search this blog

Wednesday, June 10, 2015

101 ancient Eurasian genomes (Allentoft et al. 2015)


It'll take me a while to digest all of the information in this massive new Allentoft et al. paper. But I've already noticed that, just like in Haak et al. 2015, the Yamnaya samples are again from the eastern half of the Yamnaya horizon. This time, however, not all of the Yamnaya individuals carry Y-haplogroup R1b; one of the five samples belongs to Y-haplogroup I2a (see here).

So I'm wondering what more westerly Yamnaya sites will reveal in the future, considering the predominance of Y-haplogroup R1a among the Corded Ware individuals sampled to date, and the close genome-wide relationship between the Yamnaya and Corded Ware?

Abstract: The Bronze Age of Eurasia (around 3000–1000 BC) was a period of major cultural changes. However, there is debate about whether these changes resulted from the circulation of ideas or from human migrations, potentially also facilitating the spread of languages and certain phenotypic traits. We investigated this by using new, improved methods to sequence low-coverage genomes from 101 ancient humans from across Eurasia. We show that the Bronze Age was a highly dynamic period involving large-scale population migrations and replacements, responsible for shaping major parts of present-day demographic structure in both Europe and Asia. Our findings are consistent with the hypothesized spread of Indo-European languages during the Early Bronze Age. We also demonstrate that light skin pigmentation in Europeans was already present at high frequency in the Bronze Age, but not lactose tolerance, indicating a more recent onset of positive selection on lactose tolerance than previously thought.

Allentoft et al., Population genomics of Bronze Age Eurasia, Nature 522, 167–172 (11 June 2015) doi:10.1038/nature14507

See also...

R1a-M417 from Eneolithic Ukraine!!!11

728 comments:

«Oldest   ‹Older   401 – 600 of 728   Newer›   Newest»
Aram said...

Kurti
If You believe that the Armenian R1b is not from Steppe ( and it seems Krefter agrees with You ) then this could have a dramatic effect on the question of PIE locus. Because in the current stage of are knowledge Yamna is heavily dominated by R1b-L23.
If West Asian R1b and Yamna are brothers ( and Yamna is not the parent) then things change a lot. The question of Armenian language is a minor question in all this puzzle. Let's on this point concentrate only on the genetics.


Davidski said...

Yes, you're definitely onto something there postneo. Sintashta were either Finno-Ugrics or Turkics.

And you're a teapot.

Gioiello said...

@ Davidski
You should know that your spreadsheet is an artifice. It would be the same if Nirjhar007 said that, because Indian Raza and Joshi are demonstrated to be R1b1-L389-, all the modern R1b1a2 subclades come from India. The fundamental point of my theory was just that the modern subclades descend from R1b1-L389+, which has the highest variance in Italy and is absent in India. The same is true for R-L51, absent so far in Eastern Europe, born very likely in Italy (I did the first map with Argiedude, and it is my avatar now in some FB blogs). That Bell Beakers in Central Europe in the Late Neolithic were R-L51 subclades is now a proof, demonstrated without any possible doubt.

Davidski said...

^ Italian teapot.

Unknown said...

Gioiello

Can you link your proposal for R1b Italian refuge with any specific period and material / archaeological evidence ? (whilst we wait aDNA)

Gioiello said...

I think you should look at the tree that smal (Sergey Malyshev) is doing on "R1b1a2 (P312- U106-) DNA Project", tree that I contributed to begin on Anthrogenica: after, as I am not subdue to the interests of the greatest firms in this field,...
There all is clear and I wrote a lot about that tree on eng.molgen, if once it will look at the light/life and so on.

Nirjhar007 said...

David,
I don't know but according to Kuzmina-
// There is however linguistic evidence of a list of common vocabulary between Finno-Ugric and Indo-Iranian languages. While its origin as a creole of different tribes in the Ural region may make it inaccurate to ascribe the Sintashta culture exclusively to Indo-Iranian ethnicity, interpreting this culture as a blend of two cultures with two distinct languages is a reasonable hypothesis based on the evidence.//
and we KNOW that No F-U elements exists either in Indic or Avestan etc, so the common logic will be?

Anonymous said...

@Davidski
Thanks for the "preview". They are obviously Dravidians or Altaics :). If R1a-Z93 is confirmed in Sintashta and Andronovo the OIT proponents have to give up but i doubt they will accept their defeat.

Gioiello said...

@ Davidski

I don't understand what do you mean with "teapot" beyond its known meaning. Perhaps you should use "teapoy", from Hindi "tipai", Skr. tri "three" and pada "foot". In Tuscany we say "treppiede", and perhaps our origins would be clear to all us. I understand that I am born in the town of Galileo Galilei, but you have had Koeppernick, and nobody should forget that every theory may be verified from experiments.

Anonymous said...

@Nirjhar
There is basically no Finno-Ugrian substrate or just tiny traces of it in modern Russian language except of some place/river names so do you think that Russians had no contact with Finno-Ugrians in the past? There are many Indo-Iranian loawords in Finno-Ugrian languages and they are much older than Indo-Iranian loanwords in Dravidian or Austro-Asiatic languages so Indo-Iranians had earlier contact with FUs than with native South Asians

Nirjhar007 said...

Skilur, I'm not an OIT proponent but its obvious that you have no sense practical approach at all...

Nirjhar007 said...

No most of them can be easily be termed as Vedic or even middle-IA except perhaps the word for wheel but its a superstrate influence from outside like Mitanni.....

Unknown said...

Sintashta Oracles showing a pattern. K12b has the Gedrosia.

K13 Eurogenes
88.2% Swedish + 11.8% Tabassaran = 9.56
K15 Eurogenes
88.5% West_Norwegian + 11.5% Tabassaran = 6.76
Dodecad K12b
88.2% Swedish + 11.8% Balochi = 5.33

Matt said...

David, think you'll be in a position to run qpAdm on Sintashta or Andronovo? To see how close they are to your "Corded Ware sample" as "0.73 Yamnaya + 0.27 Esperstedt_MN" finding (http://eurogenes.blogspot.co.uk/2015/03/modelling-yamnaya-with-qpadm.html)

la señora bibiloni said...

Something I find intriguing: Unetice is supposedly ancestral to Urnfield, Hallstat and La Tene, so we always assumed it was an Indoeuropean culture. But no R1a/ R1b has been found in Unetice - only samples of good ol' I2. What's going on here?

Simon_W said...

@ Colin Welling

According to Haak et al. the Scottish top the list of Yamnaya ancestry on the British Isles with 48.6% according to Figure S9.25. Unfortunately this figure doesn't include Germans and Austrians. It does include Czechs, though, who have slightly more than that, with 48.8%. Now you might rightly object that Czechs, although geographically in central Europe, have eastern European Slavic admixture.

German Corded Ware is modeled as 79.1% Yamnaya, and German Bell Beaker as 50.5% Yamnaya in chapter S9. So you can see a dilution of Yamnaya ancestry from Corded Ware to Bell Beaker. And if the Scottish were nearly pure Bell Beakers from Germany, they might just fit. However, Unetice from the early Bronze Age in central Europe is modeled as 57.1% Yamnaya, and judging from Figure 3 Halberstadt_LBA from late Bronze Age Germany has even more Yamnaya ancestry than Unetice. So whatever Yamnaya ancestry modern Germans may have, Bronze Age central Europeans had more than enough to explain the Yamnaya ancestry in northwestern Europeans.

Indeed, Corded Ware itself didn't expand to northwestern Europe. But neither did Yamnaya! It was post-Corded Ware central Europeans who expanded to northwestern Europe. Obviously they already started doing so in the Bell Beaker period.

And that there were recent founder effects and y-DNA bottlenecks in most European populations is no longer a theory, and neither is the recent rapid bifurcation of R1a and R1b:
http://dienekes.blogspot.com/2015/05/more-y-chromosome-super-fathers.html

Of course there was a massive migration that caused a dramatic autosomal change, but this doesn't preclude that the yDNAs go back to just a couple of successful patriarchs.

Indeed there was Yamnaya in Hungary, possibly with R1b. So it's not impossible that Bell Beaker R1b is from these, though it might just as well be from Corded people. In any case Hungary wasn't overflowing with R1b, that much can be said right now... And as Allentoft et al. have just repeated: The highest level of Yamnaya affinity is found in Corded people, the lowest in Bronze Age Hungary, while German Bell Beakers were intermediate.

Ultimately both Northwest European R1b and Corded Ware R1b were from Yamnaya.

Simon_W said...

The relatively low Yamnaya ancestry in Bronze Age Hungary, combined with the absence of R1a and R1b so far, seems to confirm the conclusion drawn by the archeologist Alexander Häusler, namely that Yamnaya had no lasting effect in the Carpathian Basin and that it was quickly assimilated by the local cultures.

Simon_W said...

@ Matt

In your PCAs it's interesting how the Iron Age Armenians are shifted relative to the Bronze Age Armenians. I'd say, they are at the same time more EEF and more WHG.

Krefter said...

La,
"But no R1a/ R1b has been found in Unetice - only samples of good ol' I2. What's going on here?"

There's only 3 Y DNA from Unetice. Autosomally they were similar to Bell beaker and Corded Ware. There was admixture with local people, so that's probably where the I2 comes from.

Krefter said...

Ancient West Eurasian Y DNA. I added the Allentoft data.

https://docs.google.com/spreadsheets/d/12G2cfjG0wHWarsl5bB99ridFmvUWzqlZfZ6_e_R6oIA/edit

Simon_W said...

@ Mike Thomas

If Genetiker is right and there is R1a-2124 in Sintashta (of which I'm not sure), why would you think the upstream mutations would need to be common in Europe? If they experienced major founder effects only in Asia... At least they do occur in Europe, rarely.

And regarding the map figure in the Allentoft paper, no I think the arrow from Yamnaya to Afanasievo is correct. The Afanasievo samples don't have the EEF admixture that's present in Sintashta, so they're not like Corded Ware.

Nirjhar007 said...

Simon, Z-93 and its Subcldes in Europe are too ''rare'' to be of European origin....
Afanaseivo can be an Extension of steppe cultures but we need their Y-DNA.

Kristiina said...

Skilur, yes, everybody expected to see R1a1 in Sintashta, but the “bomb” is that Sintasha R1a1 may have arrived from the west and went from Sintashta to India. Many would have expected Z93 to have developed in Central Asia without western ancestors (this is what I thought), and this new northern route is interesting from the linguistic point of view, and also from the Uralic point of view.

Mike, yes, N1c from 2500 BC was found near Smolensk, and that is why I proposed that it arose somewhere east of Belarus. Yes, I think too that FU is quite recent movement. It looks like it spread in the taiga area during the Iron Age. According to the recent Pugach et al. paper, the first Samoyedic admixture event happened in South Siberia between European and Siberian ancestry components c. 500 BC.

However, in addition to this Uralic story, N1c and N1b (as well as R1a1-Z93), took part in the Turkic invasion of Siberia and Xiognu campaigns (209 BC — 93 AD) and this opens up interesting linguistic perspectives considering the typological similarities between Uralic and Turkic languages.

As for Indo-Iranian, I think that if Sanskrit developed from the language of Sintashta people (2100–1800 BC), the Iranian branch should have developed on their way from Sintashta to India, i.e. in the east. Armenian is not an Indo-Iranian language, so it could well have spread to Armenia directly from the Yamnaya area before 1000 BC.

Nirjhar, yes, I am very patient, and I am waiting for plenty of other findings as well.

Gioiello, yes, we have to wait for more ancient yDNA from Italy and the surroundings, but I am making all my hypotheses on the basis of the current evidence which can of course change in the future. I am open to change my views.

Matt said...

SimonW: In your PCAs it's interesting how the Iron Age Armenians are shifted relative to the Bronze Age Armenians. I'd say, they are at the same time more EEF and more WHG.

Yep, by the raw f3 stats, yes. Compared to Bronze Age Armenia, Iron Age Armenia falls quite a lot closer to Neolithic Central Europe (closer than any present day West Asians / South Caucasians / North Caucasians do) and a bit closer to WHG (close as North Caucasians). I think the magnitude of shift towards EEF is larger than towards WHG. They just look shifted on a cline towards EEF to me, on those PCA of the f3.

I A Armenia also seems to be closer to Yamnaya than B A Armenia, while being a little further from Mal'ta (http://i.imgur.com/Got21JL.png). I A Armenia seems to have much fewer SNPs available for comparison though by about a factor of 5 though on the comparisons.

Nirjhar007 said...

Kristiina,
//Sanskrit developed from the language of Sintashta people (2100–1800 BC)//
you mean Indo-Iranian well Sintashta is very young the archaeology to India is a dud and highly artificial. false claims have been made by AIT proponents to Fit Sintashta with Aryan Genesis etc etc, will tell later in Details:) i'm tired now....

Simon_W said...

Kurti,

I've never said that all R1b in West Asia came from the steppe and with IEs. I've just suggested this for R1b-L23.

The Iron Age Armenian has 23.1% of the Dodecad K12b North_European component. It's true that North Caucasians have similar amounts, the Chechens for example score 23%. But Armenians don't live in the Northern Caucasus, they live south of both Caucasus chains and historically in eastern Anatolia. So I can't see why you find it natural to compare them with North Caucasian people who likely have some steppe admixture. Instead it would be more sensible to compare them with South Caucasians, aka Kartvelians. And modern Georgians have just 9% of this component, much less than the Armenians, who are in an even more southern position than the Georgians.

It's not undisputed when the Armenian language reached Anatolia, but let's suppose you're correct and it was no earlier than 1200 BC. The Bronze Age Armenian I tested at GEDmatch is dated to 1211 cal BC according to Krefter's datasheet, and the Iron Age Armenian to 895 cal BC. So your claim that they predate the arrival of Armenian by over 1000 years is wrong.

Regarding your objections to what I called a Nordic-like cranial type, I meant this purely metrical, of course we haven't seen any data on their pigmentation. And anyway, as you correctly said, the Yamnaya population was dark pigmented, while some EEF were light-pigmented, so steppe-related admixture doesn't entail light pigmentation, and light pigmentation doesn't equal North European-like DNA.

Maikop was only north of the great Caucasus chain, in the Northwest Caucasus. And how much EHG admixed they were remains to be seen. But it's definitely something different than south of the Caucasus in West Asia. The EEF in Europe had no trace of EHG ancestry, and generally EHG and WHG admixture in West Asia is low, ANE is mostly found together with the teal component which is ANE mixed with Near Easterners without WHG.

I'm not so sure that R1b diversity is high among Lezgins. To the contrary, Hovhannisyan et al. found a low diversity there.

Grey said...

Mike Thomas

"Present actual evidence / data, or do us a big favour and shut up."

And your actual evidence for

"pastoralism can only have arisen where agriculture had already existed"

is where?

Grey said...

Kurti

"R1b in my opinion has been at least in the region between the Zagros/Albroz and Taurus region for a LONG time, if not maybe evolved there."

If some R1b from Yamnaya was involved in the ethnogenesis of Armenians in one corner of the Transcaucasus it doesn't mean all R1b in the region came via that route. It might be a footnote among the big migrations - still interesting though.

.

Nirjhar

"Grey,What do you think about this:)"

I thought it was very long :)


pnuadha said...

@Simon

Indeed, Corded Ware itself didn't expand to northwestern Europe. But neither did Yamnaya! It was post-Corded Ware central Europeans who expanded to northwestern Europe.

That can not be true or else central europeans would have more yamnaya dna than northwest europeans. You realize that central europeans have more actual CW heritage, right? Also, germans and polish probably have less yamnaya like dna than belarus which has less yamnaya dna than norway and scotland.

What probably happened is that the CW, although very high in yamnaya like dna, were not able to overwhelm the population of central europe like the bell beakers were able to do in northwest europe. We may also find that some of the bell beakers had much higher levels of yamnaya related dna than 50%.

The point is simple really; central europeans have more CW heritage than northwest europeans but northwest europeans have more yamnaya like dna which means they must have had a different source of yamnaya like dna.

As an aside, i astrally suspect that part of the reason Norwegians and Scots score so high in yamnaya estimates is because of small remnant of SHG in those areas. But i haven't spent much time exploring this idea because its minuscule to the big picture.

Krefter said...

RISE00: M913021: Corded Ware Estonia.

ANE K7.
ANE: 19.34
WHG: 74.48
ENF: 5.61
East African: 0.57

M966366: Corded Ware Germany. Haak 2015.
ANE K7.
ANE: 24.68
WHG: 59.61
ENF: 12.55

There was variation in Corded Ware.

pnuadha said...

Of course there was a massive migration that caused a dramatic autosomal change, but this doesn't preclude that the yDNAs go back to just a couple of successful patriarchs.

It sure could, but that has nothing to do with who the founding population for the bell beakers were.

The founder effects with regards to r1b did not happen within the time yamnaya left the steppe and Bell Beaker established itself in central europe with loads of r1b. The population was too large for the founder effect to explain the dominance of r1b.

Maybe later, the r1b diversity in western europe shrank considerably but it wasn't in the small time frame of 2800bc to 2550bc.

pnuadha said...

@krefter

There was variation in Corded Ware.

ya, i noticed that too. How much has the level of ANE gone up in estonia after that, or apart from that, sample?

Kristiina said...

Here you go http://postimg.org/image/8idyv6tn9/full/
I hope the map is reliable. It is found here: http://www.eupedia.com/forum/archive/index.php/t-30765.html

ANE has slightly decreased but WHG has decreased dramatically and ENF gone up almost as much.

RISE00
ANE: 19.34
WHG: 74.48
ENF: 5.61

Modern Estonians:
ANE: 18.5
WHG: 50
ENF: 27

Modern Lithuanians:
ANE: 18.5
WHG: 52
ENF: 28

Krefter said...

You guys should save this in a file somewhere.

https://docs.google.com/spreadsheets/d/1684wMM_ZJFoxcxJXK0jkVgeGGEVl5Nw3-Moc_IFrlOs/edit#gid=1051326962

Corded Ware Estonia fits in the East Baltic. Her score isn't exactly like Balts(and Estonians) but it's very close. Corded Ware Germany may have mixed with Gok2-types while Corded Ware Estonia mixed with MNs who were mostly Hunter gatherer in terms of blood.

Kristiina said...

I see that ANE K7 is giving very different values from those I have usually seen, so Krefter, why that ANE_7 is giving so low ENF values to north Europeans? When I personally did Eurogenes ANE / WHG / ENF test, my ENF was c. 33%.

Matt said...

@ Krefter, the CW probably do vary, though the Baltic Bronze Age samples are very low coverage, IIRC, so I don't know if that will hold up exactly as it is. It's in the right direction though that I'd expect for the Baltic, with more WHG/SHG admixture than German CW, so I think that's right (and there is some evidence that CW had problems transplanting their southern way of life to the Baltic, I think).

Maybe that links to the discussion between Colin and Simon, re the Yamnaya estimates in the Haak paper, and what that means for Corded Ware Central Europeans as candidate ancestors for British Isles plus Scandanavia.

What we see in the Haak paper, is that where Central and East Europeans have less estimated Yamnaya they have more WHG (although some of this may in fact be East Eurasian for populations like Russians and Estonian).

So where the populations in perhaps Poland may have slightly less Yamnaya ancestry than Norway or Scotland, that seems like it may be due to refluxes of new and successive cultural trends with more EHG ancestry from Russia or the Baltic, replacing some "Yamnaya" with EHG, after the early dates of the Corded Ware (as well as a MN European resurgence). Just as at the same time in another direction the CW culture may have also been spreading MN European ancestry to Sintashta.

All of this would be diluting the similarity to Samara Yamnaya, particularly because adding more MN European and EHG into the mix is decreasing the West Asian / Gedrosia element, which is the most signature to the Yamnaya when EHG is removed (and Scots have that odd elevated Gedrosia signature).

Unknown said...

RISE00: M913021: Corded Ware Estonia

Looks odd on some of the more Modern tests. Very NW Europe, NE Europe. Just very North!

Davidski said...

These samples have to be pooled into high quality population groups to analyze them effectively, which means it's better to use ADMIXTOOLS to test them. qpAdm will be very useful here.

I should have most things up and running later today.

At some point I'll also make composites for each group or sampling location, so that at least they're not missing markers, which should give us a better idea of what these populations look like in ADMIXTURE.

a said...

@Kristiina said...

"I see that ANE K7 is giving very different values from those I have usually seen, so Krefter, why that ANE_7 is giving so low ENF values to north Europeans? When I personally did Eurogenes ANE / WHG / ENF test, my ENF was c. 33%."

In K15
Sintashta's looking a little like Spanish Aragon.

Unknown said...

I cant see an R1b Corded Ware sample on Genetiker's webpage? Has this been corrected/amended? Which one is supposed to be R1b?

Unknown said...

Matt

"So where the populations in perhaps Poland may have slightly less Yamnaya ancestry than Norway or Scotland, that seems like it may be due to refluxes of new and successive cultural trends with more EHG ancestry from Russia or the Baltic, replacing some "Yamnaya" with EHG, after the early dates of the Corded Ware (as well as a MN European resurgence). "

That fits with what is expected given the suspected demography of places like Poland and Belarus between the Bronze Age and pre-modern Era....

Unknown said...

OK found them on Jean Manco's page.

Chad said...

Unsupervised K10

https://docs.google.com/spreadsheets/d/1g-oFfeMGpTFgNH8K_iu9ZLZIjcCig97nkgBLQeIWJlQ/edit?usp=sharing

Chad said...

Sintashta and Andronovo are very hyper-Baltic. Anyone claiming roots from the South and then migrating to Europe are living in denial.

Chad said...

Andronovo looks like a Norwegian, plus 10-15% extra Siberian.

Krefter said...

R1b-U106, R1a1a1, and I1a3a1 in Sweden/Denmark 4,000 years ago. Note outside this I1 has only been found in Neolithic Hungary and Bell Beaker R1b is 2/2 P312 so far.

https://docs.google.com/spreadsheets/d/12G2cfjG0wHWarsl5bB99ridFmvUWzqlZfZ6_e_R6oIA/edit#gid=479090567

Geneticker's page.

https://genetiker.wordpress.com/2015/06/11/y-haplogroups-for-prehistoric-eurasian-genomes/

Chad said...

The R1b MBA Armenian stands out vs modern Caucasus pops as far more steppe shifted. E sample looks like modern N Caucasus. IR samples are more SE European like.

Davidski said...

Yeah, these Bronze Age Armenians look like they had recent ancestors from the steppe and/or northeast Caucasus. This is a pretty good fit for them using qpAdm (chisq 2.140, tail prob 0.543766).

baArm
Yamnaya 0.312
LBK_EN 0.688

But then again this is an awesome fit for Yamnaya (chisq 0.573, tail prob 0.902514).

Yamnaya
EHG 0.427
baArm 0.573

So working out the precise details of the nature of gene flow between Yamnaya and the Bronze Age Caucasus is gonna be tricky.

Btw, Matt, depending on the markers used, Sintashta can be modeled as around 50/50 Esperstedt_MN/Yamnaya, or around 20/80 Esperstedt_MN/Corded_Ware_LN, but the fits aren't great. There's something missing for them using these models.

I'll look at this issue in more detail soon.

Chad said...

Yeah, the BA Armenian is the most "Teal" of anyone we have...

For Sintashta, Afanasievo, and Andronovo, try Karitiana or Nganasan. They look like they could be pretty diverse, with some just a couple percent ENA, and others look 15%.

Unknown said...

David

"Sintashta can be modeled as around 50/50 Esperstedt_MN/Yamnaya, or around 20/80 Esperstedt_MN/Corded_Ware_LN, but the fits aren't great. There's something missing for them using these models. "

Could it be extra "EHG" is required ?

Krefter said...

Sintashta/Andronovo are a curve ball. A few years ago it seemed for sure Yamnaya would come out R1a. We had Corded Ware R1a and Andronovo R1a, and why wouldn't Yamnaya in the middle have it? Who would have thought Sintashta/Andronovo and their R1a were just a back migration from Corded Ware-types. It's was totally unpredictable.

Krefter said...

None Americans: Curve ball as in tricky or unpredictable.

Davidski said...

Mike,

I tried both WHG and EHG. They don't help. I can try SHG later, but I doubt it'll help.

Krefter,

Check out page 5 of the supp info...

1.5 Abashevo/Sintashta (2100-1800 BC), Andronovo (1700- 1500 BC), Karasuk (1400-900
BC), and Mezhovskaya Culture (1300-800/700 BC).

Unknown said...

Krefter

We need pre-Bronze Age samples beyond the ones at Samara and Karelia to clarify. It's interesting because R1a and R1b probably split c. 18 kya. Have they been living side by side by in almost mutually exclusive , bounded territories all that time ?

And even non-Americans understand "curve ball". :)

Davidski said...

Here are PCA plots of a few of the ancient Armenians.

https://drive.google.com/file/d/0B9o3EYTdM8lQSVlHRFJPY0laR1E/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQSU1LMlMwZTljeVU/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQSGxkSzBVLUI1WXM/view?usp=sharing

https://drive.google.com/file/d/0B9o3EYTdM8lQTmZhclotaXVaZWs/view?usp=sharing

Unknown said...

On thinking about this this does not look like modern Europe to me. I feel like there are huge gaping holes in this data.

(1) Masses of R1a when this haplogroup is NOT well represented today in most areas. It looks like it is associated with the Corded-ware folk.
(2) Bell Beaker is R1b P312 (R1a2a1a2) which is the southern component, well represented in Spain and Ireland in modern populations. This kind of fits with the proposed origin of the Bell Beakers in Iberia. I can buy that.
(3) The only clear R1b U106 (R1a2a1a1) is Swedish Nordic LN (RISE98). This is just plain weird given current distributions. My best guess is that U106 was sitting in the British Isles (no samples) and Sweden. It must have permeated out to Europe later. The distribution of this clade (expanding out of Doggerland) really does not fit with a very,very late arrival in Europe from the steppes.

In summary I am going to say that despite the absence of Iron age R1b or earlier samples it looks to me that R1b P312 expanded with the Bell Beakers from Iberia in the Bronze Age (the implication being that it was already in Iberia at the start of the Bronze age). And R1b U106 sat out the bronze age in Britain/Scandinavia (the implication being that it was already in place at the start of the Bronze age).

I declare the arrival of R1b in Europe .... still unclear.

(But probably pre-Bronze age from this data.)

Davidski said...

Keep in mind that one of the most gaping holes in the data is Yamnaya west of the Don.

This is where I predict we'll find the ancestors of the early Corded Ware and Bell Beakers, and thus of most present-day Europeans.

Krefter said...

@Annie Mouse,

There's no way in hell R1b-L11 is not from the east(East Europe, West Asia).

Neolithic Y DNA has come out with 0 R1b1a2. Look at the Y DNA from EEF-type North Italians from 5,000-3,700YBP. All typical Neolithic I2a1a and G2a(Otzi). Same with large majority EEF Bronze age Hungarians. Y DNA that will be coming out soon(maybe next year, I don't know) from El Portalon Spain from around 4,000YBP will also come out typical I2a, G2a, etc. and not R1b-L11.

Back in the day, before we had dozens of their genomes, people had room to speculate Neolithic West Europe was genetically diverse with a diverse array of Y DNA, and therefore there was room for R1b-L11. But now we know this wasn't the case. They weren't diverse, they were all the same-thing(EEF+WHG) and had the same Y DNA. Their Y DNA didn't include R1b-L11.

Considering the close relation between all Neolithic West Europeans. What are the chances there was a lone R1b1a2a-L23 lineage that hasn't been found yet in ancient Y DNA and somehow expanded and became dominate right before the Bronze age?

R1b-L11 first pops up in Central European Beaker, who were largely of Yamnya-type origin. This is no coincidence. It is also no coincidence Yamnaya belonged mostly to R1b1a2a2.

Allentoft claims there's a R1b from Corded Ware. And if true, this R1b predates Bell beaker in that region. That right there should prove that R1b in West Europe expanded from the East.

"Bell Beaker is R1b P312 (R1a2a1a2) which is the southern component, well represented in Spain and Ireland in modern populations. This kind of fits with the proposed origin of the Bell Beakers in Iberia. I can buy that."

P312 is most diverse in Central Europe which is close to where it originated. All there is in Iberia and Ireland is pretty much DF27 and L21. Unless East European P312 went down to Iberian Bell Beakers, Y DNA from Iberian Bell Beaker will be I2a, G2a, etc.

Unknown said...

"@Annie Mouse,

There's no way in hell R1b-L11 is not from the east(East Europe, West Asia)."

I am not going to challenge you, I dont have a strong view on this. I am just leave you with a distribution map as food for thought. Note L11 and L11 with major clades removed.

http://www.theapricity.com/forum/showthread.php?1814-SNP-s-and-Haplogroups/page2

"El Portalon Spain from around 4,000YBP will also come out typical I2a, G2a, etc. and not R1b-L11. "

Is this inside information or a guess?

"They weren't diverse, they were all the same-thing(EEF+WHG) and had the same Y DNA. Their Y DNA didn't include R1b-L11. "

Your opinion. My current view is much more complicated.


"R1b-L11 first pops up in Central European Beaker,

Yep specifically R1b P312 (R1b1a2a1a2). As I said.


"who were largely of Yamnya-type origin. "

A lot of leeway for history in this statement.

"It is also no coincidence Yamnaya belonged mostly to R1b1a2a2."
Which is not L11 (R1b1a2a1a) or even an ancestor, it is a great great uncle. So your point is?

"Allentoft claims there's a R1b from Corded Ware. And if true, this R1b predates Bell beaker in that region. That right there should prove that R1b in West Europe expanded from the East."
R1b in Corded ware does not prove R1b came from the east, far from it. R1b1 is found in Spain aound 5100 BCE in association with the LBK. 2000 years before Yamnaya and Corded ware. The evidence for R1b's Eastern origins lie in its parents and closest kin.

"P312 is most diverse in Central Europe which is close to where it originated. "

STR data as I recall? If my memory served me this study was trashed on publication as meaningless. STR data is only suitable for genealogy and loses almost all value in long term studies as the STRs vary back and forth too much.

"All there is in Iberia and Ireland is pretty much DF27 and L21. "

See the link above for P312 distribution (first plot). The bulk of P312 is in Iberia and Ireland This one below by Myers of P312 variance is the best I could come up with and it assumes only one expansion rather than expansions in different directions. Never-the-less it is centred on France not central Europe.

https://rokus01.files.wordpress.com/2010/09/p312.png

I dont know if R1b P312 originated in Europe or arrived from elsewhere. All I am saying is that from this data it expanded out of Iberia with the Bell Beakers. As the following article suggests.

"One of the earliest Beaker styles is the “Maritime tradition” that probably originated from the Early Copper Age Copos in the area of the mouth of the River Tajo in Portugal." from Turek, 2012

"Unless East European P312 went down to Iberian Bell Beakers, Y DNA from Iberian Bell Beaker will be I2a, G2a, etc."

Well since all Bell Beakers so far are P312 whether they went north or south they are most likely to be P312 on current data. Plus there are the 2 R1bs that were in Iberia 2000 years previously.


June 14, 2015 at 8:42 PM

Unknown said...

Correction:

"R1b-L11 first pops up in Central European Beaker,

Yep specifically R1b P312 (R1b1a2a1a2). As I said.
---------------------------------------------------
The sample is from Germany (Kromsdorf, 2600–2500 BC) which is really Western European Beaker not Central European Beaker. I assume we are talking about the same one?

capra internetensis said...

@Krefter

Bell Beaker Blogger doesn't think R1b-M269 is Paleolithic. He thinks it is Chalcolithic and arrived from the east. He just has some interesting views about the route it took.

@postneo

There are like 2 Sintashta Y DNA samples. How on earth would we expect to tell from that sample size whether there was R2, or whatever else there may have been? And of course there has been wave after wave of new people rolling through the area, so why should there be much continuity?

Grey said...

Krefter

"There's no way in hell R1b-L11 is not from the east(East Europe, West Asia)."

I think R1b came into Europe from the steppe* as part of a large migration *but* the distribution of R1b in western Europe is very odd and I think that oddness is where some of the opposition comes from.

(*not arguing how it got to the steppe in the first place)

DF27 looks like it expanded out of the Basque area

http://cache.eupedia.com/images/content/Haplogroup-R1b-DF27.gif

L21 looks like it expanded out of Ireland

http://cache.eupedia.com/images/content/Haplogroup-R1b-L21.gif

S21 looks like it expanded out of - hard to say - Frisia/Netherlands maybe.

https://thecampblogbymike.files.wordpress.com/2013/01/haplogroup-r1b-s21.gif

It might not mean anything but these are odd distribution patterns for a mass migration from the east - especially the Irish/Basque ones as most of the known population movements in later history were in the opposite direction.

So if it was a mass migration from the east then it seems to me something odd might have happened specifically along the Atlantic coast that led to those three clades expanding so dramatically in what seems like the wrong direction.

Also the Atlantic coast is a separate bio-region and I think it's the generally likely that something odd may happen when a population crosses a border between a bio-region they are adapted to and one they aren't.

http://40.media.tumblr.com/fe9613ef7cda5883c54bdb644cbeb6fe/tumblr_nigb28ir2A1rasnq9o1_1280.png

Kristiina said...

Grey, why a Y-line should spread only to one direction, i.e. in this case from east to west. R1b may well have had a founder effect in several places where, then, a particular subtype developed and started to spread, DF27 in the Basque area, L21 from Ireland/British Isles, S21 from Netherlands etc. If R1b P312 was in Kromsdorf c. 2550 BC, it had plenty of time to diversify and move around. I do not think that a mass migration is necessary, although I do not exclude it in some places. IMO, it can also be a constant trickle that goes on for centuries.

In general, my impression is that these big changes in yDNA during the Bronze Age and Iron Age are mainly due to metallurgy (+wheel, horses, long distance trade). Men who knew how to work metal or carried metal tools/weapons were successful in having offspring.

Nirjhar007 said...

Kristiina,
//In general, my impression is that these big changes in yDNA during the Bronze Age and Iron Age are mainly due to metallurgy (+wheel, horses, long distance trade). Men who knew how to work metal or carried metal tools/weapons were successful in having offspring.//
Yes You are quite right:).....

Matt said...

Interesting PCA for the BA Armenians - those agree a lot with the ADMIXTURE from the paper and the compressed PCA in the paper in suggesting a more North Caucasus like population than Armenians. Strange that the f3 stats seem to me to show a reduced affinity to other populations, particularly EEF in BA Armenians and a relatively increased one in IA Armenian, and that's not so evident there. Likewise for their qpAdm. Perhaps this is an effect of grouping in some of the lower quality samples in the paper, and at some point in the future running the same stats with a high quality subset would get a result that didn't have that combination of low affinity with a dual North European + Caucasus affinity.

Match for Yamnaya as EHG+ BA Armenian is cool, even though BA Armenian is too late to tell us if this is close how this happened.

Thanks for the other qpAdm results as well - Sintashta looks on the paper's ADMIXTURE like it has some distinction from other populations via the presence of a purple component that isn't like anything in West Eurasia. Even as a qpAdm fit those proportions are much less Yamnaya like than I'd expect. The more MN goes into Sintashta, the more puzzling it is for them to have contributed to other populations (since 50:50 MN:Yamnaya is I think pretty close to present day North-Central-East Europe). If WHG and EHG don't improve the fit, I guess it goes without saying MA-1 wouldn't either.

Davidski said...

Nope, for Sintashta adding MA1 doesn't help. I also tried adding Motala_HG and it didn't work.

But I think these are interesting...

Pathan
baSin 0.709
Georgian 0.165
Dai 0.126

chisq 0.059 tail prob 0.970891

Kalash
baSin 0.700
baArm 0.176
Dai 0.125

chisq 0.315 tail prob 0.854188

Those are some of the lowest chisq/highest tail prob combinations I've ever seen in qpAdm, which I suspect means something.

Unknown said...

Grey

A) Those lineages are all descended from S-112. Is n;t it obvious they represent individual founder effects from one common ancester, somwehere in the north-west Alpine region- like Swabia (lets say) ?

B) The proof that you require for the emergence of pastoralism may be found here

https://books.google.com.au/books?id=PPumUjpp--UC&dq=pastoralists+outside+world,+khazanov&hl=en&sa=X&redir_esc=y

https://www.academia.edu/1493571/Multiregional_Emergence_of_Mobile_Pastoralism_and_Non-Uniform_Complexity_across_Eurasia

http://aragats.arts.cornell.edu/?wpfb_dl=29

All the evidence lies on those few hundred pages of reading.

But I'd be very interested in reading your fresh new perspectives on pastoralism (which will undoubtedly overturn the academic community). Im guessing your years of field research will be getting published in this year's Annual Review of Anthropology ..??

Unknown said...

David

Wow They are interesting indeed.

Alberto said...

@Chad

Thanks for that Admixture run. Very interesting.

Regarding the BA Armenian, the oldest one (ca. 1800 BC):

baArm (RISE413)
Amerindian: 2.5%
EEF: 14.2%
Euro_HG: 6.5%
BA_Cauc: 74.5%

Karelia_HG
Amerindian: 12%
EEF: 0%
Euro_HG: 88%
BA_CAUC: 0%

baYam (RISE552)
Amerindian: 2.8%
EEF: 0%
Euro:HG: 49%
BA_Cauc: 46%

The BA Armenian, apart from getting some EEF admixture during the 1500-2000 years they must have been on the area, it doesn't look to me it has any steppe admixture.

On the other hand, it looks like a great fit for Yamnaya with the EHGs. The fact that he's the oldest one we have (though not old enough), and that he's R1b, shouldn't it be giving a strong clue as to what really happened? Sure, we'll need an older sample for 100% confirmation, but the odds are that we found the direct descendants of the "Armenian-like" population from Haak et al.

Unknown said...

Alberto/ Dave/ Matt

There is strong evidence for continuity from Copper Age/ Early Bronze Age Kura-Arax to mid to late bronz Age (although the KA 'fragments". On the one hand, there is clear evidence that the stimulus from Majkop came from the south, and perhaps the earliest Kurgans were in south Caucasia (but requires more Carbon dates, etc). On the other hand, there is suggestions for Majkop influences to south. Yet at the same time, the two regions (nth and Sth) of Caucasus who clear material differences.

Its really hard to tease this out/ make hypotheses based on archaeology. Im disappointed they didnt get any Late Neolithic samples from Armenia or Georgia, Urkain or Sth-West Russia....

Alberto said...

@David

Very interesting indeed those results about Pathans. But also extremely surprising, woudln't you say? I mean, mixing Sintashta (which are like Norwegians) with some Georgian, how can it give a Pathan? There must be a reason for it, but it will need a lot of investigation to know why.

The results using Sintashta are better than using Yamnaya, right?

Balaji said...

Skilur,

You wrote, “If R1a-Z93 is confirmed in Sintashta and Andronovo the OIT proponents have to give up but i doubt they will accept their defeat.”

Why should they give up? They will simply say that the R1a-Z93 in Sintashta and Andronovo is from India. In order to truly defeat OIT, someone must demonstrate how the WHG in Sintashta and Andronovo was “washed out” in the migration from the Steppe to India. This is no easy task and Allentoft have made it harder by showing that WHG in Sintashta is even more than in Yamnaya. Sintashta have more WHG than ANE. But in India there is almost no WHG and there is ANE. The ratio WHG/ANE is greater than 1 for Sintashta and 0 for India.

Grey said...

Kristiina

" I do not think that a mass migration is necessary, although I do not exclude it in some places. IMO, it can also be a constant trickle that goes on for centuries."

Yes in the context of Europe as a whole and over the whole time period I think there was a mass migration but as you say not necessarily in one big wave leaving room for some oddness happening in some places.

Alberto said...

@Mike

Yes, it's a bit disappointing that the samples were not older, but I'm sure they're coming soon enough. And I guess that the most likely is that they will confirm what already looks very likely.

What is indeed surprising from this paper is the Sintashta phenomenon. I don't think anyone expected them to be North Europeans. And David's qpAdm results are even more surprising. Also their Y-DNA being very South Asian is surprising. I guess we'll need samples from BMAC to understand how to make sense of all this.

Davidski said...

The same models for the Pathans and Kalash also work well with Yamnaya, especially for the Kalash (not surprisingly, the more teal population in my K9 test). But as far as I can tell, Sintashta still looks better for both.

Kalash
Yamnaya 0.263
baArm 0.627
Dai 0.110

Chisq 0.354 Tail Prob 0.837957

In any case, it's obvious that we need a Bronze Age steppe population to successfully model the Kalash and Pathans. And for Pathans we also need a little bit of Early European Farmer admixture (orange cluster in my K9 run), which I think is why Sintashta works so well for them.

Andronovo might produce better results for Pathans if I remove one of the obvious outliers with high Siberian admix. I might try that at some stage.

Unknown said...

Alberto

Yes. Im sure more data is coming.

"And I guess that the most likely is that they will confirm what already looks very likely."

Care to elaborate ? :)


Davidski said...

Page 5 of the Allentoft et al. supp info tells us how it went down.

http://www.nature.com/nature/journal/v522/n7555/full/nature14507.html#supplementary-information

Note how they talk about Corded Ware-derived Abashevo from Eastern Europe being a close relative of Sintashta.

They might as well be talking about R1a-Z282 and R1a-Z93.

Alberto said...

@David

Yes, I can understand a rather small group could migrate from Europe to the Sintashta area and be the founders of such culture. And I can understand even that they migrated south and become a ruling elite, and even change the language. But when we have a big enough society (like the IVC was), a ruling elite represents 0.1% of the population (that's why they are an elite in the first place). The masses descend from the masses. That's how societies with a pyramidal structure works.

To have a big genetic impact, a ruling elite doesn't work. We need a mass migration. And that's where a mass migration from Europe to Sintashta and then to South Asia becomes difficult to explain. Even more looking at the genetic profile of Sintashta and modern Pathans.

Anyway, the results are really intriguing and interesting. But the mystery will require more samples from Asia to understand the process.

Davidski said...

Well I really can't see any other option now than a mass migration from the eastern CWC/Sintashta/Andronovo regions to South Central Asia.

Anyway, to make sure that the relationship between the Dai and Han (as one of of the right pops) wasn't messing things up, I replaced Han with Ulchi, and got slightly better fits.

Pathan
baSin 0.634
Georgian 0.231
Dai 0.134

chisq 0.049 tail prob 0.975924

Kalash
baSin 0.578
baArm 0.284
Dai 0.137

chisq 0.240 tail prob 0.88707

Alberto said...

@Mike

"Care to elaborate ? :) "

I mean that what made more sense back when Haak et al. was published, has become even more likely now. The "Armenian-like" population that entered the steppe came from the Caucasus and they where the ones who brought R1b with them (and likely the language and culture). The other option of HGs taking Caucasus wives was never very solid. And even Anthony's suggestion of a Mesolithic population from the North Caucasus was a wild guess at best.

Now we'll just need older samples to confirm it. And then samples from Central Asia to confirm that they originally came from there (not that I have a crystal ball, but there are just not many other options). Though alternatively they could have been in the Caucasus all along and share the same genetic profile with Central Asians. But that's less likely, I think.

Unknown said...

Alberto


So where did the (admittedly later) Bronze Age Armenian R1b samples sit in the overall phylogeny compared to moderns and yamnaya and BB samples ?

Kurti said...

@David

bA Armenians did not get any type of Yamna ancestry. I gave at least 3 reasons for why. MAYBE there was some kind of genetic input from other regions of the world but surely not Yamna.


Yamna did not have Atlantic_Med type ancestry and had only 5% of Caucasus. This is the major difference to modern West Asians and what causes the "European" shift not freakn Yamna type ancestry.


The major "European" shift in this samples is the much smaller frequency of Southwest Asian and the significantly higher Atlantic_Med ancestry (over 30% vs modern populations with ~10%) and Caucasus (30%).


You are like trying to explain the blueness of the sky with the greenhouse effect.

Armenia was populated by Northeast Caucasian type people and every freakn linguist believed that Northeast Caucasians arrived from further South and not the opposite.

MAYBE there was minimal Yanma type input already, but certanly Yamna type ancestry is not the explanation for EEF type genes in BaArmenia (It is rather the remnant of the ancient farmers) and I doubt it is the explanation for "North Euro" type ancestry either because it doesn't exceed what modern Northeast Caucasians have and is very well in accordence that Northeast Caucasians are settlers from further South.

Davidski said...

I never said Yamnaya admixture explains the EEF in Armenians. Of course it doesn't.

But Northeast Caucasians don't derive all of their ancestry from the Near East. A lot of it had to have come from the steppe, Siberia and/or Central Asia at some point.

Kurti said...

@Alberto

"The BA Armenian, apart from getting some EEF admixture during the 1500-2000 years they must have been on the area, it doesn't look to me it has any steppe admixture."


Exactly, The EEF couldn't come from Yamna which lacked it. Ba Caucasus (ENF) couldn't come from Yamna because they had 1/3 less of it.

EHG is in accordance to North Caucasic speakers who are known to have come from South.

Absolutely nothing points to an Yamna influence.

Alberto said...

@Mike

You mean regarding the R1b tree? I'm absolutely no expert in those phylogenies and easily get lost. But isn't Armenian R1b mostly L23? Which is ancestral to Yamnaya R1b and to L51?

And isn't the parent (or grandparent or whatever) of R1b-L23 mostly in Turkmenistan?

Surely someone else can give a much better and detailed explanation about it.

Kurti said...

@David


"But Northeast Caucasians don't derive all of their ancestry from the Near East. A lot of it had to have come from the steppe, Siberia and/or Central Asia at some point."

Yep maybe 20000 years ago, who says that this ancestry didn't reach already the regions further South too? For example through the Iranian Plateau. Fact is linguists agree that Northeast Caucasic languages have a relation to Hurro_Urataen. Fact is also that Ba Armenian "NOrth European" type ancestry is perfectly in line with what is found among North Caucasians.

Mark my words. Bronze Age North Caucasian samples will turn out more EHG than South Caucasian once. From historic perspective it makes sense that there was a Southern migration towards North. Ancient populations from South(Anatolia/Iran, Mesopotamia) replacing ancient populations from North Caucasus.

Davidski said...

Kurti,

1) No one claimed that Armenians got their EEF-like ancestry from Yamnaya.

2) Northeast Caucasians obviously don't derive all of their ancestry from the Near East.

Alberto,

Ancient Samara R1b is, at the very least phylogenetically, ancestral to both Armenian and Yamnaya R1b. Modern Armenian R1b is not ancestral to Yamnaya R1b.

R1a and R1b are clearly EHG markers. EHG lacked Near Eastern ancestry. So your proposal that Yamnaya R1b derives from Armenian R1b must involve two major migrations; first one south from the steppe to the Transcaucasus, then a second one from the Transcaucasus to the steppe soon after.

This is possible, but unlikely. It's more likely that if Armenian R1b isn't from Yamnaya, then at some level it represents shared paternal EHG ancestry between Armenians and Yamnaya.

In other words, the Bronze Age Armenians need not be part Yamnaya, but can be, say, Hurrians, who might have been part EHG. Actually, didn't Hurrians live in the Transcaucasus at that time?

This of course doesn't solve the issue of the Caucasian admixture in Yamnaya, but female mediated migration does, and is backed up by what we know about the Bronze Age from archeology.

Kurti said...

David

You said there is Steppe type ancestry in South by comparing it to Yamna.

That in some point of time there was North Eurasian typ influx into South was not even on debate.

The point was that you were making it look like the BaArmenian samples received relatively recent (Indo European) Yamna type ancestry. And I explained why this is not possible.

That was the point. I personally believe that ANE type ancestry also came via the Iranian plateau, that explain better the ASI affinity in BaArmenia.

I also believe that some WHG type ancestry existed in Western Asia in combination with farmer ancestry this explains the Atlantic_Med type ancestry which is 1/4 WHG imo.


But whatever

Unknown said...

Independent of what moved where and which direction between south of the caucasus and north, I think its pretty clear that the WestMed-Atlantic component is likely to be a Mesolithic Anatolio-Caucasian marker. This regions was linked more to Europe than the Levant in the UP.

Alberto said...

@David

Yes, we have an old sample of Samara HG carrying R1b. It seems R1b was not confined to a small region, since it was also found in a Spanish farmer. And yes, EHG and the "Armenian-like" population do have a shared ancestry, sharing R1 and ANE. And they were probably neighbours around SE Kazakhstan.

Actually that's also a possibility as to why Afanasievo and Yamnaya are so similar, being a mix of 2 populations while being quite distant apart and contemporary. If this Central Asian population was from SE Kazakhstan to the Caucasus and EHG were north of them in the steppe, it could have been a parallel movement from Central Asians going north into the steppe and mixing with EHGs.

But although we need more samples, my feeling is that R1b was probably a minority in the steppe and a majority among Central Asians. And the BA Armenian R1b is much closer to Yamnaya R1b than that of Samara HG.

But yes, we'll need more and older samples to confirm it. You can still hang onto the Caucasus wives theory for a little while longer if you really wish :)

Nirjhar007 said...

Balaji,
//Why should they give up? They will simply say that the R1a-Z93 in Sintashta and Andronovo is from India. In order to truly defeat OIT, someone must demonstrate how the WHG in Sintashta and Andronovo was “washed out” in the migration from the Steppe to India. This is no easy task and Allentoft have made it harder by showing that WHG in Sintashta is even more than in Yamnaya. Sintashta have more WHG than ANE. But in India there is almost no WHG and there is ANE. The ratio WHG/ANE is greater than 1 for Sintashta and 0 for India.//
It will all depend on aDNA from SC Asia but yes the Autosomal structure is not helping AIT and neither the clades actually.....

postneo said...

@balaji if sintashta is from cw/yamnaya we need a trail of z93 ancestors there.
So Far older East yamnaya is r1b. Is cw ancestral to z93? If not it's not compelling.

Davidski said...

Corded Ware R1a is indeed ancestral to Z93, because it's M417*. And Eastern Corded Ware is archeologically ancestral to Sintashta.

Some of you guys still can't see the writing on the wall? Do you actually need writing on the wall literally to get this?

Unknown said...

This is an the elephant in the room for me.

The dominant mitochondrial haplogroup in Europe is H (~50%). It is chiefly H1 and H3.

H (total) = 18 (18% of the 101)

H(other)= 2%
H1 = 1%
H2 = 4%
H3 = 2%
H5 = 2%
H6 = 3%
H8 = 1%
H11 = 1%
H44 = 1%
H46 = 1%

The H1af (RISE 569) is a Czech Bell Beaker. This has been found in Denmark, Norway and the UK according to Phylotree. It happens to be Empress Alexandria Romanov's haplogroup.

The H3b (Rise 71) and H3v +16093 (Rise 42) are both Danish Bell Beakers.

But H1 plus H3 is about 1/10 of modern levels, and it is most probably flowing up from Iberia (probable source of the Bell Beakers) and it is completely absent from the bronze age steppes or Italy (possible Med route). So the questions are: is it that H1 +H3 have not fully arrived yet? Are they in an unsampled region? Or do they tend to dispose of corpses by methods other than interment? Very similar pattern to R1b P312 (R1b1a2a1a2).

H1 and H3 have been in Spain, France and Germany since the neolithic. So why arent we seeing more of it in the Bronze age?

The H2 is H2a and I cant find any pattern to it. The H8a is reasonably Armenian. The H6b is also Armenian. H6b seems to be middle eastern H6 (common in Saudi Arabia also).

I personally found it interesting that the other 2 H6 were H6a1b. So we now have H6a1b in Yamnaya (according to Brotherton), Okunevo (RISE516), Unetice Germany (according to Brandt) and Unetice Poland (RISE145). The modern Chuvash (in the heart of the steppes, IndoEuropean language) are about 30% H6 but I have never been able to find out which H6. A good candidate for R1a 's feminine partner. Or perhaps the Finno Ugric feminine substratum.

Krefter said...

I think your opinions about mtDNA H clades are too confident. Keep in mind most of the mtDNA from this study is from people of largely Steppe-decent.

"So the questions are: is it that H1 +H3 have not fully arrived yet? Are they in an unsampled region? "

I'll have to check again, but I'm pretty sure many Neolithic Hs are H1 and H3. Few have been tested for H1 and H3, but the ones that have often come out positive. H1 and H3 were very rare in the Steppe, we know this because of Wilde. 2014.

BTW, my opinion is that H2, H6, and H4 are mostly of Steppe-decent in Europe. I don't know though.

Nirjhar007 said...

David, First we see Samples of Similar age of CWC in Asia then we will found out, so far there is 0 Z-93 in Ancient Europe and a good chance of R1a-417 to be there in Some Significant ancient sites in Asia....

Nirjhar007 said...

// Eastern Corded Ware is archeologically ancestral to Sintashta.//
Sintashta is a clear mismatch to CWC Civilization wise.

Simon_W said...

Some observations from the ADMIXTURE analysis in the paper:

The South Asian-Central Asian-Caucasus component which appears at K = 8 is strong in BA Armenians and in Yamnaya and it's present in all populations influenced by these. It's absent in early in middle Neolithic European farmers. So it might be a marker of IE expansions. Notably this component is absent in BA Hungary and Montenegro. At K = 8 Remedello is like Gok2, and BA Hungary and BA Montenegro likewise, they only have a few individuals with increased HG ancestry.

However, at K = 15, the South Asian-Central Asian-Caucasus component splits into a South Asian and a Caucasus component. Now the Caucasus component is also present in BA Hungary and Montenegro. The Caucasus component was already present in the early European farmers, but it had disappeared in some middle Neolithic farmers, probably because of slight extra WHG admixture. At present, neither Sardinians nor Basques have any trace of the Caucasus component. Yet in BA Hungary and Montenegro it had increased again, to a higher level than in the early farmers. BA Montenegro has a lot of it, among the modern Greeks only one has more than them. So, whereas the combined K = 8 South Asian-Central Asian-Caucasus component looked like associated with R1a/R1b people, this K = 15 Caucasus component behaves more like associated with J2 expansions.

But then, at K = 19, the Kalash component appears, and now those populations who had the combined K = 8 South Asian-Central Asian-Caucasus component get some Kalash admixture. While those who only had the K = 15 Caucasus component, don't get Kalash, they keep their Caucasus component. Only three indiviuals among BA Hungary have some Kalash, presumably from the Yamnaya there. BA Montenegro doesn't have it. Modern Greeks have only minute traces. And so do modern Armenians. But interestingly, BA Armenians had a stronger Kalash signal! It has already disappeared in IA Armenians. Yamnaya has even more than the BA Armenians.

Krefter said...

The I2a from Yamnaya is I2a2a1b1b2-S12195.

https://genetiker.wordpress.com/2015/06/11/y-haplogroups-for-prehistoric-eurasian-genomes/

Simon_W said...

By the way, RISE98 who is classified as baSca, is actually Swedish Battle Axe / Nordic LN, he's dated to no later than 2032 BC. That's long before the Bronze Age started in Scandinavia, and as we know Sweden wasn't part of the Bell Beaker realm either. So he's largely from a Corded/Battle Axe derived background. And he's R1b!

Nirjhar007 said...

Simon, Kalash Comp. = Archaic PIE component?

Krefter said...

"That's long before the Bronze Age started in Scandinavia, and as we know Sweden wasn't part of the Bell Beaker realm either. So he's largely from a Corded/Battle Axe derived background. And he's R1b!"

Yes and the Eastern branch of L11: U106. So, Beaker can't be the source of all R1b-L11.

Simon_W said...

Quite awesome. And the Corded Ware R1b was from central Poland and no later than 2578 BC, i.e. not in a particularly western position and certainly predating Bell Beaker arrival.

@ Nirjhar
It depends on where PIE was located, I'm not sure right now. There is some chance that it was associated with PPIE, if we want to speculate.

Krefter said...

Simon,

Sorry, I should have pointed out the Corded Ware R1b-U106 is from Sweden. The reported Corded Ware R1b by Allentoft is from Germany.

capra internetensis said...

@Mike and Grey

Here is what Frachetti says in Pastoralist Landscapes and Social Interaction in Bronze Age Eurasia, which I have been reading:

"In the western steppe we may confidently look to agricultural societies for the earliest regional evidence of animal domestication, but the extent to which pastoralists were at all marginalized or directly derivative from agricultural socioeconomic systems is unclear. In fact, recent work in the central steppe at Botai seems to suggest that pastoralism emerged there as an independent innovation that had little to do with agricultural societies in its early stages.... ties between agricultural economies and early forms of specialized mobile pastoralism are without evidence in the central and eastern Eurasian steppes.... we may conclude that the 'agriculture to pastoralism' developmental model is dubious for the eastern Eurasian steppes, because the current archaeology seems to support a direct transition from mobile hunting/fishing economies to mobile pastoralism."

Whether Botai people actually domesticated the horse, or acquired it from livestock herders to the west, is still unknown. It would be nice to have some Botai aDNA - there are human remains, and the preservation conditions should be decent.

Matt said...

@ Davidski: Re: the qpAdm Sintashta plus BA Armenia models vs the Yamnaya plus BA Armenia.

Wow. Yes. It looks like Sintashta is actuallly really pretty good (better than Yamnaya) at acting as a proxy for Pathan and Kalash in their outgroup f4 relationships, with only pretty small input from Caucasus and Southeast Asian.

So Sintashta like populations as mass contributors to the ancestry of at least people from the Pakistan-Hindu Kush region got a lot more compelling.

I suspect the reason why Sintashta takes up a lot more space compared to Yamnaya when the other populations are the same is that as Sintashta is a population that behaves more Basal Eurasian / Neolithic population than Yamnaya, it's a better fit, while the ANE-ness of Sintashta probably makes it better than B A Armenian, which has less. Pathans / Kalash on a genetic level *want* to have an ancestor / proxy that has the right level of ANEness vs Basal Eurasianness and Sintashta plus a little BA Armenian fits the bill better than Yamnaya plus a lot of BA Armenian.

I guess a couple of remaining questions on that are

1) Whether it fits the ingroup relationships. Basically, whether the relatedness to WHG and others is right on that basis. Haak used a “World Foci 15 + Ancients"* rather than World Foci 15 set in some of their analyses, and this showed some divergence between different models (demonstrated a difference in behaviour between Esperstedt_MN plus EHG and Esperstedt_MN plus Yamnaya, which were otherwise the same with World Foci 15 alone). I'm not sure whether this would help to check whether the mixing populations fit right, or whether this would just complicate matters.

2) Whether "teal" people actually existed (and if they did, we will probably find them in Neolithic Iran or the pre-Bronze Caucasus), how well they would contribute. If they fit better than Georgia / BA Armenian they could end up taking a larger share from Sintashta. Pakistan-Hindu Kush is very "teal" in the ADMIXTURE.

3) Whether using Onge instead of Dai might expand that fraction a little, as in the ADMIXTURE from the paper where Pathans seem to have a little more of the Onge cluster than they have Dai here. Onge rather than Dai should affect the intra-East Eurasian f4 stats and that should have some effect on the fit. Shame the waiver prevents you using it.

Really cool though. From the ADMIXTURE in the paper, I'd have said the Pathans looked about a 1/3 Sintashta, plus 1/3 teal, plus 1/3 Onge (but that depends on teal being real, which is not certain by any means). This is quite different while being a close fit. I'm more persuaded of the possibility of a large scale mass migration (more than elite) from Sintashta than I would've been before seeing that.

* World Foci 15 + Ancients = Ami, Biaka, Bougainville, Chukchi, Eskimo, Han, Ju_hoan_North, Karitiana, Kharia, Mbuti, Onge, Papuan, She, Ulchi, Yoruba, Starcevo_EN, LBKT_EN, LBK_EN, Spain_EN, HungaryGamba_EN, Stuttgart, Spain_MN, Loschbour, LaBrana1, HungaryGamba_HG, Motala_HG, SwedenSkoglund_MHG, SwedenSkoglund_NHG, MA1, Kostenki14, Ust_Ishim.

Simon_W said...

The problem with the so called Iron Age Armenian is that he isn't really younger than many of the other samples. He's inbetween 1209 and 1009 BC. No less than four of the other Armenians have lower age estimates younger than 1009 BC. So whatever difference there exists between RISE408 and the others, it's probably random. Matt found him shifted towards EEF, so maybe he has just a little more of the indigenous ancestry, in line with his lower Kalash component.

Judging from Krefter's datasheet, the oldest BA Armenian sample which dates to about 1800 BC has already R1b-M269. That his slightly younger fellow RISE423 dating to about 1300 BC has a Dodecad K12b North_European share comparable to North Caucasians and unlike South Caucasians allows for a northern origin of this R1b-M269. Though I'm not sure, Armenian-like people also carried ANE-related ancestry to the steppe, otherwise it would have been more diluted.

Simon_W said...

@ Krefter, yes I've grasped that it was from Sweden. According to the list posted by David the Corded Ware R1b is RISE1, and according to the Supplementary Table 1 of the paper RISE 1 is from Poland, from Oblaczkowo in Malopolska.

Simon_W said...

@ Nirjhar

Well, since the Kalash component seems related to R1a/R1b and the Caucasus component to J2, I would guess that PPIE is rather related with the Kalash component, because of the relationship with Uralic. J2's ultimate origins were far from any Uralic-related languages.

Simon_W said...

Sorry, in Wielkopolska (greater Poland, not lesser).

Seinundzeit said...

Matt,

For what it's worth, the ASI construct as conceptualized in Reich et al. shared more drift with Dai than it did with Onge. Also, our very own Chad has the Onge samples, and he hasn't been able to create a decent ASI cluster with only them in ADMIXTURE. He always needed to add East Asians, as ASI seems to be more East Asian-shifted in relation to the Onge.

In addition, using d-stats, it seems that the Onge are an unambiguously ENA population, nothing distinct about them. Please take a look at these results:

Chimp Onge Loschbour Japanese 0.0586 12.384 16786 18874 351075
Chimp Kostenki14 Japanese Onge -0.0045 -1.106 15836 15693 332566
Chimp Loschbour Japanese Onge -0.0057 -1.466 16978 16786 351075

As the individual who posted these noted:

"Onge significantly prefer East Eurasian over West Eurasian, and West Eurasian has no significant preference between East Eurasian and Onge."

So, Dai is good enough.

Also, David has used a rather Spartan set of pright populations, which is considered better by those who developed this software (too many pright populations isn't good).

At the end of the day, it's amazing how great these models work. The Sintashta-Pashtun model is the best I've ever seen, for any population, in terms of stats.

Alberto said...

@Seinundzeit

"At the end of the day, it's amazing how great these models work. The Sintashta-Pashtun model is the best I've ever seen, for any population, in terms of stats."

Indeed, and that's the most mysterious part of it. On one hand, the fit is too good to be ignored. Then we know through archaeology of the historical links between Sintashta and BMAC. And finally the Sintashta samples have the nowadays Asian specific R1a subclades.

The obvious caveat is that unless we throw to the bin all we thought we knew, it's hard to explain that if Sintashta people are something similar to Estonians or Lithuanians, by mixing 70% of them with a bit of Georgian and a smaller amount of Dai you can really get a Pathan.

And then there's the f3 stats in the paper with an outgroup which show Sintashta's better match being Andronovo, followed by WHG, SHG, Lithuanian, Afontova Gora, Icelandic, CW, Estonian... and we find Pathan in position 80, much below Spanish, Sardinians, Ashkenazi Jews, Maltese, Cypriots or Druze, and just above Moroccan_Jew.

So while the links are very intriguing, there's also something strange going on. I guess that more tests might reveal something more, though ultimately we might need more Asian DNA (from BMAC and other relevant places) to be able to put this finding into its right context.

Unknown said...

Alberto

"can understand a rather small group could migrate from Europe to the Sintashta area and be the founders of such culture. And I can understand even that they migrated south and become a ruling elite, and even change the language. But when we have a big enough society (like the IVC was), a ruling elite represents 0.1% of the population (that's why they are an elite in the first place). The masses descend from the masses. That's how societies with a pyramidal structure works.

To have a big genetic impact, a ruling elite doesn't work. We need a mass migration. And that's where a mass migration from Europe to Sintashta and then to South Asia becomes difficult to explain. Even more looking at the genetic profile of Sintashta and modern Pathans."

If this modelling is correct; which appears to be so, not to mention the shallow age of Z93 (going by full sequencing & phylogeny), then we need to explain the apparent conundrum.
But maybe not. Wasn't the Indus plain largely depopulated after the Harappan collapse ?

Unknown said...

Chad

Looks like you're right about I2a2. Looks eastern
But that could be anywhere between Carpathio-Balkan regions and Russia..

Alberto said...

@Mike

"If this modelling is correct; which appears to be so, not to mention the shallow as of Z93 (going by full sequencing & phylogeny), then we need to explain the apparent conundrum.
But maybe not. Wasn't the Indus plain largely depopulated after the Harappan collapse?"

Yes, it's a possibility that the collapse was previous to an invasion and that the mass migration needn't be so large because they filled a void. I don't know the details good enough to have an opinion about it.

But that still leaves the purely genetic contradiction question open. It's not only that a mass migration from Europe to India is strange in itself. It's that Pathans genetically have never looked like 70% Baltic, plus a small part of Georgian and a smaller part of Dai. So I think that this needs further investigation to understand what is really going on.

Krefter said...

@alberto,
"The obvious caveat is that unless we throw to the bin all we thought we knew"

I don't think anyone is taking the percentages literally. Maybe it has something to do with shared Teal or another type of ancestry.

Unknown said...

Capra

Frachetti has further developed his theories on pastoralism a later article- which I linked. He clearly sees a role for agricultural communities for the west (Pontic) steppe. Central steppe (Botai, etc) truly seemed to have gone from hunting wild horse to domesticating it for food resource. Different mechanisms again were operant at the eastern most extent- where he suggests a role for contacts with Central Asian agriculturalists; as well as more western steppe regions.

I would be very interested in a propper assessment of Afansievo aDNA .

Unknown said...

Alberto
You're of course correct. Cannot wait for actual aDNA from Central Asia and northern India.
Maybe the modelling is picking up common ancestry rather than cause/effect ?

whatever the case, the peculiar distributions of R1b vs R1a in North eurasia requires explanation.


Alberto said...

@Krefter

"I don't think anyone is taking the percentages literally. Maybe it has something to do with shared Teal or another type of ancestry."

Yes, I guess we just can't take them literally, no matter if it's with those numbers with which the program get that great fit. When using Yamnaya and BA Armenian (for Kalash), the populations and proportions do make more sense from a genetic point of view. So we have to find out why with Sintashta it gives those results. There has to be a good reason for it, but no idea what it is exactly.

Chad said...

A better test would be Chimp Paniya Onge Dai. Onge and Dai do not plot even close. The ASI cluster usually takes all 11 Onge and 4-6 Kinh.

Seinundzeit said...

Alberto,

Actually, long ago, Everest was able to model the HGDP Pashtuns as 66% Lithuanian, using a program in the older ADMIXTOOLS release. With Georgians, he only got 23%. So, that program had Pashtuns as 66% Lithuanian and 23% Georgian.

At the time, I dismissed the results. Not anymore though.

You guys are still thinking in terms of ADMIXTURE. ADMIXTURE just isn't good for this sort of thing, the clusters it produces are "ethereal", and you never know when the output makes sense, since there is no standard via which one can judge the output!

By contrast, if a model can't work, qpAdm simply has it as "infeasible". If a model is inaccurate, or poor, we can see that in the qpAdm output. In this case though, the output is amazingly good.

Also, we need to remember that the IVC collapsed, and that a massive void was left in the region.

DMXX said...

Davidski,

After you're done arranging your new teapots, any chance you could check the Iranian and Kurdish results with Sintashta in qpAdm please? Thank you in advance.

I also cosign Seinundzeit's point about rigid "ADMIXTURE thinking". Another issue is stable components forming from unspecified genetic drift, as is the case with "Kalash" and likely so with "Gedrosian". Applying these hybrid components onto samples which likely contributed to their development is bound to produce spurious results.

Unknown said...

Davidski'a PC Plot of the BA Armenians looks very much like the modern Lezgi and Chechens who speak a NE Caucasian language with similarities to Hurrian/Urartian from the Russian linguists like Sergei.
I'm thinking that the J2b2f (if true and if on the Indian branch) might reflect an Indo-Aryan adstrate like the Mitanni. We found in the Turkey paper (Cinnioglu et al) a sample who is J2b2-M241 from Diyabakir, Turkey who from his YSTR pattern (dys461/A7.2) is like the Pakistan/India J2b2s not the Balkan J2b2s and maybe a relic Mitanni Indo-Aryan.

Nirjhar007 said...

Simon,
''because of the relationship with Uralic''
The relationship with Semetic and Sumerian is very telling IMO which are basically ignored..

Aram said...

Again, haplogroup J2a4b*-M67(xM92) comprised 51-79% of the Y
chromosomes in the Ingush and three Chechen populations (North-East Caucasus, Nakh
linguistic group), while, in the rest of the Caucasus, its frequency was not higher than 9%
(average 3%). Finally, haplogroup J1*-M267(xP58) comprised 44-99% of the Avar, Dargins,
Kaitak, Kubachi, and Lezghins (South-East Caucasus, Dagestan linguistic group) but was less
than 25% in Nakh populations and less than 5% in the rest of Caucasus.
.....

Therefore, linguistics explained a larger part of Y-chromosomal variation in the Caucasus.
These analyses indicated that linguistic diversity is at least as important as geography in
shaping the Y-chromosomal landscape, and suggested that the pronounced genetic structure of
the Caucasus might have evolved in parallel with the diversification of the North Caucasus
languages.
.....
http://mbe.oxfordjournals.org/content/early/2011/05/13/molbev.msr126.full.pdf+html

also

Chechens are connected with the Middle East on the Y-DNA side, but closer to Europe in terms of mitochondrial DNA.
[I. Nasidze, E. Y. S. Ling, D. Quinque et al., "Mitochondrial DNA and Y-Chromosome Variation in the Caucasus," Annals of Human Genetics (2004) 68,205–221.]

Clearly linguistics in North Caucasus correlates with Y DNA markers. It was the result of male founder effect. The womens are more Steppe related that is why they are shifted to North Europe.
The most probable reason that this J2 and J1 moved to high mountains of Caucasus is the invasion of another population. Most probably the R1b.

I think the greatest irony of this 'Hurrian' R1b story will be the finding that this R1b-Z2103 came from Balkans. This will nicely explain their Atlantic shift. :)

Gioiello said...

@ Aran Palyan

"I think the greatest irony of this 'Hurrian' R1b story will be the finding that this R1b-Z2103 came from Balkans. This will nicely explain their Atlantic shift. :)"

Aram said...

One can take a look at the J2a-Y3612 subclades and see how young are the TMRCAs of this people. Clearly a Middle Bronze Age ( some after Iron Age) founder effect. They are even branches who start in Current Era.
J2a Hurrians fleeing to Caucasus from R1b invador Hurrians :)

http://www.yfull.com/tree/J-Y3612/

Aram said...

Oh I forget to say that J2a-Y3612 people are Chechens and Ingushes. ( Nakh people )

Davidski said...

Here are some decent qpAdm models for Armenians, Assyrians, Iranians and Kurds.

Armenian
baArm 0.263
Iraqi_Jew 0.704
Han 0.033

chisq 1.689 tail prob 0.429678

Assyrian
baArm 0.022
Iraqi_Jew 0.948
Han 0.030

chisq 1.581 tail prob 0.45365

Iranian
baSin 0.202
Iraqi_Jew 0.743
Han 0.056

chisq 0.536 0.764725

Kurdish
baSin 0.163
Armenian 0.819
Han 0.018

chisq 0.298 tail prob 0.861593

I could probably do a lot better for Armenians and Assyrians (and maybe get rid of the Han noise), but not for Iranians and Kurds at this time. In any case, the really outstanding models might have to wait until we get some Neolithic genomes from the Near East.

Alexandros said...

Aram,
interesting points about J2a in the Caucasus. Would you then place the origin of J2a somewhere in Hurrian land? If yes, where exactly would that be? North Levant? Anatolia? further inland towards Mesopotamia? The presence of J2a in Iron age Altai is making the situation regarding the origin of this haplogroup further complicated. To me it appears that at some point during the Bronze age there was a 2-directional migration of J2a people, reaching both southeast (i.e. Crete, Italy, etc.) and central (i.e. Hungary) Europe, as well as central Asia. It is very likely that such migrations were linked to metallurgy.

pnuadha said...

@Matt

Maybe that links to the discussion between Colin and Simon, re the Yamnaya estimates in the Haak paper, and what that means for Corded Ware Central Europeans as candidate ancestors for British Isles plus Scandanavia.

What we see in the Haak paper, is that where Central and East Europeans have less estimated Yamnaya they have more WHG (although some of this may in fact be East Eurasian for populations like Russians and Estonian).

So where the populations in perhaps Poland may have slightly less Yamnaya ancestry than Norway or Scotland, that seems like it may be due to refluxes of new and successive cultural trends with more EHG ancestry from Russia or the Baltic, replacing some "Yamnaya" with EHG, after the early dates of the Corded Ware (as well as a MN European resurgence). Just as at the same time in another direction the CW culture may have also been spreading MN European ancestry to Sintashta.

All of this would be diluting the similarity to Samara Yamnaya, particularly because adding more MN European and EHG into the mix is decreasing the West Asian / Gedrosia element, which is the most signature to the Yamnaya when EHG is removed (and Scots have that odd elevated Gedrosia signature).


Northwest Europeans (Norwegian 24.5/53.5 EHG/YAM, Scottish 22/48.6 EHG/YAM, Icelandic 22.1/49 EHG/YAM) have more EHG and Yamnaya ancestry than East Central Europeans (Belarusian 20.6/46 EHG/YAM, Czech 22.4/48.8 EHG/YAM) so dilution by Russian like folk is not a valid explanation.

There could be nuances to this story. Maybe, however unlikely, Germans and Polish actually have more EHG/YAM than Belarusians (whom i tend to view as an upper bound for the former) because the CT culture had a greater demographic impact on the eastern CW than the western CW. Maybe Central Europe has experienced a greater degree of dilution of their EHG/YAM due to migrations from the south, after the bronze age, than northwest Europe. In any case, we can't expect the northwest europeans to have more (true) CW heritage than the central europeans and seeing how northwest europeans don't have less EHG/YAM, but instead have more, the northwest europeans need an additional, or better yet, a different source of EHG/YAM.

Yamnaya to hungary, hungary to bell beaker, and bell beaker to northwest europe is the best explanation in the sense that there is a strong archeological and genetic link between these groups. Its also a good explanation for the transfer of language, which is something hasnt been emphasized enough in this thread. At the same time CW is just a bad explanation for autosomal dna, ydna, and language of northwest europe. I will point out that the chain i describe above is still wanting of evidence that the yamnaya in hungary had the demographic potential to dramatically change northwest europe and that they actually became fully incorporated into the eastern bell beakers. Going beyond a strict adherence to the above chain, we may find that the yamnaya themselves actually moved beyond hungary and into austria and the czech republic before being absorbed into the bell beakers. It also might be the case that the yamnaya impact on northwest europe was strongly mediated by a migration that went north of the carpathians; but again, this will be a different phenomenon from the corded ware whose language, ynda, and reach don't add up for northwest europe.

pnuadha said...

btw, to simon and those making similar comments about the ba hungarians, I don't expect hungary to be genetically homogenous nor exhibit genetic continuity around the time the yamnaya were in hungary. in other words, the ba hungarians were not like the yamnaya in hungary who's descendants probably kept moving to the northwest.

Davidski said...

It looks like Corded Ware expanded in both directions.

In the west they swamped Old Europe, and in the east they swamped Yamnaya, and eventually a large part of Asia.

Alberto said...

@Seinundzeit

"You guys are still thinking in terms of ADMIXTURE".

The problem is that it's not just admixture, especially in an unsupervised run and based on modern samples that could have been misleading us. Even with calculators run in supervised mode and based on ancient genomes (like K8), the results simply don't match.

But there are formal stats too (I referred to the f3 stats on the paper above). And David's favourite IBS. We still haven't seen IBS of Sintashta, and there could be a surprise, but if CW is anything to go by, there also the first places are occupied by Estonians, Norwegians, Polish,... and Pathans appear in position 96, 5 below Moroccan_Jew, for example.

That's why I said that "unless we throw to the bin everything we thought we knew...". And I'm not trying to discard the results of qpAdm as just an oddity, mostly because we also have archaeological evidence of contact and because of the haplogroups. All I'm saying is that we need to further investigate the relationship between these populations (which surely exist in some significant way) to better understand it and put it in its right context. We can't just accept the results at face value and think that's the whole truth to it.

Davidski said...

Affinity tests, like IBS or f3/f4 stats, are interesting, but not very practical for populations with complex histories, because even a small amount of admixture from a more diverged population can dampen affinity significantly. Pathans have around 10% of ASI, plus a couple per cent of East Eurasian ancestry, which means they'll be pulled down the list in affinity tests to ancient groups with no ENA ancestry, like Sintashta.

The best methods in such cases are either extremely well designed ADMIXTURE models, which are hard to achieve even now with the samples we have, or qpAdm, which is based on allele correlations deep in the phylogeny, so it probably picks up a lot of ancient shared ancestry in some cases, but is overall probably the most practical solution in this context.

Aram said...



Thanks Davidski

No need to get rid off of Han and force them to be more similar to modern Armenians.
I think the situation in Armenia is quite clear and it was expected from Marc Haber study.
A lot off heterogeneity in Armenian highlands during Bronze Age. We will have great surprises from places like Van, Sasun, Erzrum, Malatya etc if we have aDNA from there. We will see a lot off SWA in south west a great deal of ENF in Western parts were G2a is high. Let's not forget that Sasun Armenians have 20% of T and 18% of R2a! Only God knows what we will see there. All this populations was homogenized by IE tribes culturally, and later politically by the iron rule of Urartu. Et voila!

Anonymous said...

@Colin Welling

Belarusians have more ANE, WHG and "teal" than Germans and less EEF than even Scandinavians. I am not sure where you got the idea that Belarusians have more Neolithic farmer ancestry than Germans.

Aram said...

Alexandros
The J2a is very old. Yfull tree shows a 18000 years of TMRCA. I think the most probable origin is in Zagros maybe deeper in Iran. They had a demic diffusion, reached India. I think You should expect some old J2a in Greece and Italy. It seems that J2a was a seafarer. And it was there most certainly before Phenicians started their sea faring. I can't say anything about their linguistics because this is a so big areas.

I just can say that Nakh branch is only a subbranch of J2a4b. Most certainly their original homeland was in Nakhchivan or somewhere near lake Urmia a place known as Dzurdzuk.
Armenian historians mention their presence as Nakhchamats. So if they are the remnants of Hurrian populations then most certainly we can say that some branches of J2a spoke Hurrian.
But only a branch and not all J2a.

It could be also Greek historians also knew them as Gargareans. Ingush's endonym is Galgay.
https://en.wikipedia.org/wiki/Gargareans

Balaji said...

From page 36 of the Supplementary Information of Allentoft.

“K = 6
Populations from Oceania form their own component, which is present at low levels in BA Armenians and Yamnaya / Afanasievo, as well as some later BA Asians.”

This is similar to what was found in Haak. This must mean something.

Grey said...

capra

Interesting stuff. I can see how crop-raising might more easily lead to animal domestication but I wondered if there might be exceptions.

The main reason for wondering about the possibility is if a population did domesticate something like goats / sheep before they figured out growing crops then for that moment in time they could be incredibly mobile.

Just a "what if?"

pnuadha said...

@skilur

What???

Davidski said...

Skilur is right.

Alberto said...

@David

Yes, I agree about the shortcomings of methods that estimate overall affinity when a population has some small amount of admixture from a very divergent population. And that the best method would be a very well designed ADMIXTURE with the exact ancient samples. And that's why I think that until we have ancient DNA from Asia (for this specific case BMAC looks relevant), we can't know exactly what is what.

But the best we can do till then is to look at all the available tests, trying to understand the whole picture. In this case I do think that the qpAdm results are relevant and mean something, but qpAdm also has shortcomings and cannot do magic to guess without the right input. You see how using Yamnaya instead of Sintashta gives very different proportions. So I still think that the wise thing is to further investigate this to try to understand it better.

About those f3 stats, here's what the authors say about them:

"f3(Outgroup;Population1,Population2)
This “outgroup”-f3 statistic9 is expected to be proportional to the amount of shared genetic
drift between Population1 and Population2 in their common ancestral population until their
divergence. Unlike methods based on pairwise distances such as FST, genetic drift specific to
Population1 or Population2 does not affect this statistic."

Simon_W said...

http://jpst.it/zA5m

The map on the left shows the archeological cultures of Italy at the beginning of the Iron Age. The dark and the grey areas are dominated by cremation burials. The white areas are almost exclusively with inhumation burials. The map on the right shows the incidence of R1b-U152/S28. There is a remarkable correspondence between cremation burials and high incidence of R1b-U152.

In Northern Italy the cremation custom became common in the Middle Bronze Age. In the early Bronze Age, inhumation was still the standard. RISE486 lived around 1950 BC in Northern Italy, that is, in the Early Bronze Age. And he's still I2, not R1b. So chances are high that R1b arrived in northern Italy during the Middle Bronze Age, together with cremation.

In central Italy cremation didn't get common before the Final Bronze Age. It arised there with the advent of the Protovillanovan. The exact origin and ethnic affiliation of the Protovillanovan is still uncertain, but it's related with the Urnfield cultures of Central Europe, and hence it makes sense that it introduced the central European haplogroup R1b-U152 to the rest of Italy.

Two years ago, a paper on the y-chromosomal variation of Italy showed that there are two principle components of yDNA variation, and the first PC splits Italy into a northwestern and a southeastern part. See the map on the left side: http://jpst.it/zA5z
This first PC is strongly loaded by R1b-U152, to a lesser degree also by R1b-S116 and R1b-L2 (all R1b variants which must have come from the north).

But in case someone thinks that these central European influences associated with cremation explain the advent of the Italics in Italy, this would be quite wrong. Because the Italic tribes are strongly associated with the southeastern yDNA pole of PC1, and with inhumation burials, as this map shows (Italics in yellow): http://jpst.it/zA5J

Therefore I think that the Protovillanovan was ethnically Ligurian IE (which is related with Celtic) and that the Italics arrived earlier, from the southern Balkans. Hence their high incidence of J2 and the strong West Asian admixture. (Etruscan then must be from some pre-existing group of central Italy which assimilated the Ligurians.)

rozenblatt said...

Does anyone know if we can hope to see samples from Ancient Near East or Central Asia anytime soon? Like from BMAC or Sumerians? I know about ancient Anatolian from Kumtepe, but anything else?

Davidski said...

Yeah, I'm getting hints and hearing stuff second hand that there are a lot of ancient samples from west Asia and also southeastern Europe being sequenced right now. But I can't give you any details, and I haven't a clue when they'll be featured in a paper.

Suffice to say these samples are not the ones listed on the Pinhasi-ERC site.

Simon_W said...

If western European R1b is from the Corded Ware, and at present it strongly looks like it is, then it gets tricky to explain either the presence of Italic in Italy or the relatively close relationship between Italic and Celtic. It seems quite impossible to construe an archeologically plausible migration originating in the Corded Ware and ending up in Central and Southern Italy while leaving out the North. On the other hand, if Italic and Celtic split already at the time when Yamnaya went to the Balkans and Corded Ware to Central Europe, then Italic and Celtic would be much less similar.

But as I suggested above, I think Italic came from the Bronze Age Southern Balkans. And thus it may well be derived from Yamnaya movements to the Balkans. But then it's inevitable that Celtic didn't originate in the Corded Ware, but moved to central Europe later, coming out of Bronze Age Southeastern Europe, following up the Danube. I'm thinking here specifically of the Tumulus culture which rapidly spread over large parts of central Europe at the start of the Middle Bronze Age. This would elegantly explain the Italian-like admixture in modern central Europeans without necessitating an unrealistically strong effect of the Romans. And as Krefter has calculated, even modern Northwest Europeans have some slight extra East Med admixture that's not explained by Bell Beakers. And needless to say that J2 is present at low levels almost everywhere, even in Sweden and northwestern Europe. Oh and we already have a J2 male from the probably Celtoid Urnfieldculture.

This would mean that the crazy R1b duplication of the Bell Beakers had nothing to do with Celts. Actually we don't know what languages were common in the Bell Beaker complex. But this would also very plausibly explain why R1b looks more associated with Basques, ancient Aquitanians and northern Iberians than with IE southwestern Europeans – if it spread before Celtic related languages spread.

Grey said...

A possible reason why something odd needed to happen along the western and northern edges of Europe is soil pH.

Crops like wheat generally prefer a neutral pH.

Two of the reasons for acidic soil are 1) heavy rainfall (e.g. atlantic coast) via leaching and 2) conifers (high altitude or latitude e.g. northern forests).

Maps showing the range of LBK show gaps along the western and northern edges so I'm thinking that maybe the acidity of the soil was the barrier to further expansion which required some kind of innovation.

Cultures like Funnelbeaker and Globular Amphora could then be seen as attempts to create a form of farming/herding that would work in those northern forest conditions leaving a question mark over the solution that worked along the Atlantic coast.

(nb millet and oats do better than wheat in acidic soils.)

(I haven't found a good single page that lists all this together but some of the random links with pieces of it are:

http://soilquality.org/indicators/soil_ph.html

http://www.buys-online.net/soil_ph.htm

http://www.soilquality.org.au/factsheets/soil-ph-south-austral

http://www.eea.europa.eu/publications/report_2002_0524_154909/biogeographical-regions-in-europe/the_atlantic_region.pdf
)

Grey said...

Simon_W

A possible explanation might be large scale tribal migrations combined with traders running out ahead like a bow wave. This would fit better if the tribal migration side wasn't one big continuous wave but more piecemeal i.e. a bite sized piece of territory was taken and then things were peaceful for a while until the population grew then another bite of territory followed by another period of peace etc.

If so the people running ahead might have CW dna but speak the languages of the people they were living among.

Alexandros said...

@rozenfag

I am pretty sure that the UK/Spanish group who recently published mtDNA data of Neolithic Near Eastern farmers from northern Levant (http://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1004401) would be interested in autosomal analysis of their data. They must be currently working on that. If for any reason they are not, then they should pass these samples on to a different lab. Those samples are too precious!

Seinundzeit said...

Alberto,

If it were merely shared affinity via something else (between Pashtuns and Sintashta), Iranians and Kurds would also display huge percentages of Sintashta, as they tend to overlap with Pashtuns in those sorts of cases. But here, that just isn't happening. Pashtuns are consistently 60%-70% Sintashta, while Iranians and Kurds are, respectively, only 20% and 16% Sintashta.

Also, the fact that Sintashta + Georgian for Pashtuns is the best fit we've seen ever produced by this software, just can't be ignored. The fit is an outstanding model, incredibly good. There is just no going around that. The people of South Central Asia seem to be predominantly steppe Proto-Indo-Iranian, with the addition of some extra West Asian + ENA admixture.

This probably explains why Pashtun + Stuttgart, or Pashtun + BedouinB, worked great in "threepop" tests for West Asian highlanders and Caucasians. Armenians, Iranians, Kurds, and eastern Turks were consistently best modeled as South Asian + European Neolithic/BedouinB. I guess the reason for this is because South Central Asians are direct descendants of Bronze Age steppe populations, and can proxy as the Indo-European ancestors of modern West Asians.

At the end of the day, everything makes sense now.

*A side note, but I'm sure everyone knows that Pashtun R1a1a, which constitutes around 50%-70% of the modern Pashtun y-DNA profile, is directly descended from Sintashta R1a1a. So, the uniparental markers are congruent with everything else.

Alexandros said...

Aram,

When you talk about presence of 'old J2a' in Italy and Greece, are you referring to the Neolithic? To be honest if you asked me a couple of years back, I would say that I am almost certain that J2a was a Neolithic hg of the Levant, Anatolia and Crete/Cyprus and spread from there to the rest of Europe through several successive migrations from the Neolithic up to the Roman times. The aDNA data we have so far do not seem to support this, as BA Italy appears to be a plain mix of I2 and G2 and I am pretty sure Greece was something similar. Also no J2 among ENF samples from the rest of Europe. I do not see any room for J2 in Neolithic Levant and SE Europe and we certainly do not have the evidence so far for such a presence. Therefore, based on the data currently available, I am leaning towards a BA/IA expansion of this haplogroup.

pnuadha said...

Skilur is right.

Of course he is, but I would never contest such an obvious fact. It also has nothing to do with the purpose of my post. Hence the ....??? WTF

Davidski said...

Also, another thing, what makes you think Icelanders, Norwegians and the four Scots from Argyll (two of them kind of related) are representative of Northwestern Europe?

Why didn't you use the English from Kent for your example? Even Orcadians would be a more sensible choice than Icelanders and the four Scots.

pnuadha said...

If western European R1b is from the Corded Ware, and at present it strongly looks like it is

Thats completely against the evidence. I just explained how corded ware can't be the source or western r1b, let alone the sheer lack of evidence for such.

Northwest europe (Scotland, Iceland, Norway) has more EHG/YAM then east Central Europe (Belarus, Czech), and surely more than Germany and Poland too, despite having less true CW ancestry. There must be a separate source for yamnaya like genes in Northwest europe. The autosomal change happened so quickly that founder effect on ydna could not have taken place.

The r1b yamnaya and early italo-celtic branching from PIE steppe is literally staring us in the face and it doesn't point to CW.

Anonymous said...

@Colin Welling
Maybe i misunderstood you but you talked about dilution by CT in the eastern part of Corded Ware, which makes NW Europeans more Yamnaya/Corded Ware like. This is not true and Ukrainians(except of Carpathians), Belarusians and Poles have all less EEF and more ANE on average than Germanic Scandinavians. Haak estimates for Yamnaya ancestry among modern Europeans are not accurate and Yamnaya and Corded Ware ancestry peaks not in Norway but in Lithuania.

Krefter said...

The R1b-U106 from Swedish Corded Ware is U106*. He's negative for all basal subclades of U106, and has his own private mutations, so is apart of his own clade.

http://www.anthrogenica.com/showthread.php?4664-Request-Y-DNA-haplogroup-results-from-Allentoft-2015&p=90192&viewfull=1#post90192

Unless U106 isn't well defined and U106*s are common, this is evidence U106 originated around Scandinavia and expanded recently(Not something crazy like Neolithic).

pnuadha said...

@david

Also, another thing, what makes you think Icelanders, Norwegians and the four Scots from Argyll (two of them kind of related) are representative of Northwestern Europe?

Why didn't you use the English from Kent for your example? Even Orcadians would be a more sensible choice than Icelanders and the four Scots.


I can take the extreme northwest (norway, Iceland, and Scotland) if I want, they still have more r1b, more EHG/YAM, and less corded ware than east central europe. So the argument is there in its entirety! (btw, iceland is half viking half irish, fully northwest european, and high in EHG/YAM).

I can also include England to Northwest Europe, and Northwest Europe will still topple Germany and Poland, who's upper bound is belarus and czech. The point stands.

Davidski said...

Poland is surrounded by the Czech Republic, Lithuania and Belarus, but closest to Scandinavia, and you picked the Czechs and Belarussians as the upper bound for Poland?

pnuadha said...

@skilur

Maybe i misunderstood you but you talked about dilution by CT in the eastern part of Corded Ware, which makes NW Europeans more Yamnaya/Corded Ware like

Just ignore that comment. I was only exploring the more complicated senarios to make the point that even in the relatively unusual cases, CW can not be the main source of western r1b. The unusual case i was referring to is the a senario in which Belarus had less EHG/YAM than Poland and Germany, because Belarus got their CW influence from the CT while their base was overwhelmingly WHG. It does not contradict the fact that Belarus has less EEF than Poland and Germany. Anyways, if you understood my post you would know that I used Belarus and Czech's as an upper bound for the EHG/YAM and this weird senario i mention above is a case when the germans and the poles actually exceed that upper bound. So it is a worse case senario in my argument. If you think that is impossible, so be it. It in no way hurts my argument.

Haak estimates for Yamnaya ancestry among modern Europeans are not accurate and Yamnaya and Corded Ware ancestry peaks not in Norway but in Lithuania.

Where do you get this? Ill come out and say that I trust these research papers more than i trust the work of bloggers.

pnuadha said...

Poland is surrounded by the Czech Republic, Lithuania and Belarus, but closest to Scandinavia, and you picked the Czechs and Belarussians as the upper bound for Poland?

Ok lets completely surround poland and take the max of its neighbors. That would be Norway. In that ideal case, Norway has less CW ancestry and equal EHG/YAM ancestry. Hence, we need another source of EHG/YAM in Northwest Europe.

Anyways, i don't see why Belarus and Czech R. is not a good upper bound for Poland and Germany, the only source populations for CW in Northwest Europe. How do you think Poland and Germany fit into said pecking order?

Davidski said...

I think your whole pecking order is going to implode when we get steppe samples from west of the Don.

Unknown said...

Dave, what collection of Hgs do you hypothesize we might get from the Nth Pontic steppe/ north Carpathian region ?

Davidski said...

R1a and I2a.

Alberto said...

@Seinundzeit

"Also, the fact that Sintashta + Georgian for Pashtuns is the best fit we've seen ever produced by this software, just can't be ignored."

I agree and that's what I've been saying. I won't quote myself here, but you almost repeated my words exactly (just read it up there).

But of course, there is a long, long way from not ignoring a piece of evidence to cherry picking that single piece and ignoring everything else. I'm avoiding both extremes.

I'm honestly surprised that for you it all makes perfect sense and happily take those results literally. I myself am surprised by them, but cautious about their real meaning. At least for now.

Where do you think that Pathans got their ~70% more ANE than what is accounted for in that mix?

Davidski said...

These ancestry proportions are estimated from f4 ratios, and f4 ratios don't see EHG as a mixture of ANE and WHG. Sintashta should be mainly EHG, I guess.

So it's hard to say how we can reconcile this model with the ANE model, because ANE might not even exist in this model.

a said...

^^^
another teapot theory bites the dust.

Alberto said...

@David

Yes, the model doesn't seem to work here. For ANE and for WHG levels. Unless we've just discovered by chance that ANE only is a composite of ASI and WHG?

Nirjhar007 said...

Yeah SC Asians and Iranians etc have significant EHG-WHG Ancestry which greatly corresponds to their R1a frequencies....

Gioiello said...

@ Nirjhar007 said...
"Yeah SC Asians and Iranians etc have significant EHG-WHG Ancestry which greatly corresponds to their R1a frequencies...."

Nirjhar007 said...

Gioiello, someone from your province tells me that you are an Italian Nationalist:D.

Gioiello said...

Others say that there are Polish Nationalists etc. Do you remember how eng-molgen finished and Anthrogenica was born? And who said that, was a Nationalist of which country? My name is Gioiello Tognoni, I published books and papers, and you? I am Italian, that's all.

Nirjhar007 said...

I'm? Just a nobody but the fella who said knows you very well but i can't tell due to privacy reasons;)...

Gioiello said...

My theory od an Italian Refugium has nothing to do with the fact that I am Italian and it will be right or wrong independently from my beung an Italian. You seemed agree with my theory that R1a descends from the Northern European hunter-gatherers. Am I a Scandinavian Nationalist too?

Gioiello said...

I didn't ask you his/her name. I know too many morons. It would be difficult to choose one... but I'd prefer to speak only about genetics, and linguists if you want, and many others. Do you know Latin? We could speak about Cicero et alii...

Nirjhar007 said...

No i didn't agree with the R1a descends from the Northern European hunter-gatherers just NC Europe and Central Asia can be the two candidates for the origins of R1a, well its not my opinion, so you don't have to be angry on me:p and about your theory i have no problems...

Gioiello said...

Of course, as I think being very smart, I don't need ask you, because I know perfectly that person... but I could send you many Nationalist writings of his. That infltrators usually do.

Kristiina said...

You say that Pashtuns can be modeled as 66% Lithuanian and 23% Georgian or consistently 60%-70% Sintashta. I cannot find Pashtuns in test.10.odds of Chad but I can instead find Kalash and Pathan with the following components:
Kalash: EEF 0.00, Near East 0.00, Euro HG 0.09-0.11, BA Cauc 0.60-0.64, S Indian 0.21-0.26
Pathan: EEF 0.07-0.10, Near East 0.02-0.03, Euro HG 0.10-0.11, BA Cauc 0.47, S Indian 0.25-0.28
Sintashta: EEF 0.19, Near East 0.00-0.03, Euro HG 0.47-0.48, BA Cauc 0.27, S Indian 0.00

If Kalash usually do not have any EEF, can they have a really big Sintashta-related ancestry?

IMO, Kalash and Pathan cannot be modeled without South Indian component as it forms one fourth of the Pathans, so I do not get how Pashtuns can be modeled without input from India.

Is that 70% Sintashta possible if, for example, Kalash have no EEF, 1/5 of Sintashta EuroHG, twice as much BA Cauc and even ¼ of S Indian?

Are Pashtuns very different from Kalash and Pathan?

Nirjhar007 said...

I'm too tired to talk on anything now btw i know you are a prominent scholar btw can you link me your website? if you have any?.

Gioiello said...

Thus did Western European Hunter-Gatherers come from Central Europe and not from the Isles? And all those R1a-M420, one of them examined from Yfull and old, it seems, 17200 years, but possibly also 24400?

Kristiina said...

Correction! I meant to say 1/5 of Euro HG.

Gioiello said...

I thought that the "eminent scholar" were you. i am only a litterate, no career, no political links... no website. Nothing of nothing.

rozenblatt said...

Kristiina - Wikipedia says that Pathan is just another name for Pashtun people: https://en.wikipedia.org/wiki/Pathan_%28disambiguation%29

Alberto said...

@David

"f4 ratios don't see EHG as a mixture of ANE and WHG"

Did you, or anyone, ever try to model EHG as a mixture of WHG and something South Asian (Dai, Malay, Sakilli,...) to see what happens? Any f3 signal of admixture, qpAdm models,... This D-stat is significant:

Loschbour EHG Dai Chimp -0.0137 -2.506 341763

Though this one also:

Loschbour EHG She Chimp -0.0144 -2.665 341763

It would be interesting to know what ANE is exactly. If the cluster was removed from K8 and the synthetic sample run through the resulting K7, would be be WHG + SE (+ a bit of EE)?

Matt said...

@Sein, hi, re: Onge, as I understand it qpAdm works by residuals fitted to masses of f4 stats.

So unless, e.g. f4(Yoruba,Dai;Sintashta,Ulchi) is exactly the same as f4(Yoruba,Onge;Sintashta,Ulchi), which it won't be because Ulchi and Dai are much closer in the phylogeny, then there will be a difference in the residual fit.

It may be a small difference, it's hard to tell.

And Onge is of course probably phylogenetically closer to the actual admixing population for Pashtuns / Kalash.

So the dispute is not whether Onge are on the ENA branch.

Alberto: The obvious caveat is that unless we throw to the bin all we thought we knew, it's hard to explain that if Sintashta people are something similar to Estonians or Lithuanians, by mixing 70% of them with a bit of Georgian and a smaller amount of Dai you can really get a Pathan.

Yeah there is the strangeness of the how this relates to either a WHG-ANE-ENF model, or an WHG-ANE-Basal model, as you note, if Sintashta is in fact anything like Corded Ware, as the paper seems to claim.

Is the position of Pashtun on f4(Yoruba,Pashtun;WHG,ANE) much like the expectation for the combination 70%Sintashta,10%Dai,20%Georgian? That's some information qpAdm doesn't take into account without including WHG and ANE, except indirectly through Karitiana vs Dai/Han. Another example is f4 comparisons involving EEF, etc.

Davidski: Pathans have around 10% of ASI, plus a couple per cent of East Eurasian ancestry, which means they'll be pulled down the list in affinity tests to ancient groups with no ENA ancestry, like Sintashta.

This is a good point. Trying to quantify this effect, just for a comparison under f3(Ju_Hoan_North,Test,Sintashta), going least up from there:

Mansi = 0.2582345, Pathan= 0.2590559, Kalash = 0.2607518, Nogay = 0.2605475, Tajik = 0.2635420, baKarasuk= 0.2648039, Georgian = 0.265330, Chuvash = 0.265804, Mordovian =0.2725367, Corded Ware = 0.2755243.

According to various ADMIXTURE, Chuvash have about 20% ENA. Still closer. Nogay, 24% ENA. Still a little closer than Pathan, as close as Kalash. Tajik around 10%. Closer than Pathan or Kalash. Mordovians are around 6%, and of course closer.

Some of this is Siberian ENA though. Trying to take account of that f3 with Sintashta, Nganasan is 0.2408751, Xibo is 0.2367805, while Kinh is 0.2353016, Dai = 0.2347445, while Onge is only 0.2318051. There's a difference, but I wouldn't expected that alone to have a very large effect. This is hard to quantify though.

@ Colin Welling, nah, I think those EHG+WHG+EN stats aren't useful because they can't register the teal survivor fraction of Yamnaya (which later, more EHG diluted arrivals from East wouldn't have as much of) accurately, so "bung" its "ANE" into EHG, leading to a curious effect of the more EHG populations seeming more WHG (which makes no sense in terms of geography).

pnuadha said...

@Matt

Colin Welling, nah, I think those EHG+WHG+EN stats aren't useful because they can't register the teal survivor fraction of Yamnaya (which later, more EHG diluted arrivals from East wouldn't have as much of) accurately, so "bung" its "ANE" into EHG, leading to a curious effect of the more EHG populations seeming more WHG (which makes no sense in terms of geography).

Ive already addressed your points. Regardless of your proposed Russian like migrations or whatever into Belarus and Czech republic, both of them still have more CW heritage but not more EHG or Yamnaya.

Also, what are you even talking about? If EHG in northwest europe comes from yamnaya then the northwest europeans necessarily have more yamnaya because equivalent yamnaya would imply that the Belarusians and Czechs would have at least as much EHG as northwest europe. I can't believe I have to spell that out. You actually think that the Belarusians and Czechs have more yamnaya than Northwest Europeans (my argument only relies on that not being the case), which is the primary source of EHG in northwest europe, AND that russians mixed with belarusians raising their EHG levels but that somehow Northwest Europe is still estimated to have more EHG than belarus??? Think about the implications of what you say in order to save time and confusion.

Finally, migrations mess up the geographic regularity of the distribution of ancestry. Thats why all the yamnaya samples have come back r1b despite r1b having a huge wester bias in modern times. Its the same with yamnaya. Central Europeans are closer to the PC steppe but they don't have more yamnaya.

Seinundzeit said...

Alberto,

I take this evidence to be determinative, mainly because you don't see such good fits everyday. We are dealing with "perfection" here, from a stas-based perspective.

From a purely pragmatic perspective, nothing shown via ADMIXTURE or PCA can beat this.

For what it's worth, if one is going to dismiss these results, one has to dismiss the whole Haak et al. paper, and thus the whole idea of mass Indo-European migration into Europe. Although, I know you aren't dismissing these results, your'e simply exercising caution.

Matt,

The Onge might make a slight difference in the residual fit, but the general model of 60%-70% Sintashta isn't going to change (as people with the Onge samples can testify). F4-stats cut very deep into the phylogeny, which is where the advantage lies.

Also, I think it's time to scrap the ANE-WHG-ENF model. When the Haak et al. preprint came out, you yourself noted that the model probably needed to be modified substantially.

Kristiina,

Your'e comparing apples to oranges, very different analyses at play here. You are talking about Chad's ADMIXTURE run, and I was talking about qpAdm, and another program in the older ADMIXTOOLS release.

For anyone who's curious,

Chad had some supervised ADMIXTURE run with a Yamnaya component, a Corded Ware component, a Bedouin component, and a EEF component. Despite the presence of all these components, South Asians latched onto Corded Ware (I think Chad said that Sindhis were 70% Corded Ware).

As RK noted a while back, we are at a point in which the formal modelling must come first, with ADMIXTURE as a secondary sanity check. And in this case, ADMIXTURE confirms the formal modelling.

Davidski said...

Guys, relax a bit. I'm going to double check these results with the Sintashta sequences taken from the bams at the ENA.

Unknown said...

Dave

Any word of the breakdown of specific R1a1 subclades found in present study ?

Unknown said...

K15

Swedish Battle Axe Culture

1 North_Sea 39.93
2 Atlantic 18.64
3 Eastern_Euro 16.41
4 Baltic 15.82
5 West_Med 7.64
6 South_Asian 1.07
7 Sub-Saharan 0.36
8 Oceanian 0.12

Single Population Sharing:

# Population (source) Distance
1 North_Swedish 5.69
2 Swedish 6.66
3 Norwegian 6.81
4 West_Norwegian 8.26

80.4% West_Norwegian + 19.6% Erzya = 5.15
77.2% North_Swedish + 22.8% West_Norwegian = 5.38
69.4% North_Swedish + 30.6% Norwegian = 5.38

Dude Manbro said...

Here is the least squares distances between my Eurogenes K15 results and those of the current samples I have from the recent paper:

RISE00 (CWC)
17.93349409

RISE94 (Battle Axe)
15.03828579

M690970 (Sintashta) RISE#?
13.28885145

RISE98 (Nordic LN)
11.91426405

I'm closest to the Nordic LN sample and then the Sintashta sample. Being of Northern Euro descent, I find it very surprising that I am closer to the Sintashta sample than I am to the Corded Ware and Battle Axe samples that are available.

Krefter said...

Dude ManBro,

Several K15 components are more popular in certain region today probably because of drift. Drift that occurred after these ancient people were born. All K15 can do is show what big region these ancient people fit in, not what specific little ethnic group they're closest too. ANE K7 is another Eurogenes test at GEDmatch and it's much better, because it breaks down ancient ancestry.

Unknown said...

Krefter
You're becoming an expert in genomics bud

pnuadha said...

@david

Does the recent paper give estimates/models for the proportion of yamnaya in modern europeans like the last paper did? I don't see it explicitly and I'm wondering if its subtly given.

If not, whats your estimate of the yamnaya/CW contribution to modern europeans based on a ENF/WHG/YAM model or maybe just MN + WHG + YAM based on the results of this paper? (without using a hypothetical "western yamnaya")

People keep saying that this paper does not support a mass migration but I'm still seeing evidence for a 1/3+ contribution of YAM to northern europe which i would define as a mass migration.

Davidski said...

Where did I say there was no mass migration?

Both Haak et al. and Allentoft et al. clearly show that there was a mass migration of Corded Ware people across the North European Plain.

But there is no evidence in either paper that there was a mass migration into Europe directly from the Yamnaya horizon.

Can you show me this evidence if it exists?

Dude Manbro said...

I used the least squares distance formula to compare my K7 values to those of the same samples from my previous post:

RISE00 (CWC)
12.7515640351

RISE94 (Battle Axe)
4.7329577969

M690970 (Sintashta) RISE#?
10.220167858

RISE98 (Nordic LN)
10.239348124

The results vary greatly between the two runs. The Battle Axe sample's values are very close to my results now, but the Sintashta sample is still close, coming in second, slightly edging out the Nordic LN sample.

pnuadha said...

Where did I say there was no mass migration?

you never did. by people i meant maju, bell beaker blogger, and the one who we don't mention. of course they are biased as heck but for some reason they say this paper is less supportive of mass migration.

pnuadha said...

Also, im not aware of any more evidence than you are. I don't see what evidence we will get that is more direct than we already have. It pretty obvious there had to be a mass migration from the steppe bringing EHG and new ydna to much of europe.

Im just wondering what yamnaya, direct or indirect, input you would expect based on this paper.

Unknown said...

Colin

A few of the Hungarian kurgans spilled over into Lower Austria etc. This eastern BB zone was the contact point with the East. Samples from here would be interesting, as well as more Hungarian ones.
I'm sure there forthcoming.

Unknown said...

*they're**

«Oldest ‹Older   401 – 600 of 728   Newer› Newest»