and the founding sequence of D2 is also found in a Yukaghir. (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2427195/)
”The genealogy of the haplogroup D2 shows that the founding sequence marked by the coding region transition at 11215 gives rise to two star-like branches and previously unreported D2* sequence variant [in Kolyma Indigirka Yukaghir]. Most prominent is the node that gave rise to D2a containing two mutations, 3316 and 9536. Next is the D2a1 cluster distinguished by the presence of 7493-8703-16129-16271. It splits into two distinct lineages, the D2a1a defined by the transition at 11959 and D2a1b distinguished by the motif 195–9181, with the former being restricted to Beringian populations [Chuvantsi Yukaghirs, Chukchi and Eskimo-Aleuts], whereas the latter occurred in the ethnically heterogenous Altaic-speakers of Siberian interior, though in very low frequencies.
Of 36 Commander mtDNA samples, three non-Aleut, probably of the Tlingit origin, harbored the D2a1a root type, thus delineating a founding haplotype for all Central Beringian haplogroup D2 mtDNAs. It is apparent that the Aleut 8910A lineage is only a small portion of larger radiation of D2a1a, which gave rise to the Na-Dene, Aleut, and Chukchi-Eskimo lineages.
In my opinion, the expansion of D2 in Beringia clearly predates the recent Altaic and Uralic linguistic expansions. According to Wikipedia, Yakuts originally lived around Olkhon and the region of Lake Baikal. But beginning in the 13th century they migrated to the basins of the Middle Lena, the Aldan and Vilyuy rivers under the pressure of the rising Mongols, where they mixed with other northern indigenous peoples of Russia such as the Evens and Evenks. Tungusic languages are older in the area, but their main yDNA is C2c and not Q, and I do not think that they have anything to do with Saqqaq culture. Nganasans arrived on the tundra quite recently, and they do not share any mtDNA with Eskimos (including not any D2), and only common yDNA is C2c (and Q?). According to Wikipedia, ”by the middle of the 17th century, Tungusic peoples began to push the Samoyedic peoples northward towards the tundra Taymyr Peninsula”.
@Tobus
"You mean the Fig 5C that is captioned "A known migration edge is inferred from MA-1 to the root of Native Americans and Inuit, but this gene flow event excludes the high-coverage Saqqaq individual"
Yes, the one that shows an arrow from Saqqaq to MA-1 (5C) and from the root of Amerindians and Eskimos to modern West Eurasians (5A). Of course, Raghavan has issued a disclaimer that his evidence doesn't mean what it shows, so hence his account is not credible, as I pointed out. His insistence that Amerindians derive from MA-1 is ridiculous in the light of all the arguments I presented apropos Raghavan 2013. Evidence from Raghavan 2014 further confirm my earlier critique as Saqqaq clearly behaves like a New World population (and not an East Asian population) vis-a-vis MA-1 but the directionality of admixture is different from the one presented in Raghavan 2013.
"You must know all this since you are referring to the paper and have obviously read it, so what do you hope to gain by misrepresenting it in this way?"
I'm not misrepresenting Raghavan 2014. I'm criticizing Raghavan 2014 from the point of view of his own data. You again treat everything you read as a gospel. His bootstrap support for MA-1 to Amerindians direction means nothing because no West Eurasian alleles are found in Amerindians and Amerindians have ancestral alleles in common with east Asians that MA-1 doesn't have, hence the direction must be from Amerindians to MA-1 and from Amerindians to East Asian. He treats MA-1 as an unadmixed population, which is contradicted by his own data in Ragahvan 2014, where MA-1 is close to a highly admixed European-Amerindian (unmasked) Aleutian population (Fig. S10A).
Thank you Kristiina for the additional references.
FWIW, I was noting merely that the new article did not resolve the issue, without making a statement about the state of the genetic literature generally.
@German: That's all very well and good (actually it is factually incorrect in a number of key points, but let's stay focused), but this is *your* interpretation not Raghavan et al's, underscoring my point. The theory presented in the paper is actually the *opposite* of what you are saying, hence you are misrepresenting it.
@Tobus
You fundamentally lack any understanding of what a scientific debate is. I'm not misrepresenting Raghavan. I'm providing a 180 degree better interpretation of his own data in the light of how one data point relates to the other (e.g., Saaqaq behaves like Amerindians toward MA-1 and it shows an admixture edge into MA-1 - a point that Raghavan sweeps under the carpet) and in the light of all the other data. And I always represent facts accurately. The real issue is the interpretation methodology that Raghavan uses.
@German: I'm not misrepresenting Raghavan
According to your latest post your original statement is logically equivalent to "Raghavan 2014 shows a 180 degree better interpretation of his own data"... an impossible situation. Clearly what Rhagavan 2014 actually shows is different to what you say it shows, and that's "misrepresentation" in anybody's dictionary.
@Tobus
" what Rhagavan 2014 actually shows is different to what you say it shows, and that's "misrepresentation" in anybody's dictionary."
Raghavan 2014 provided a flawed interpretation of the data he extracted from DNA. I provide an interpretation that's 180 degrees better because it accounts for the data pattern that he chose to dismiss because he didn't have a right theory to work with. There's no misrepresentation here. Geneticists are like machines: they don't think, they just replicate the program they were taught is right. I provide a human(istic) alternative to this mechanistic science.
@German:There's no misrepresentation here
So if I think Einstein was wrong and come up with my own interpretation of Relativity it's OK for me to say "Einstein shows it's possible to travel faster than light speed"? .... or to say "Newton shows gravity doesn't exist" if I invent my own reason for why things fall down?
If you don't like "misrepresent" then call it whatever you want, just please take care to present your original ideas as your own next time.
@Tobus
"If you don't like "misrepresent" then call it whatever you want, just please take care to present your original ideas as your own next time."
Interpretation is mine. The data is Raghavan's. Are you proposing that Raghavan should privatize and patent his data? Typical Tobus's nonsense.
Post a Comment