search this blog

Wednesday, August 19, 2015

No significant genetic substructures within (eastern) Yamnaya


It's interesting and perhaps important that there's practically no difference between the two sets of Yamnaya samples published to date, despite the fact that they're from regions separated by ~1,000 kilometers. Yamnaya_Rise is from between the Black and Caspian seas, while Yamnaya_Haak from just north of the Caspian.

Note the lack of significant Z scores (around 3 or more) in these D-stats. Although the almost significant Z score with Loschbour, a Mesolithic Western Hunter-Gatherer (WHG) from Luxembourg, does look somewhat curious.


It's difficult to know what all of this means exactly without seeing any ancient samples from archaeological cultures that preceded Yamnaya on the Pontic-Caspian Steppe. However, it could mean that the Yamnaya nomads arrived north of the Caspian from the south, which is also what preliminary Y-chromosome data is hinting at (see here).

Yamnaya was succeeded northeast of the Caspian by its offshoot, the Poltavka Culture, which in turn was replaced by the Sintashta Culture. The most widely accepted theory, based on archaeological data, is that Sintashta formed from a chain of cultures derived from the late Corded Ware horizon of East-Central Europe. This is backed up by the D-stats below, which suggest some western admixture in Sintashta that is missing in Yamnaya.


In this analysis I used samples from the Allentoft et al. (Rise Project), Haak et al. and Lazaridis et al. datasets, all of which are publicly available. The latter two are found at the Reich Lab site here.

See also...

A multidimensional approach

44 comments:

Alberto said...

Thanks David.

I agree with what you wrote. But perhaps the most surprising stat is the second one of the second set. Corded Ware is closer to Yamnaya than to Sintashta? I would have expected that stat to be significantly negative, not almost significantly positive.

Sintashta indeed looks very western, but so western (compared to CW from Germany) as to explain that stat above?

It would be interesting to compare Corded Ware to Sintashta directly (as Pop1 and Pop2)? The PopTests you chose should be fine to get an idea of where the differences between them might lie.

a said...

Many thnx David, your work and dedication is unsurpassed. Who would have guessed these results?

Sites like Samara & Khvalynsk[proto-Kurgan?];look like they have C14 dates from 2 relatively close sites and plenty of samples. Wow, some of the dates are 3000+/- years older than Sintashta farmers and or Corded[can you believe it?], who look like the new kids on the block in and around that region[East&West]. Although, going by pottery, I'd venture that Samara might push into 6k+/-B.C. just a wild guess. I guess we will have to wait for more samples.

The Khvalynsk culture extended from Saratov in the north to the North Caucasus in the south, from the Sea of Azov in the west to the Ural River in the east.

A good sprinkling of calibrated C-14 readings obtained from material in the graves of the type site date the culture certainly to the approximate window, 5000-4500 BC. This material is from Khvalynsk I, or Early Khvalynsk. Khvalynsk II, or Late Khvalynsk, is Late Eneolithic.

"Some regard Khvalynsk I as Early Eneolithic, contemporary with the Samara culture. Gimbutas, however, believed Samara was earlier and placed Khvalynsk I in the Developed Eneolithic. Not enough Samara culture dates and sites exist to settle the question.

The Khvalynsk type site is a cemetery, 30 m by 26 m, containing about 158 skeletons, mainly in single graves, but some two to five together. They were buried on their backs with knees contracted. Twelve of the graves were covered with stone cairns. Sacrificial areas were found similar to those at Samara, containing horse, cattle and sheep remains."

"A good sprinkling of calibrated C-14 readings obtained from material in the graves of the type site date the culture certainly to the approximate window, 5000-4500 BC"

Arch Hades said...

"The most widely accepted theory, based on archaeological data, is that Sintashta formed from a chain of cultures derived from the late Corded Ware horizon of East-Central Euro".

There is archaeological evidence for this? I didn't know that. Got a link to any sources? Anyway, the genetic data from Allentoft sure backs it up.

a said...

"The most widely accepted theory, based on archaeological data, is that Sintashta formed from a chain of cultures derived from the late Corded Ware horizon of East-Central Euro".

Missing link?



Comparing Sintashta and Vedic: rá-tha denotes a war vehicle?
Comparing Sintashta design to Vedic and Neša designed/chariots.
2 persons- Sintashta?
3 person- trivandhurá (1.41.2; 7.71.4)
3 person- Neša &battle of Kadesh.
Does not look like a solid link.How about C.W. Chariots?

Matt said...

Alberto: But perhaps the most surprising stat is the second one of the second set. Corded Ware is closer to Yamnaya than to Sintashta?

True point.

For other stats, in Allentoft 2015 - Sintashta got the same as Bell Beaker in the stat D(Yoruba,Yamnaya,Hungary,Pop), indicating that it was less close to Yamnaya than Corded Ware and as far as from Yamnaya Bell Beaker was

http://www.nature.com/nature/journal/v522/n7555/fig_tab/nature14507_ST1.html

This is also reflected directly in their supplement where D(Yoruba, Yamnaya, Corded Ware, Sintashta)= -0.0036, Z=-1.45, at a low level, while D(Yoruba,Yamnaya, Bell Beaker, Sintashta)=0.0008, Z=0.292 (no significant difference).

... but at the same time, on those stats, Sintashta seems to be a little further from Neolithic Central European farmers than Corded Ware, judging by the above comparison of their stats D(Yoruba,Neolithic Central)(Pop,Yamnaya)... and directly via D(Yoruba, Neolithic Central European, Corded Ware, Sintashta) = -0.0032, Z=-1.15. Likewise D(Yoruba, Neolithic Central European, Bell Beaker, Sintashta) = -0.0142, Z=-4.85 (which is probably significant).

Suggests possibly some possible minority non-MN European and non-Yamnaya gene flow into these Sintashta samples. Otherwise hard for Sintashta to be both further from Yamnaya than Corded Ware and same distance as Bell Beaker, while also further from Central European Neolithic than either of these two. Although I would stress this would be minority (and Sintashta certainly is closer to European Middle Neolithic than Yamnaya). There's also that peculiar "purple" in their ADMIXTURE in Allentoft et al. May still just be a statistical oddity - I'm not sure that Allentoft only ran the transversion SNPs like Davidski has.

On another note, also interesting that Sintashta looks no closer to or further from Iraqi Jew or Georgian than Yamnaya.

Davidski said...

As per above, Sintashta is an offshoot of late Corded Ware. The Corded Ware samples from Haak are from fairly early sites, so they're very Yamnaya-like, and hence more similar to Yamnaya than to Sintashta. The Corded Ware from Rise are more of a mixed bag, and thus symmetrically related to Yamnaya and Sintashta, although a little closer to the latter.

Yoruba Corded_Ware_Haak Sintashta_Rise Yamnaya_Haak 0.0177 3.251
Yoruba Corded_Ware_Rise Sintashta_Rise Yamnaya_Haak -0.0023 -0.377
Yoruba Corded_Ware_Haak Sintashta_Rise Yamnaya_Rise 0.0064 1.081
Yoruba Corded_Ware_Rise Sintashta_Rise Yamnaya_Rise -0.0016 -0.295

Chad Rohlfsen said...

I have some of the Rise samples as more West Asian like, with the Haak being slightly more EHG on average. I'll try to get something posted ASAP.

Romulus said...

Is there any differentiation between the catacomb I2a2 and the rest of the Yamnaya in terms of autosomal affinity?

Davidski said...

Strictly speaking there aren't any differences between the RISE552 I2a individual and the combined Yamnaya set. The differences in affinity to Corded Ware and Loschbour are close enough to significant, but the fact that it's an individual might be skewing things a bit.

Yoruba Afanasievo_Rise RISE552 Yamnaya 0.004 0.624
Yoruba Corded_Ware_Haak RISE552 Yamnaya 0.0192 2.923
Yoruba EHG RISE552 Yamnaya 0.0133 1.654
Yoruba Koros-LBKT-Starcevo RISE552 Yamnaya -0.0028 -0.208
Yoruba Loschbour RISE552 Yamnaya 0.0224 2.878
Yoruba MA1 RISE552 Yamnaya -0.001 -0.111
Yoruba Iraqi_Jew RISE552 Yamnaya 0.0086 1.956
Yoruba Georgian RISE552 Yamnaya 0.0081 1.925
Yoruba Kalash RISE552 Yamnaya 0.0063 1.515
Yoruba Lezgin RISE552 Yamnaya 0.0055 1.323

Also, the sample from the Stalingrad quarry doesn't really deviate much from the combined Yamnaya set.

Yoruba Afanasievo_Rise RISE555 Yamnaya -0.0035 -0.28
Yoruba Corded_Ware_Haak RISE555 Yamnaya 0.0207 1.57
Yoruba EHG RISE555 Yamnaya -0.0017 -0.11
Yoruba Koros-LBKT-Starcevo RISE555 Yamnaya -0.0253 -0.989
Yoruba Loschbour RISE555 Yamnaya 0.0198 1.424
Yoruba MA1 RISE555 Yamnaya 0.0072 0.43
Yoruba Iraqi_Jew RISE555 Yamnaya 0.0033 0.402
Yoruba Georgian RISE555 Yamnaya -0.0019 -0.255
Yoruba Kalash RISE555 Yamnaya 0.0059 0.757
Yoruba Lezgin RISE555 Yamnaya 0.0016 0.215

a said...

Comparing using K6 averages.Using rough calc's
R1b-Z2103 cluster[5samples]3k-3.3k+/-B.C.]
9.66/79.42

Rise I2a sample 552
18.73/71.25

Rise 397 sample [Armenian]
59.56/36.26

Samara H.G.
0.000/80.79

Yamnaya I0429
7.78/81.43

Clearly using K6 R1b-Z2103 cluster average are closer to Samara H.G. both in time/au- ratio and in some cases geographical position[
Yamnaya I0429 sample is within 1.5+/-km.]than the R1b-z2106[18249219snp] twig and I2a.
One of the major branches in Armenian today is R1b-z2103+


https://docs.google.com/spreadsheets/d/17FnR-OeYN4m-XCzWxChiLPJR6QW_ehmR3I0Mrz6OjDI/edit#gid=1930035171
http://eurogenes.blogspot.com.au/2015/07/high-res-r1b-tree-featuring-16-ancient.html

a said...

RISE_baAfan average are also more closer to R1b-Z2103 K6 cluster when compared to I2a and R1b-z2106.

RISE_baAfan-
10.19/78.16

R1b-Z2103 cluster[5samples]3k-3.3k+/-B.C.]
9.66/79.42

Rise I2a sample 552
18.73/71.25

Rise 397 sample [Armenian]
59.56/36.26


Samara H.G.
0.000/80.79

Yamnaya I0429
7.78/81.43

a said...

Or put another way-k6 separated,by 3k kilometers and 23oo years-;
M.E._10.19%+/-& E._2.63+/-% variance.
Samara H.G.R1b
0.000/80.79[7650 +/-B.P]
Yamnaya Z2103-R1b[avg.5]
9.66/79.42[5300+/-B.P.]
Afanasevo[avg.3]
10.19/78.16[5300+/-B.P.]


Balaji said...

Thanks. This is more evidence that Sintashta or Andronovo were not ancestral to Kalash or Pathan.

a said...


Blogger Balaji said...

" Thanks. This is more evidence that Sintashta or Andronovo were not ancestral to Kalash or Pathan."

What about option "C" R1b-Z2103? Something along the lines of R1b (ht35) Haplotype #5 found in the project below?

19 389i 389ii 390 391 392 393 385a 385b
14 13 29 24 11 13 12 - -
Geographical Locale
%
Pakistan [Kashmiri] 16.67
Macedonia, Greece 7.14
Ankara, Turkey 5.13
South Caucasus [Armenian] 5.00
http://freepages.genealogy.rootsweb.ancestry.com/~gallgaedhil/haplo_r1b_ht35.htm

Davidski said...

Settle down, no one's saying that R1b, and especially R1-Z1203 isn't a Proto-IE lineage.

It very likely is. The fact that R1a badasses Sintashta pushed into the Trans-Ural steppe and then swept across Asia doesn't mean it isn't. It's just another chapter of the early IE story.

a said...


^^^
Davidski said...
As I recall it wasn't that long ago fine badasses Steppe specimen's as yourself were fending off every possible origin other than Steppe[Sintashta blowing away R1a boys highest expectations ]. Both, you and I know very well the general 1k question can only come to one logical conclusion; why these guys are so stubborn and refuse to take up Anthony/Mallory in light of genetic evidence you have produced? Only the details remain/have to be worked out[IMO]. I'm looking forward to when Eurogenes can get some more ancient samples; specifically the higher up early branches of our ancestors, perhaps even including R2 & V88 boys, that would be fantastic and perhaps resolve all lot more of our ancestral questions Steppe origin compared to other regions.

Anyway, whether you use K6 or K15, the fact remains, properly parsed Yamanaya/Sok H.G. R1b[future L51/z283 still to be resolved withing Yamnaya sphere] data points to the highest % and oldest Eastern European cluster; while Sintashta & C.W. is the highest % recent European cluster, that in my opinion can only point to our ancestors being/moving around the Steppe supporting our autochthonous status to the Steppe region, and feeding R1a/R1b/R2 lines South/West and East[ politically correct monikers?][IMO].
BTW,it would be nice to have an exclusive R1a&R1b&R2 board to discuss this without all the looons/teapots,who do not have any personal connection to the subject, other than a passing whim, and to call someone Euro-centric and/or some other name[one does get tired of all the name calling], for no reason. After all we are just trying to find out our ancestry like everyone else [I may joke around, but you know exactly what I'm talking about].
But I know it's not going to happen anytime soon. just my 2cent

Mike Thomas said...

Blogger "a"

Sorry to burst your bubble, but your ancestry goes back to some peasant farmer who lived c. 1500 AD. You personify the 'loons' & 'teapots' who give genetics a bad name, because you conflate a population genetic prehistory with some trumped up personal genealogy.

a said...

@ Mike Thomas said...

Okay Mike I will add that to the long long list. BTW how did you become interested in Yamnaya,as a subject? Are you looking into your family history also?

Mike Thomas said...

@ "a"

What list ?

"as a subject? Are you looking into your family history also?"

No . I'm not under any pretense that someone from 5, 10 or 20, 000 years ago has anything directly to do with my "family history".

Krefter said...

"Part 1: Looking For mtDNA Diversity in West Eurasia"

http://www.anthrogenica.com/showthread.php?5225-Part-1-Looking-For-mtDNA-Diversity-in-West-Eurasia&p=104125#post104125

I just began my own Blog named "mtDNA Atlas' to provide mtDNA data from around the World.

BTW, Yamnaya's closest maternal relatives are Volga/Ural people. The frequency of "Steppe mtDNA" is consistent with autosomal DNA.

a said...

Krefter said...


I just began my own Blog named "mtDNA Atlas' to provide mtDNA data from around the World.

Good job:)
You have a much better understanding/interest in mtDNA. In fact I think we share the same line :)U5b2a2[Polish/German/Prussian].
A couple of really odd questions, maybe you can answer;maybe not.
Question 1, is there any relationship to early pottery and certain mtDna branches?
Question 2, Is Georgian/Altain/Kennickwick man & {Ojibway} the same branch X; if so what is the date separating all three locations/branches[roughly]?

a said...

Mike Thomas said...

@ "a"

What list ?

My list of flaws, naturally : )

Krefter said...

@a,
"Question 1, is there any relationship to early pottery and certain mtDna branches?
Question 2, Is Georgian/Altain/Kennickwick man & {Ojibway} the same branch X; if so what is the date separating all three locations/branches[roughly]?"

I don't know.

"You have a much better understanding/interest in mtDNA. In fact I think we share the same line :)U5b2a2[Polish/German/Prussian].'

Yes I have U5b2a2b1. Mine is Prussian. U5b2a2 is most popular around Germany, Poland, BeloRussia, and Scandinavia.

Krefter said...

New ancient mtDNA from Romania. Early Neolithic-Late Bronze age.

http://forwhattheywereweare.blogspot.com/2015/08/ancient-mtdna-from-romania.html

Mike Thomas said...

"A"

I know you're joking, but nevertheless , a "flaw" implies something personal, like being messy. I don't know anything about that

All I was initially critiquing was your belief that just because R* in Mal'ta, it doesn't mean that Yamnaya descend directly from him, or that they were part of the sample population group/ "kin", or that there has been continuous, uninterrupted habitation of Siberia for 25000 years, and much less that they spoke the same language. That's all

Krefter said...

There's an Indian specific R7 from Late Neolithic Romania. That's a one in a million. Although it reminds me of Indian-specfic mtDNA found in Bronze age Near East.

Roy King said...

@Krefter
very interesting article on the mtDNA of Middle/Late Neolithic Romania. I do wish there could be more discussion in this blog of other, non-steppe related migrations and demography. I've always said and felt that a more nuanced approach carries merit. The movement of populations from NW Anatolia post-Early Neolithic is intriguing and bears major relevance to the present distribution of Y chromosome SNPs and mtDNA markers (and ultimately autosomal PCAs and admixture patterns)in the Balkans. The linguistic correlates are also potentially fascinating and, IMO, independent and different from the swathe of IE language replacement in Europe.

Davidski said...

Which Romanian archaeological culture does it come from? And which ethnic groups in India is it found in?

Mike Thomas said...

That article isn't so new
We already know from other studies (eg Unterlander's abstract) that there was more movement into Europe during the middle Neolithic. This might correlate with Y haplogroups J2 and E-V13 otherwise not yet found in early Neolithic.

Gioiello said...

It is true that these mutations found in SMR-2 :G16319A T16362C and C16261T, are found in R7a’b, but lacks C16260T and above all R7 has 73G and we have 73A, thus I doubt that this sample is an R7. Anyway Maju said: that these samples “reflect greater degree of ‘Europeanization’”, explaining in detail why.

Chris Davies said...

"There's an Indian specific R7 from Late Neolithic Romania. That's a one in a million. Although it reminds me of Indian-specfic mtDNA found in Bronze age Near East."

Indian-specific mtdna R7 was also found in one Tutsi in Rwanda.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2720951/

[The same study also mentions mtdna U9 found in both Ethiopia and India, and M33 in both Egypt and India].

Simon_W said...

What the currently available ancient DNA evidence in my opinion suggests with regards to the PIE question (sorry @ Roy King!):

- The Indo-Iranians didn't come from West Asia. R1a variants typical for them have been found in Sintashta people who were Northeast European-like in their autosomes. Moreover, the fact that they were more like regular Europeans than the Yamnaya people suggests that they were intrusive, but from the west – not from the southeast.

- The Armenians didn't come from West Asia. The only Iron Age Armenian available has a slight shift towards Bulgarians and northern Greeks in the genomic West Eurasian PCA. That's hardly a coincidence. (He also has a yDNA which today is most common in southeastern Europe.)

So the only candidates left for genuinely West Asian IEs would be the Hittites/Anatolians and the Tocharians. It would be very interesting to see an early Hittite genome.

However, until this will come true I note the following which favours a steppe origin of the PIE:

- R1b-P297 was present in an EHG with no West Asian autosomal admixture. It seems to be a steppe marker, possibly derived from central Asian predecessors.

- Early Bronze Age Hungarians, who were presumably IEs of an early Italo-Celtic variety, had a hefty dose of western steppe admixture, but only very slight Armenian-like admixture, perhaps 4% (according to Alberto's 4mix analysis with K8 input). While it's still possible that their language had come from these 4%, it has to be said that they were not dominated by West Asian yDNA.

Gioiello said...

@ Simon_W

Whereas I totally agree with your analysis of the Sintashta origin (not sorry @ Roy King), I think that your analysis of the aDNA from Samara is a little premature.
“R1b-P297 was present in an EHG with no West Asian autosomal admixture. It seems to be a steppe marker, possibly derived from central Asian predecessors”
I0124 is an R-P297 (13300 years old) intermediate to R-M478 (7300 years old) and doesn’t demonstrate anything about R-L51 and subclades (that I think won’t be found in Eastern Europe) and it isn’t said that the same R-M478 is older in Asia than Europe.

Simon_W said...

Well, L51 is still derived from P297, so it can't said that it's completely unrelated. Every person with L51 also has P297. It might be speculated that P297 had a distribution that stretched from the North Caspian via central Asia to the South Caspian and to West Asia. But I just wonder why the Samara EHG had P297 if he was completely a dead end. Was there something that important about this marker that caused it to be that widespread in Paleolithic foragers? Hardly. If the Samara P297 was completely a dead end, I would rather expect him to have been R1b1d or something like that.

Simon_W said...

Some further thoughts:

People like Nirjhar might suggest as an alternative scenario for Sintashta that they were European intruders ruled by a thin elite of South-Central Asian origin who dominated their yDNA. However, this is unlikely for two reasons: First, their yDNA haplogroups didn't split from European R1a long ago, to the contrary, they're very recent spin-offs. And second, it's unlikely that recent intruders from Europe with a demic impact met recent intruders from SC Asia without demic impact at about the same time.

My personal guess is that the early Hittites will be similar to the Bronze Age Armenians, and thus will be suggestive of an origin close to the North Caucasus. Maybe even the Bronze Age Armenians spoke an unattested language related to Anatolian IE! I think this hypothesis makes sense because in general European admixture in Anatolia is weak. In contrast Caucasus-like admixture is very strong, if you bear in mind that the early farmers from Anatolia were still close to Sardinians. People speaking Anatolian IE may also be responsible for at least some of the Caucasus-like admixture in modern Greeks. I've read there are place names of Anatolian IE origin in the Aegaean.

Gioiello said...

I didn’t say that I0124 was a dead line. It may have been in the bush who brought to the R-M478, but for what we know so far only that line survived. Thus it is important having found it in Samara about 7500 years ago. But the ancestor of R-M478 may have lived also elsewhere, both in Asia but also in Europe. We don’t know. Of course R-L51 has had its pathway, and the map that Argiedude and me did and also that of Rocca’s demonstrate that it is a Western or Central European haplogroup. I said that Eastwards Italy it is at 0,00% percentage, and the smal’s tree (R1b1a2 (P312- U106-) DNA Project” demonstrates so far that the subclades of the parallel R-PF7589* are above all in Italy, and mostly in Tuscany, so I expect that it will be found in the ancient DNA. If not, very likely I’ll have to correct my theory.
But too many approximation in the peer reviewed paper. After the R-L51 found in the Balkans which were R-L23 “Balkan cluster” also this mt R7 found in Romania. Nobody replied my post, but they use the restriction enzymes yet, and also here perhaps too much approximation.

Krefter said...

Mark my Words: The upcoming Neolithic Anatolian/Greek genomes will belong to what today are European-specific mtDNA haplogroups(Incluing H subclades typical of Europe). They might be the biggest contributor to Modern European mtDNA.

Their mtDNA will probably be mostly made up of: H1, H5, T2(especially T2b) J1c, K1a, and HV0.

Krefter said...

Something I think isn't very well known is: H1 takes up almost 50% of modern European H and It had a strong presence in Neolithic Europe. There's no way they got their H1 from hunter gatherers, they introduced it from Anatolia.

Of Neolithic Hs tested for the H1 SNP.
2/8 LBK have H1
10/20 from Early Neolithic France.
3/6 from Middle Neolithic France.
1/1 from Megalithic Spain.

But from Yamnaya/Catacomb.
2/19 are H1.

Mike Thomas said...

Krefter
Prediction duly noted
What specific H clades were mesolithic samples from Iberia and the Baltic ?

Krefter said...

It looks like the one from Karelia was H2a2b. None of the Rs from Western Russia(where R1a, N1c were found) could find an H subclade with HV1 coverage. One of the Upper Palaeolithic Hs from Spain looks like H6.

But I'm very skeptical of reported Hs from pre-Neolithic Spain, even from a study published this year. I'd be very excited if pre-Neolithic H was found, so I'm not driven by an anti-H attitude.

http://www.anthrogenica.com/showthread.php?4246-19000-year-old-mt-DNA-Haplogroup-H-from-the-Franco-Cantabrian-site-of-el-Miron!!!!

Mike Thomas said...
This comment has been removed by the author.
Gioiello said...

@ Krefter

“Vladimir Tagankin who's [whose] username at this forum is smal”
Are you saying that Vladimir Tagankin and Sergey Malyshev are the same person? Are you like the woman whose name we cannot name here?

Gioiello said...

Perhaps you did mean "s[e]ma[rg]l".

mementas said...

Kind of surprised to read this. Yamna archeology shows there were all sorts of cross interacting across the Pontic river basins. Some assumed these people had differing ancestral backgrounds. Looks like that wasn't the case and early Yamna were by and large the same people.