search this blog

Saturday, May 2, 2015

Complex admixture history and recent southern origins of Siberians


On a brighter note, here are some interesting Identity-by-Descent (IBD) maps from a very competent new preprint at bioRxiv on the admixture history of Siberians. You'll find the full collection in the paper's supp info PDF here.

The results suggest that Russians share minimal IBD with Siberians, despite the well documented presence of significant European admixture in many parts of Siberia. But this isn't really a contradiction because:

Furthermore, our results demonstrate that the European ancestry component detectable in many populations of southern, central, and northern Siberia is not the result of post-colonial Russian admixture as may have been expected [37] and as was suggested on the basis of ALDER analyses [38]. Rather, with the exception of the Dolgans and the Samoyedic-speaking groups of western Siberia, the European ancestry represents one of the most ancient components [dating to not more than 4500 years ago] of the complex admixture history of Siberian populations.

The ~4,500 YBP date makes a lot of sense. This is the Late Neolithic/Early Bronze Age period, which also saw massive population movements from the steppe deep into Europe (see here).

Citation...

Pugach et al., The complex admixture history and recent southern origins of Siberian populations, bioRxiv, Posted April 30, 2015, doi: http://dx.doi.org/10.1101/018770

32 comments:

Davidski said...

This paper references an interesting ancient DNA study from 2006, which appears to catch the moment when south Siberian genetic structure probably shifted from ANE to ENA around 6,000 to 7,000 years ago.

http://onlinelibrary.wiley.com/doi/10.1002/ajpa.20247/abstract

spagetiMeatball said...

That seems to coincide with the time that the secondary (more east asian) native american wave entered the americas.

I think they were Na-Dene speakers? One of the related languages is still spoken on the yenisei river.

Kristiina said...

”Rather, with the exception of the Dolgans and the Samoyedic-speaking groups of western Siberia, the European ancestry represents one of the most ancient components [dating to not more than 4500 years ago] of the complex admixture history of Siberian populations.”

I do not know if that statement is correct if you take note of p. 41 (Fig S11) of the supplementary material. The first admixture event in Selkups and Yamal Nenets is dated 2700 years ago and in Yamal Nenets 2490 years ago as an admixture between the European and Central Siberian components. The next event is dated 1590 years ago in both groups between the first formed ancestry component and the West Siberian ancestry component.
The West-Siberian component is typical of Khanties (nearly 100% in Khanties) and it is formed in K5 from the admixture between European and Central Siberian components. I would assume that the language shift to a Samoyedic type language coincides with this Khanty like Western Siberian admixture event dated 1590 years ago. Moreover, Hungarian, Mansi ad Khanty are on the same East Uralic branch together with the Samoyedic languages.

According to this new paper: http://genome.cshlp.org/content/suppl/2015/02/18/gr.186684.114.DC1/Supplemental_Figures.pdf
the age of Samoyedic N1b is just 942-2653 years, so it could fit with the above-mentioned admixture 1590 years ago. Also the Samoyedic N1c seems to be shared with Khanty and Mansi (N-L1034) and the age (TMRCA) of all N-L1034 is 2,415 years and this could fit with N1c first developing in Khanty and then later mixing into Samoyeds.

Similarly, the age of Yakut and Even N1c is 821-2580 years, and that fits with the second Yakut admixture event 1710 years ago with the Central Siberian ancestry component. The first admixture event happened in a more southern location between a European and a Han like population 2600 years ago.

I really hope that a yDNA analysis will be conducted on Baraba Steppe specimens and other West Siberian specimens in order to better identify the populations that underwent the first admixture event.

Krefter said...

BTW, a Mesolithic mtDNA sample from the neighborhood of Samara Russia, but 1,000-2,000 years older than the one in Haak 2015, is a HV1 match with the one in Haak 2015. Some-type of family or ethnic connection there.

Marnie said...

@Krefter

" a Mesolithic mtDNA sample from the neighborhood of Samara Russia, but 1,000-2,000 years older than the one in Haak 2015, is a HV1 match with the one in Haak 2015."

Which site is it from?

Krefter said...

Lebyazhinka Russia.

Kristiina said...

As regards the Samoyeds, another scenario occurred to me which is probably more plausible. As we know that Samoyedic Ianguages were first spoken in South Siberia, it is possible that the first admixture event happened there 2700 years ago, and the second admixture event happened when they headed north. This interpretation would be in line with what Jaska has explained regarding development of Samoyedic languages. In this model, there would not be a direct genetic relationship between Khanties and Samoyeds although they share the same Western Siberian component.

Moreover, according to the paper, ”all our analyses indicate that the Nganasan, the northernmost indigenous people of Siberia, are quite distinct from the other Samoyedic populations. They lack the Western Siberian and the European ancestry components, which are both characteristic of the other Samoyedic groups. Genetically, they appear to be much closer to the neighboring Tungusic speakers and the Yukaghirs than to the Selkup and even the neighboring Nenets. This leads us to conclude that the Nganasan have a different genetic history than the other Samoyedic groups and that they were linguistically, but not genetically, assimilated by the Samoyedic populations who around 1,000 years ago migrated from the south to the northern Arctic.”

Krefter said...

Just plugged AG2 into the ISOGG Y-Tree AddOn.

What it tells us is he had F. And he is not G or IJK(an ancestor of R). He's negative for several R mutations, and upstream mutations of R.

He is positive for a GHIJK mutations, an HIJK mutation, and an H mutation. He could be HIJK(xIJK) or H. I'd say he was for sure F(xG, IJK).

Krefter said...

Plugged in Sf11(Mesolithic island by Sweden) he is I(xI1, I2a1b, I2a2). Also his mtDNA is for sure U5a1 and probably U5a1f1a.

Davidski said...

AG2 is heavily contaminated, so you might be testing the Y-DNA of someone who dug him up or handled him later in a museum.

Krefter said...

True.

Mike Thomas said...

David

Looking at your original K8 spreadsheet, there doesn;t seem to be many ancient sampls in it - eg Corded Ware or Yamnaya. (Maybe I missed a different one ?)

Is it possible for you to do a couple for CWC, Yamnaya, Unetice, + ust-Ishm, Kostenki, Mal'ta, Hungary KO1, Hungary C01, Hung BR1, La-Brana, Starcevo & an LBK ?
(no rush, whenever you can

Davidski said...

K8 results for some of the Haak et al. samples are here...

http://eurogenes.blogspot.com.au/2015/03/bell-beaker-corded-ware-ehg-and-yamnaya.html

But I'm pretty sure you've already seen them.

Also, all three WHG genomes are basically 100% WHG, while the SHG samples are 15-20% ANE and the rest WHG. The EEF genomes, including Starcevo_EN, are around 75% ENF and 25% WHG. Most of these samples should be in the original spreadsheet.

The K8 isn't relevant for samples older than the Mesolithic.

Aram Palyan said...

"Since the peoples with western and eastern Eurasian ancestries co-existed in the Altai for a prolonged period of time [22], it is plausible that the admixture date of ~2,300 ya estimated for the Altaians captures admixture between these local groups and is not associated with any additional migration from the outside."
From this study.
And this one.

"The regular-sound-change tree estimates a mean divergence time between the outgroup Chuvash and other Turkic languages of 204 BCE, with a 95% credible interval of 605 BCE to 81 CE. This compares to proposals from glottochronological analyses that suggest dates of 30 BCE to 0 CE [21] and 500 BCE to 50 CE from historical data [18, 21, 22]. The sporadic-sound-change model estimates the mean age of the tree to be more than two millennia older (2408 BCE, 95% CI = 3994–1279 BCE), because it wrongly assumes that the many occurrences of regular sound change along the outgroup Chuvash branch are multiple instances of independent phonological change."

http://www.cell.com/current-biology/fulltext/S0960-9822%2814%2901373-6

An interesting case were genetics and linguistics are consistent.

Helgenes50 said...

If the Steppe component is 100% Yamnaya.
This means that we share less with them than in the Haak paper.
The percentage is always high for the CCW, but not for the current Europeans,
nor for the North Europeans.
I don't know where is the truth, but these last results seems to me more credible

Davidski said...

I don't think that the steppe nomads who invaded Central Europe during the Late Neolithic were 100% like the Samara Yamnaya.

I suspect they were similar to that mysterious Karsdorf LN individual.

Helgenes50 said...


I suspect they were similar to that mysterious Karsdorf LN individual.


Those who invaded Western Europe, or their ancestors, probably had an amount of Hindu kush, what is not the case for Karlsdorf.
This component seems to be more related to R1b, like for the BBC.
Is it an idea ?

Davidski said...

Yes, it's possible that they came from further east, closer to Central Asia, but still weren't exactly like the Samara Yamnaya, so the algorithm has to compensate.

Matt said...

Helgenes If the Steppe component is 100% Yamnaya. This means that we share less with them than in the Haak paper.

Good observation, yeah, estimates from this ADMIXTURE are close to Haak's model in some instances (Karsdorf_LN is almost exactly the same), on the other hand, some populations like Norway / Iceland / Bell Beaker have Yamnaya cut in half (for Norway 53.5% in the three way EN, WHG and Yamnaya model or 48% in the five way model plus BedouinB and Nganassan to population average 26% in the Steppe 10). Almost everyone other that Karsdorf_LN has less Yamnaya related ancestry (if that is what the steppe component represents).

http://textuploader.com/x34u

Maybe some of this may be if WHG-UHG is shifted towards the "East" compared to the WHG samples used in the Haak models, so Euro_HG plus Near East works well enough for many populations that there is no need for as much Yamnaya related ancestry. (That's how it looks from PCA) http://i.imgur.com/3BDXlV0.png

Mike Thomas said...

@ Matt

"Maybe some of this may be if WHG-UHG is shifted towards the "East" compared to the WHG samples used in the Haak models, so Euro_HG plus Near East works well enough for many populations that there is no need for as much Yamnaya related ancestry."

That is what I have entertained as possible. But we need full-sequenced need Mesolithic samples from E. Germany, Poland, Ukraine, Moldavia, ..

But i don;t think its just a matter of being EHG/ ANE admixed. Its about the combination of EHG / 'West Asian' , isn't it ?

@ Davidski

"I suspect they were similar to that mysterious Karsdorf LN individual."

yes he has very high Yamnaya-type ancestry, but what are you getting at ?

Helgenes50 said...

The first genetic data on Neolithic settlers in France, in the southern of the Paris Basin.


http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0125521#sec015

Davidski said...

The Bronze Age Pole belongs to R1a1a...

http://polishgenes.blogspot.com.au/2015/05/r1a1a-from-early-bronze-age-warrior.html

Mike Thomas said...

Btonze Age pole:

Would like to see pre-BA Y chromosomes in Poland and (non-Russian) EE in general


French Neolithic data :

0 out of 16 Y chromosomes amplified. Ouch

truth said...

Btw new mtDNA from Neolithic France -->
http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0125521

Marnie said...

Wall Street Journal Article about ancient DNA studies, including the Reich Lab Haak paper:

http://www.wsj.com/articles/ancient-dna-tells-a-new-human-story-1430492134?mod=e2tw

"It turns out that, in the prehistory of our species, almost all of us were invaders and usurpers and miscegenators."

That's funny, I don't recall this scenario of universal and non-stop genocide and invasion being discussed at the SAA. The Wall Street Journal no doubt loves this scenario.

Doesn't the Rockerfeller Foundation fund a lot of ancient DNA studies?

Marnie said...

@WSJ

All now the shameless corporate plug (it is, after all, the WSJ):

"Before these technical innovations, reading DNA required the laborious amplification of short segments, one at a time. By 2008, companies such as 454 Life Sciences in Branford, Conn., and the San Diego-based Illumina began marketing machines that could read millions of DNA samples in parallel. In the past, researchers wanting to study ancient or modern DNA had to sip from raindrops; now they can drink from fire hoses."

Mike Thomas said...

Dave
You said you earlier stated were reading an interesting abstract on Siberia before the colonisation of Americas. Remember what it was ?

capra internetensis said...

This is an awesome paper and the method they are using seems promising.

Sanderson et al have released a tool called adwave to do this kind of wavelet admixture calculation, but I think at present it only does one admixture event. They seem to be using PCA rather than ADMIXTURE to assign ancestry blocks.

http://www.genetics.org/content/early/2015/04/06/genetics.115.176842.abstract

Davidski said...

capra,

PCA isn't very useful in this context when looking for admixture from closely related populations, like for a specific type of European admixture in other Europeans. You need haplotype information for that, and even then the outcomes are often uncertain.

Mike,

First post in this thread.

Mike Thomas said...

Thanks Dave
No it was from a couple weeks back. About palaeo hunter gatherers diffusing to America and Europe ?

Mike Thomas said...

Found it
From your IBS thread

capra internetensis said...

@Davidski

What kind of haplotype method are you referring to?

AFAIK all of these admixture dating methods will have trouble with closely related ancestral components.