search this blog

Tuesday, February 10, 2015

Massive migration from the steppe is a source for Indo-European languages in Europe (Haak et al. 2015 preprint)


I'll probably end up writing a whole series of posts on this paper. But for now, here's the abstract and a PCA.

We generated genome-wide data from 69 Europeans who lived between 8,000-3,000 years ago by enriching ancient DNA libraries for a target set of almost four hundred thousand polymorphisms. Enrichment of these positions decreases the sequencing required for genome-wide ancient DNA analysis by a median of around 250-fold, allowing us to study an order of magnitude more individuals than previous studies and to obtain new insights about the past. We show that the populations of western and far eastern Europe followed opposite trajectories between 8,000-5,000 years ago. At the beginning of the Neolithic period in Europe, ~8,000-7,000 years ago, closely related groups of early farmers appeared in Germany, Hungary, and Spain, different from indigenous hunter-gatherers, whereas Russia was inhabited by a distinctive population of hunter-gatherers with high affinity to a ~24,000 year old Siberian6. By ~6,000-5,000 years ago, a resurgence of hunter-gatherer ancestry had occurred throughout much of Europe, but in Russia, the Yamnaya steppe herders of this time were descended not only from the preceding eastern European hunter-gatherers, but from a population of Near Eastern ancestry. Western and Eastern Europe came into contact ~4,500 years ago, as the Late Neolithic Corded Ware people from Germany traced ~3/4 of their ancestry to the Yamnaya, documenting a massive migration into the heartland of Europe from its eastern periphery. This steppe ancestry persisted in all sampled central Europeans until at least ~3,000 years ago, and is ubiquitous in present-day Europeans. These results provide support for the theory of a steppe origin of at least some of the Indo-European languages of Europe.


Wolfgang Haak et al., Massive migration from the steppe is a source for Indo-European languages in Europe, bioRxiv, Posted February 10, 2015, doi: http://dx.doi.org/10.1101/013433

438 comments:

«Oldest   ‹Older   201 – 400 of 438   Newer›   Newest»
Davidski said...

postneo,

The R1 Karelian and Samara forgers are about 40-50% ANE.

So the "farmers' wives" would have brought Near Eastern ancestry to the steppe.

Grey said...

I thought it would be R1b west and R1a east but it seems maybe more south and north instead but otherwise similar - with either R1a pushing R1b south and west or R1b migrating south and west creating a vaccuum.

.

"So ANE came to Europe via "farmer wives" then."

mountain mamas.

Alberto said...

@David

"The R1 Karelian and Samara forgers are about 40-50% ANE"

I still didn't have time to look at all the papers in detail, where are those numbers?

Just looking at figure 2 on the main PDF it looks like the most ANE are 4 Yamnaya samples, which would suggest that the Armenian-like population had higher ANE than the EHG.

Alberto said...

Actually looking now at the K=16 graph below Figure 2 it looks like Karelia and Samara had really low ANE. But Yamnaya about 50%.

Cpk said...

Did the IE language come from Armenian like Middle Eastern population or hunter gatherer like Eastern European population?

Davidski said...

Alberto,

The K=16 graph doesn't show ANE admixture. Go to page 84.

"All three of these 2-way mixture models arrive at a similar inference of 38-40% ANE and 60-62% WHG ancestry in Karelia-HG, which is higher than the ~20% ANE ancestry inferred for Motala12 in a previous study4, consistent with the fact that EHG are the population sharing more alleles with “Ancient North Eurasians” (SI 7) than any other."

Chad Rohlfsen said...

Damn!! I was about spot in for EHG. Must be close to Yamnaya too. I can't wait for the genomes!

Krefter said...

Davidski, the R1b1 sample from Spain is about as old as the R1b1 sample from Samara Russia. R1b and R1a could have existed in west Asia at that time. I don't see why you're excluding that possibility.

It does seem likely the R1b and R1a of Yamna, Corded ware, and Andronovo are EHG-descended, but who knows.

Ebizur said...

Simon_W wrote,

"Projection, you mean the PCA? I don't deny that Yamnaya had a lot of West Asian ancestry, but it was from the maternal side. The males all had R1b-P297 derived yDNA, and the latter was present in an EHG from Samara. So paternally they were EHG descended. This means they displaced the West Asian males. And patriarchalic as they were, did they take language lessons with their West Asian women? Rather not..."

Where have you obtained this information that the Samara hunter-gatherer belongs to R1b1a-P297? According to the preprint, I0124 (Samara_HG) belongs to R1b1-L278(xR1b1a1-M478, R1b1a2-M269). This is essentially the same designation as I0410 (Spain_EN): R1b1-M415(xR1b1a1-M478, R1b1a2-PF6399&L265&M269, R1b1c2-V35, R1b1c3-V69). Judging solely from the Y-DNA results, there is no reason to assume that the Yamnaya R1b and modern Western European R1b is descended from hunter-gatherers of Samara rather than from Early Neolithic people of Spain.

Krefter said...

Here's something important for everyone to remeber.

Now we have dozens of Y DNA samples from the steppe ranging the Mesolithic to the Iron age. All are R1a or R1b, except for a few that may be from foreign admixture. Add to that all Bell Beaker and Corded ware samples are R1b and R1a.

Also, Don't forget about MA-1. Also, Q in native Americans.

If there was Y DNA in Yamna besides R1b, R1a is the best bet.

Krefter said...

Geneticker, still thinks the Kurgan theory is baseless.

https://genetiker.wordpress.com/2015/02/11/the-kurgan-hypothesis-is-dead/

Chad Rohlfsen said...

Ebizur,
Everything under l278 can't magically appear in two places at once! Western European R1b is Steppe derived... Bell beaker makes that clear, as if the p297 and l23 didn't... This is only Samara. We're missing a great deal of the Yamnaya area. L51 will dominate another region. Guarantee it!!

Davidski said...

Krefter,

I don't find the idea of R1 migrations from the Near East to Eastern Europe very parsimonious, because Eastern Europe already had plenty of R1 at least since the Mesolithic, while ANE, which is obviously linked to R1, was probably missing from the Near East until the late Neolithic.

Sure, almost anything is possible. But for the time being, it looks like arguing that up is down and down is up.

Davidski said...

Also, Genetiker is a crackpot, so who cares what he thinks?

Mike Thomas said...

@ Skilur

"I think that yamna was also rich in r1a or r1a males killed almost all r1b males there because r1b is just not really frequent there anymore but ancient populations were not the same like modern populations."

Stop talking out of your ass, as usual.

@ everyone "Farmer's wives"

My advice to everyone - stop watching reality TV dating shows and stop talking BS. It's really quite comical

Davidski said...

Mike,

The Near Eastern mtDNA in the Yamnaya obviously does come from farmers' wives.

But the farmers appear to be missing, because R1b was already in the Samara before anyone from the Near East set foot there.

You clearly don't like this, but it's what the data show.

Matt said...

Obviously, for uniparental, R1b and R1a in Mesolithic Russians is the big story here.

For the rest looking at p25, all 3 out of 3 Unetice males were I2? Is that probable or improbable for the location?

On uniparental the German Beaker samples are all women, surprising.

Lots of I2 in early Neolithic and Middle Neolithic Spain as well. Seems like patrilineage groups dominated regions, even before IE, and I2 was *the* haplogroup by the Middle Neolithic (although there are those weird C6 [La Brana lineage] guys in Hungary as well), after an initial T1 and G2 Early Neolithic.

I'm not really a y-dna person, so what do you guys think of the "R*?" haplo in the Middle Neolithic Baalberge sample? Steppe runoff, or something "SHG" (which may indirectly be steppe runoff)? Does "R*?" indicate an outgroup to R1a / R1b (that did not contribute much to modern Europeans) or that status cannot clearly be resolved?

For y-dna, it looks like I2 and R1 are fairly well explained in West Eurasia no, and the major groups still needing an explanation are J2 and E3. J2 shows up by Gamba's late Bronze Age sample (1,110–1,270), not before so far.

From their presence in the Caucasus and LBK, and the possible survival of T among East Africans, it looks like G (G2) and T (T1) are the early Neolithic groups, and then J2 and E3 may have swept later. That seems like the crucial piece left in the puzzle, now it seems that R1 spread with Indo-European, certainly off the steppe and maybe also from West-Central Asia (we'll have to wait and see on that one). (E.g. Armenians seem be substantially R1 and also E3 and J2, so it would be interesting to know if it seemed likely that expansions of J2 and E3 had changed their balance). Anything else will just be from Uralic or Turkic dominant groups at the margins.

Chad Rohlfsen said...

It's interesting as you get up to K-20 that Loschbour starts resembling EHG more than LaBrana. Scratching my head on that one.

Chad Rohlfsen said...

Matt,
I has a couple hot spots in Germany. Not a big surprise with the hunter bounce back. One German MNE is 15% more WHG than Gok2.... maybe more.

Balaji said...

Pastoralism is a skill that is passed on from father to son by a long process of apprenticeship. This is even more true for metal working. Therefore it is highly likely that the Yamnaya derive most of their paternal ancestry from the "Near East". It is not the "farmer wives" who brought "Near Eastern" ancestry to the Yamnaya. Along with these skills, the "Near Easterners" would also have transmitted their language.

Chad Rohlfsen said...

I2 that is, sorry.

Chad Rohlfsen said...

Balaji,
R1a and R1b are EHG and native to the steppes.

PersonaMan said...

Perhaps I2 increassed in frequency with the resurgence of WHG ancestry?

Shaikorth said...

The EHG fit as a WHG-ANE mixture is actually a marginal failure, but the difference to being fit is small so they included it.

"No models could accommodate Karelia_HG as a simple (not
admixed) population (|Z|>9), or a 2-way mixture (|Z|>3). However, three models failed only
marginally (3.1<|Z|<3.7) and we show one of them in Fig. S8.6."

These models gave the 60-40 fits.

One issue appears to be that EHG is symmetrically related to MA-1 and Karitiana. All modern Europeans and of course WHG/SHG are more related to MA-1 than Karitiana including far NE-Europeans (probably not Chuvash though).

Grey said...

@Mike

take me home
country roads
...

.

@Balaji

"passed on from father to son by a long process of apprenticeship. This is even more true for metal working. Therefore it is highly likely that the Yamnaya derive most of their paternal ancestry from the "Near East"."

Or the direction arrows are pointing the wrong way.

.

"It's interesting as you get up to K-20 that Loschbour starts resembling EHG more than LaBrana. Scratching my head on that one."

Could more than one source of (slightly different) ANE ancestry explain it - like an eastern branch and a western?

Grey said...

@personaman

"Perhaps I2 increassed in frequency with the resurgence of WHG ancestry?"

That would be my guess.

postneo said...

So these "farmer wives" had zero ANE and the EHG males replaced the farmers. That means yamnaya is 38/2 = 19% ANE.

later these horse riding farmer wives potentially mated with Iberian R1b further diluting ANE.

The Rip wan winkle Karelian R1a woke up and came in at the tail of the riders to eastern Europe.

Grey said...

@postneo

"So these "farmer wives" had zero ANE"

I must have misunderstood then as I thought they had a lot?

Matt said...

@personaman: "Perhaps I2 increassed in frequency with the resurgence of WHG ancestry?"

Yes, I think you may be on to something here. It's present in one of the early Neolithic Iberians as well though, at similar time depth to the early LBK in Germany. Like R1b in the early Iberian Neolithic its another break to the pattern of y dna from Germany.

@CR: I has a couple hot spots in Germany. Not a big surprise with the hunter bounce back. One German MNE is 15% more WHG than Gok2.... maybe more.

You mean in the modern day? I2 seems all over the place in the Middle Neolithic (Germany and Spain and also there in the 4,360–4,490 NE7 Lengyel sample in Gamba et al, on the cusp of the Middle Neolithic). But with more of a range of samples we might find out more.

Any thoughts on the Baalberge Middle Neolithic with "R*?" what that may mean? The Baalberge samples also seem slightly displaced towards the east compared to the other MN Germans.

CR: One German MNE is 15% more WHG than Gok2.... maybe more.

That's surprising, where's that? On the PCA and ADMIXTURE it still looks like the Skoglund MN farmer is the top MN for HG ancestry. On the PCA there are other MN samples, which hail from Spain, comparably "north" but further "west", which displaces them further away from HGs, meaning less relatedness to hunter gatherers (as well as less relatedness to Near East, etc.).

CR: It's interesting as you get up to K-20 that Loschbour starts resembling EHG more than LaBrana. Scratching my head on that one.

Yeah, I wonder about that too. La Brana, the Hungarian HG and the SHG all form a neat cluster while Loschbour and the EHG are split across clusters. A matter of genome quality?

Skilur said...

@ Mike Thomas
Calm down, nomadic steppe scythians there were r1a and r1b is today not frequent anywhere in the steppe so either r1b was replaced by r1a or r1b was not dominant there. Indo-iranians lacked r1b and they were the longest indoeuropean population who stayed in the steppe

Shaikorth said...


"Yeah, I wonder about that too. La Brana, the Hungarian HG and the SHG all form a neat cluster while Loschbour and the EHG are split across clusters. A matter of genome quality?"

Certainly, those genomes show elevated homozygosity because they are of lower read depth than Loschbour. Homogenous admixed populations form clusters that make them appear unadmixed. As the most extreme example, Karitiana who are definitely admixed appear to only have a single component at any given K value.

Davidski said...

postneo,

OMG, that's very funny. Do you feel better now? Obviously the R1a and R1b in the Eastern Euro foragers was quite a shock. Maybe you should lie down for while?

The farmers' wives were probably already ANE admixed by the time they made it to the Samara Valley. A lot of mixture could have taken place around the lower Don.

And R1a arrived in Central Europe in the vanguard of the steppe expansion, around 4,600 years ago. The ancient DNA shows this.

postneo said...

Well the near eastern origin females (60-70% of the female population) should not have had ANE as per David?

but native pre yamnaya females presumably had ANE at 40 % as well. which means expected ANE in yamnaya becomes slightly higher ~ 25 %

Matt said...

Shaikorth: Certainly, those genomes show elevated homozygosity because they are of lower read depth than Loschbour. Homogenous admixed populations form clusters that make them appear unadmixed.

The logic seems plausible. Though shouldn't the low read depth make them all homogenous in different ways, so it would then be incongruous why they'd all join the same cluster?

Chad Rohlfsen said...

Matt,
It's in a table in there.. I'll find it.. a k value one..

Those trees with Loschbour being admixed and Kostenki make my head spin.. They have Loschbour as a mix of something EHG like and other West Eurasian.

Marnie said...

"The farmers' wives were probably already ANE admixed by the time they made it to the Samara Valley. A lot of mixture could have taken place around the lower Don."

This is garbage.

Sorry guys, but at some point, the whole process of group think degenerates to the point where one bright, objective person would do better.

Very few of you seem to be objective.

Shaikorth said...

It should indeed and lower K's, never mind PCA's and other methods do show these samples aren't identical. We'd need high quality sequences of these samples to truly know how much and to what direction low read depth changed things in this case.

Low depth or not, formal testing can differentiate La Braña from SHG's easily so the fault likely lies with the program. ADMIXTURE has a tendency to form uniformative clusters even from modern samples.

Chad Rohlfsen said...

Okay, pg 23
Okay, it's not as much as I thought. They're about 40%WHG with their model. So about 5% more than Gok2, maybe??

It's hard to tell, until David can run them. Considering the WHG part of Yamnaya, they could still be as much WHG as Estonians.

Mike Thomas said...

@ Davidski

"The Near Eastern mtDNA in the Yamnaya obviously does come from farmers' wives.

* The mtDNA clearly does come from more southern women who arrived sometime 6-5 KYA. Unlike you, however, i'm not equipped with a crystal ball and a penchant for tall-tales, so I cannot state if in fact these women definitely came from communities which were farmers, or indeed if they arrived further north due to being married off.

Only further studies incorporating isotopes and genealogy -specific tests will refute or confirm your scenario. Until then, it is a blow ass theory.

To make such as theory hold, you'd need massive institutionalized practice of marching south to acquire women every year - a highly unlikely scenario given the scale of it, and the fact that these Yamnaya communities were each independent chiefdoms who most likely did not share one and the same universal laws and decrees banning local marriage !


"But the farmers appear to be missing, because R1b was already in the Samara before anyone from the Near East set foot there."

David, do you realise that this is a non-sequitur statement ?

Unlike some of the other posters here (eg @ Grey), I surely would have presumed that you at least have an IQ over 100. Just becuase there was R1b in the Mesolithic Samara region (n=1), and indeed beyond, and likely earlier, it doesn't mean that there wasn't ongoing admixture of new but similar R1b derived males from the south. Sounds like yur just trying to hammer the evidence into your pre-conceived conclusion rather than analysing it for its own merit.

And David, for the umpteenth time - look at a Map ! There is a difference between the Levant- Sinai region and central Asian- Caucasian highlands. !!

"You clearly don't like this, but it's what the data show."

Why should I care who the 3000 BC eastern - Yamnaya dwellers married ? What I don;t like is the talking out of one's ass approach to theory people here invoke without strong evidence for it. Especially given that, at least to me, the evidence shows that those buried in Samara appear genetically related to those further south, if not partially derived from. In fact, as someone said on page 1 here, the pattern reflective here is one that has been recognized time and again in pastoralist and herding communities - seasonal latitudinal migrations north to south and vice-cersa in winter/ summer cycles, accreted over time. Ie the Samara region and the nearby south were in fact, one and the same genetic-ecotone.

So on the basis of this evidence, who's to say who was marrying who.

Perhaps it is you who is uncomfortable with the idea that your far eastern European ancestors descend from people of Asian origin, and are exonorating the genetic ingusion by passing it off as tall, green-eyed white warriors acquiring Kardashian-esque trophy wives :)

Grey said...

"To make such as theory hold, you'd need massive institutionalized practice of marching south to acquire women every year"

Not necessarily. Say the steppe was in three concentric circles and mutual patrilocal bride-swapping then the southern women would come from the southern border into the outer circle and then in the next generation some of their daughters pass into the middle circle and then in the next generation some of their daughters pass into the inner circle.

Say it's 10%

g for generation

g1)
10% outer circle

g2)
20% outer circle
1% middle circle

g3)
30% outer circle
3% middle circle
0.1% inner circle

g4)
40% outer circle
6% middle circle
0.4% inner circle

g5
50% outer circle
10% middle circle
1% inner circle

g6)
50% outer circle
15% middle circle
1.5% inner circle

etc

not exact (obviously) but you can see how it could work

if it was mutual then steppe mtdna would also be increasing in the other population

if it was raiding then it might be mostly one way

either way it the trading or raiding only has to happen at the border not everywhere

Krefter said...

I just started looking at the soul of this paper(autosomal analysis), and Yamna and CWC are much less ENF than I expected.

Yamna and CWC are about a northern as Balts in fig 2a's PCA. Also, Bronze age and Late Neolithic Germans weren't uniform at all. They cover various locations in northern Europe.

Matt said...

@Chad: Thanks. We're talking about the Esperstedt_MN sample aren't we? I'm not sure I'd be comfortable saying that sample has more WHG than the Gokhem sample from that table, as the Gokhem sample isn't included and they've been tested under different conditions. The PCA plot and admixture doesn't recapitulate that greater WHG relationship for Esperstedt_MN.

On that table generally, I think some of the order for modern populations maybe to be due to slight swings in EHG (and thus Yamnaya) affinity via SHG admixture - e.g. Norwegians vs Lithuanians or English vs Orcadian. Some of the relationships again seem strange compared to the PCA again - seems you wouldn't expect Greek vs Spanish to cluster where they do on PCA with those levels, Bulgarians vs Spanish North, Bergamo vs Basque, etc.? It doesn't really seem to obey much geographical pattern. (No reason a priori why it should - genetic integration of HGs and Yamnaya might not obey a geographic pattern, but primary genetic differentiation in Europe seems to under other methods so that's strange.)

That table's based on "We estimate mixture proportions using a method that gives unbiased estimates even without an accurate model for the relationships between the test populations and the outgroup populations (SI9). Population samples are grouped according to chronology (ancient) and Yamnaya ancestry (present-day humans).". Looking at Supplementary Info 9, it's based on some pretty serious and new f4 statistics functions, using measurements of residuals to model fit (this is the resnorm function?), combined with cluster analysis, which I don't understand at all ATM.

@Shaikorth: Yeah, ADMIXTURE's pretty great at forming clusters that explain population level variance and differentiation. Whether that reflects actual population admixture though.... is subject to the problems you mention.

Shaikorth said...

"On that table generally, I think some of the order for modern populations maybe to be due to slight swings in EHG (and thus Yamnaya) affinity via SHG admixture - e.g. Norwegians vs Lithuanians or English vs Orcadian. Some of the relationships again seem strange compared to the PCA again - seems you wouldn't expect Greek vs Spanish to cluster where they do on PCA with those levels, Bulgarians vs Spanish North, Bergamo vs Basque, etc.? It doesn't really seem to obey much geographical pattern. (No reason a priori why it should - genetic integration of HGs and Yamnaya might not obey a geographic pattern, but primary genetic differentiation in Europe seems to under other methods so that's strange.)"

It's quite likely not just those three ancestries that affect positioning. Look at figures S9.24-S9.27. Adding certain complexity beyond just three ancient populations into the model improves the fit (reduces residuals) to some degree for almost everyone, not just outliers, and the Spanish/Greek issue for instance becomes quite clear.

I don't think the fits including modern populations should be taken too literally though, as it seems that their dataset doesn't have optimal proxies. I wrote a bit about that here:

http://www.anthrogenica.com/showthread.php?3807-David-Reich-lecture-9-February-2015/page48&p=68587#post68587

Alberto said...

Yes, the K=16 graph doesn't show the ANE admixture, but the usually called West Asian admixture, that was missing in WHG, EEF, and almost completely absent in EHG too.

Yamnaya has 50% West Asian admixture.

For the authors it's all about damage control. Jokes about stolen wives and lost hunters are not well understood in academic papers.

Grey said...

Actually in the system outlined above the proportion of mtdna might tell you if it was mutual of not.

If it was mutual bride swapping at the border then the daughters of steppe women who'd gone south coming back north might balance the daughters of southern women who'd gone north going back south so it might hit an equilibrium at around 50%.

If the steppe mtdna went above 50% (at least at the border) then maybe it wasn't mutual.

Just a thought.

Matt said...

@ Chad, just one more (final) comment on the MN_Esperstedt sample, it turns out looking at the supplement that this was a Bernburg Culture sample.

From the previous Brandt mega-analysis of mtdna this was the culture where there was the strong change in mtdna frequencies from previous Middle Neolithic cultures to a more HG profile. See here where it's marked BEC - http://4.bp.blogspot.com/-t_KXpJ8f80w/Ulbx5NVzpmI/AAAAAAAAJLE/4ygZg4byqIc/s1600/timeline.jpg. So that could support higher HG in this sample.

Shaikorth: It's quite likely not just those three ancestries that affect positioning. Look at figures S9.24-S9.27.

Those figures seem to be about adding additional West Eurasian, world and ancient populations to improve residual fit. I think that's quite possible, as I said re SHG ancestry above, thanks for the pointer to the specific figures.

saman sistani said...

How can R1b originate in mesolithic Europe and still make it to the Southern Sahara in a couple millennium?

Tomenable said...

It is disappointing that the Janisławice Culture and the Globular Amphora Culture were not covered by this study.

Grey said...

@Saman Sistani

"How can R1b originate in mesolithic Europe and still make it to the Southern Sahara in a couple millennium?"

Maybe it didn't but if it did it seems unlikely to have been massive tribal invasions or someone probably would have noticed.

So if it did it was more likely small groups.

If it was small groups why weren't they killed along the way?

Maybe because they were useful.

Like copper smiths for example.

Mike Thomas said...

@ ALberto

"For the authors it's all about damage control. Jokes about stolen wives and lost hunters are not well understood in academic papers."


Very true. The one earlier "an EHG chased a mammoth all the way to Spain and got lost" takes the cake. I feel sorry.

Mike Thomas said...

@ Grey

"Maybe it didn't but if it did it seems unlikely to have been massive tribal invasions or someone probably would have noticed."

Yes maybe Entertainment Tonight would have done an exposé on it, in 7000 BC.

Mike Thomas said...

@ TOmenable

"It is disappointing that the Janisławice Culture and the Globular Amphora Culture were not covered by this study."


Im sure the Polish palaeoarchaeologists are onto it. I suspect they'll come up at least with some R1a

Marnie said...

If the paper could have just stuck to some of the central findings, like focusing on the population process in Germany, where most of the samples are from, or discussing the Samara samples in their local context on the Steppe, or even examing the signifance of R1b in Neolithic Iberia, then the authors probably could have gotten three papers out of this data.

But the whole Neolithic mass migration thing from Samara, and in particular, the chariots (remember last December?), isn't supported.

Regarding the above comment on "mutual bride swapping at the border" . . .

http://www.scientificamerican.com/article/living-in-an-imaginary-world/

Chad Rohlfsen said...

Because, it looks like some L278 EHG's ancestral to V88 made it to West Asia before 5000BCE. Not a big surprise. That Iberian is just off V88. Not even close to European m269, and not ancestral to us.

Chad Rohlfsen said...

Marnie, that L278 in Iberia is not important. You have no understanding of the R1b tree. L278 could be an EHG group that migrated into West Asia before 5000BCE. That is ancestral to V88. You can't have two populations, one in Samara, and another in Spain, having the same SNP's develop at the same time. It is impossible. Go study how this works and come back.

postneo said...

@davidski
"OMG, that's very funny. Do you feel better now? Obviously the R1a and R1b in the Eastern Euro foragers was quite a shock. Maybe you should lie down for while?"

well I proffered a straw .. you took it and gave it flesh and blood, called it "farmers wife" perhaps in jest but looks like in earnest. You should test the hypothesis and see if it works.

As for the foragers why should it be shocking. I have explained before why such finds have less significance than you think even before this paper was out.

Marnie said...

@Chad

"Not even close to European m269, and not ancestral to us."

How would you know?

Chad Rohlfsen said...

Western European R1b m269 is from Yamnaya. Bell Beaker genomes and the Yamnaya clades nailed it shut.

Chad Rohlfsen said...

R - 22
R.1 (Mal'ta Boy) - 4
R1 - 44
R1b - 2
R1b1 - 2
R1b1.1-L389 - 2
R1b1a - 2
R1b1a2 - 7
R1b1a2a - L23 - 3
R1b1a2a1 - L51 - 4
R1b1a2a1a - L11 - 7
R1b1a2a1a2 - P312 - 2
R1b1a2a1a2c - L21 - 4

Here Marnie. That Iberian is 11 snps short of L23 in Yamnaya. It is what was known as P25. Ancestral to V88. Way too far back to be ancestral to Europe. You don't have 11snps happening in two populations separated by a continent... sorry.

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

Raises some interesting possibilities for V88 though.

ZeGrammarNazi said...

@ Marnie

We have two R1b1* individuals, one from a Samara hunter-gatherer(~5600 BCE) and one from an early neolithic Iberian farmer (~5100 BCE), and then we have an R1b1a individual (~3100 BCE) and R1b1a2a individual (~3000 BCE) from Yamnaya, and that makes the suggestion that the majority of modern western European R1b arose from this lone Iberian subclade unlikely. Unless somehow these western R1b1* descendants managed to mutate the same markers as their eastern kin from R1b1 to R1b1a2a before developing divergent markers.

Either that, or shortly after this lone R1b1* migrated from Samara, the Near East or where ever to Iberia, there was then a back flow of R1b men east, into the Samara region to account for the presence of R1b1a2a in Yamnaya remains, followed by another migration to western Europe to account for R1b1ba2a1 being dominant there.

Marnie said...

@Chad

"Marnie, that L278 in Iberia is not important. You have no understanding of the R1b tree."

Chad, darling, the R1b tree is easy peasy. My ten-year-old daughter could both understand it, the all sacred "R1b tree", as well as autosomal inheritance, and just because there isn't much V88 in Western Europe, She could probably give a slide presentation on Mr. R1b-Neolithic-Iberia's possible contribution to the ancestry of Western Europeans.

I think you've completely lost your marbles with this whole y-dna thing.

Personally, I couldn't give a rat's ass what haplogroups my family shares. None of us have ever been tested. It's not just me. When I asked my husband what he thought about the whole y-dna thing, he responded, using an old Star Trek episode to illustrate:

He said "As long as I'm from Earth, it's OK."

Chad Rohlfsen said...

Well, you obviously don't understand it. Everyone can see that now.

Mike Thomas said...

Chad,

The Yamnaya samples are mostly M269+ and L23+

So why do the authors warn that they are not the ancestors of modern Europeans ?

Mike Thomas said...

Marnie- please explain your position as to R1b and its relevance here

Chad Rohlfsen said...

Because there wasn't an L51. But there is L23 that is one step from L51, there. That is the only caution. But the fact that the Iberian is way further back in the tree means that he can't be ancestral to us. That is all. They stated that testing all of Yamnaya and not just the Samara valley would likely tell a better story. P297, L23, and z2103 are there. That's L51's great grandpa, father, and brother. It will be there. That is a certainty. Samara is only 25% of Yamnaya.

Marnie said...

@Davidski

The all sacred R1b y-dna tree is on my blog, Cowboy.

http://linearpopulationmodel.blogspot.com/2015/01/r1b-l23xm412.html

I'm only pointing out that Mr. Iberia could have contributed to the ancestry of Western Europe, but his y-dna could have been lost at some point.

Rather obvious.

So, Chad's comment that this R1b in Iberia is "not ancestral to us" is quite probably WRONG.

Marnie said...

@Mike Thomas

"please explain your position as to R1b and its relevance here "

A lot of y-dna is lost, due to purifying selection, among other things, yet leaves an autosomal signature.

Chad Rohlfsen said...

There are a few in Iberia with P25. I am saying it is not what led to L51. That is from Yamnaya. Maybe it contributed a sliver of aDNA.. whoopity doo... European L51 is from Yamnaya and brought in 50% of our aDNA.

Chad Rohlfsen said...

The Iberian L278 is not the father of our L51 lines. Plain and simple.

Marnie said...

@Chad

"The Iberian L278 is not the father of our L51 lines. Plain and simple"

But your so obsessed with the male "line" that you've completely lost any sense of the bigger picture.

Mike Thomas said...

Chad. YOu might be right. But the fact remains, at present at least, is that Samara-Yamnaya is clearly oriented to western Asia +/- Eastern Europe.

ITs 'brother' clade - L51 - which dominates Europe could, as you say, have been from somewhat more western Yamnaya sites. But maybe L23* was more widespread than the western steppe. Until this excluded, then its still up in the air.

Marnie-
You are right. Just becuase the R1b from Spain 's Y-DNA died off, it doesn;t mean some portion of their autosomes didn't pass on. I think few would dispute this.

Alberto said...

@Mike

Yes, I do feel sorry too. But I think that personal preferences should not let this opportunity pass for some people.

When you think about it, this is HUGE.

Just think that Samara is in the north eastern border of Yamnaya with nowhere. Presumably the place where most HG ancestry was kept. Once they sample core areas of Yamnaya the West Asian admixture might rise to 60% or 70%. In Maykop it will be higher (and results are on the way).

What does this mean? That cultures that Maykop and Yamna didn't come out of no where. They where founded by newcomers who "replaced" (as the authors like to say) the previous population.

And this means that the number 1 candidate for decades for being the PIE homeland is now the only one that can be scratched. Because this culture was founded by recent migrants who came from "somewhere else" (Asia for sure, probably some now deserted place around Iran or Turkmenistan).

Is it too big to not take it.

PersonaMan said...

@Marnie

That early and middle Neolithic Europeans (and Iberians) contributed ancestry to modern Europeans is self-evident, that's not the issue. The point is that the R1b we find so rampant across western Europe today is extremely unlikely to have it's origins in the kin of the early Neolithic Iberian R1b1.

While it's important to consider all possibilities, people should invariably try to find a solution that explains as much of the information at hand as possible until absolute proof of anything is gained, something which the Iberian R1b1 being ancestral to modern western European R1b-M269 seems very unlikely to do.

Grey said...

@Marnie

the admixture on the steppe labeled "near eastern"

option 1) female mediated

option 2) hermaphodites

option 3) the dna tested so far has big gaps and the y dna is missing

Unless I'm missing something the data so far implies 1 or 3.

Chad Rohlfsen said...

There's no L23, or anything close, in 70 Neolithic Europeans. I've stated that L51 was in the west, and was probably dominant in Cotafeni. That looks like a stronger possibility now.

Marnie said...

And I'm not just making a technical point.

Culturally, there are a lot of men who don't have male descendants, yet contribute both genetically and culturally to the historical record:

-Men who are killed in battle before having children
-Men in the priestly class who don't have children
-Men who have daughters, but not sons
-Men who have sons who don't have sons

That's a lot of men.

To get anywhere, men have to cooperate.

So maybe this wave of R1b Iberians was important to later waves of R1b men who followed.

Very conjectural, but I'm making the point generally that just chasing around after surviving y-dna haplogroups probably doesn't give a very clear picture of what happened in history.

Mike Thomas said...

@ Alberto

Certainly, that was my take. The seminal paper which was supposed to PROVE the Kurgan hypothesis appears to disprove it, yet the authors still claims it "proves" it.

Raises some serious questions.

PersonaMan said...

But how would they themselves perceive a link between their peoples? The Neolithic Iberian R1b descendants and the steppic R1b descendants would have been foreign to each other in every respect, and would not have knowledge of their distant yet common descent.

Mike Thomas said...

@ Chad. I think we might find the western Euro branch of R1b in the Balkans, as you sketch out. But this ancient DNA thing keeps throwing curve balls !

But if we do find R1b-L23 in The Balkans, and earlier than 3200 BC, then it ain't the Kurgan hypothesis. Just ain't

Mike Thomas said...

PErsonaman

They wouldn't of course. People did have access to ftDNA back then

Which is why I called Skilur's explanation (that R1a men later killed their R1b cousins) idiotic.

Marnie said...


"But how would they themselves perceive a link between their peoples? The Neolithic Iberian R1b descendants and the steppic R1b descendants would have been foreign to each other in every respect, and would not have knowledge of their distant yet common descent."

I am increasingly of the opinion that even in the Early Neolithic, there was an elite class that was navigating across the Mediterranean and up and down the Atlantic Coast.

If this is true, simply knowing that a distant place existed, having the knowledge about how to get there, and what was there, would have faciliated movement and cultural development.

Chad Rohlfsen said...

It doesn't all have to be kurgan. Obviously we have an R1b Mesolithic that doesn't have a kurgan. Some EHG R1b with some Balkan Neolithic in Ukraine could be coming into Cernavoda. There are Pontic like burials there, as Varna fades away. Cernavoda is part of the ancestral composition to Cotafeni. It could be that L51 shows up in late Cernavoda, Kemi Oba, and Cotafeni. Or, we could have no R1b until Kemi Oba Kurgans show up west of the Carpathians. We will have to wait and see.

Mike Thomas said...

@ Chad

Varna did not fade away. It flourished and ended suddenly, within 1 - 2 generations. The "pontic-liek" burials in Varna are actually "Varna-like burials in the steppe", you see ? (simple chronology)

"It doesn't all have to be kurgan."

INdeed. So, to all (academics and bloggers) - its best to keep analyzing the data independently, and gather more, before making sensationalist Paper Titles, especially given that there is no actual evidence for what is claimed.

Chad Rohlfsen said...

The individual we refer to as ‘Samara hunter-gatherer’ x I0124/SVP44 (5640-5555 calBCE, Beta-392490) is an adult male from grave 1 in a Neolithic-Eneolithic settlement producing artifacts from the Elshanka, Samara, and Repin cultures. The specific site is Lebyazhinka IV, on the Sok River, Samara oblast, Russia. (‘Neolithic’ here refers to the presence of ceramics, not to domesticated animals or plants.) The radiocarbon date of this individual, based on a femur, is centuries before the appearance of domesticated animals in the middle Volga region. Lebyazhinka IV and the neighboring Lebyazhinka V site were occupied seasonally by multiple cultures between 7000-3500 BCE; a few graves were found in the settled areas

There's a decent chance that R1b is involved in that Repin stuff going on.

As far as Varna type burials, that's harder to prove. They lack the richness and the pottery is different. Varna burials aren't all the same either, as far as positioning. It's a bit of cherry picking.

Marnie said...

@Chad

"Cernavodă culture is ca. 4000—3200 BC"

"Coţofeni culture 3,500 to 2500 BC"

Hmm.

Varna may be out, but the Balkans was and is highly stratified.

Mike Thomas said...

Chad ; Varna had richer burials than yamnaya ; despite its earlier age

I wasn't saying 'Varna-yans' invaded the steppe; but that the M4 inhabitants of the westenr steppe incorporated and modelled certain aspects from the 'more developed' societoes in the Balkans and Cucuteni.

If you are going to talk about cherry picking; then you should re-evaluate your entire understanding of the Kurgan hypotheses

Chad Rohlfsen said...

Okay, I can agree with that.

I understand what the Kurgan hypothesis is. I don't agree with all of it(Baalberg, Baden, etc.) All I am saying is that some R1b L23 could've seeped out as that steppe burial type in Cernavoda circa 4200BCE. It doesn't necessarily have to be the one that bore L51, but it is linked to Cotafeni. L51's age is right about the beginning of Cotafeni.

Chad Rohlfsen said...

This whole Cotafeni thing looks more like Kemi Oba than anything else, so maybe that is the place to look. Either way, I see what is ancestral to Europe going on in the west. Obviously it's complicated and involves a lot of Yamnaya ancestry. Samples will resolve it.

Romulus said...

I'm disappointed that there was no pigmentation or other trait analysis in the paper. I thought there would be a section on height but that was only mentioned once and that was not even in the context of an autosomal trait.

We know from the other paper that the Yamnaya had dark hair and eyes. The light skin genes from the farmers, light eye genes from the WHGs, and the blonde hair and blue eyes in 5,000 B.C. NE7, to me are looking like concrete evidence that the physical traits unique to Europeans were established in the Neolithic (with the exclusion of red hair).

Mike Thomas said...

@ Chad

"More samples..."

Agree

Marnie said...

The Orientation of Dolmens in Bulgaria

https://www.academia.edu/3113449/THE_ORIENTATION_OF_DOLMENS_IN_BULGARIA

Marnie said...

The role of the Kokino megalith in the life of the Bronze age agricultural community

https://www.academia.edu/3500046/The_r%C3%B4le_of_the_Kokino_megalith_in_the_life_of_the_Bronze_age_agricultural_community

Mike Thomas said...

Chad
(Given your interest )
It might not be wholly unusual for some aspects of material culture (metal; sceptres, etc) and even people to have come directly to NE Balkans from the nth Caucasus region ; as if bypassing Yamnaya
Certainly, this has been postulated for much later periods (eg the Cimmerian bridles). Granted we cannot equate what happened in 700BC with 4000 BC, but it's nevertheless an interesting analogy


Nirjhar007 said...

@Guys
There were massive Migrations from West Asia-Caucasian Area towards Europe from ~4000 BC which Intensified after 3000 BC .
There Is NO other Explanation than Intrusion from Maykop which was heavily tied with Neolithic-Post-Neolithic SC-Asian-Iranian cultural tradition where we of course find High ANE and all the Subclades of Y-DNA R.
We also interestingly see the South Central Asian type ancestry making its notable appearance in late Neolithic/Bronze Age LN/BN Yamanaya and CWC Culture samples The components that is LACKING in earlier European farmers and hunter-gatherers this only Strengthens the most likely and simplest scenario...
Though i have kept aside the impressive Materialistic evidence....

postneo said...

@mike
"ITs 'brother' clade - L51 - which dominates Europe could, as you say, have been from somewhat more western Yamnaya sites."

Perhaps L51 was the ANE rich near eastern "farmer" himself. Samara is in the NE of Yamnaya may not be representative of whats spread to europe.

Mike Thomas said...

Nirj
U must be celebrating

Nirjhar007 said...

@MT
What about U?.

Chad Rohlfsen said...

I don't understand the celebration. R1a and R1b were already in Samara before the Near Eastern part. If anything goes your way.. it would be Samara hunters going into the Caucasus... then back out with Near Eastern mtdna.. and some adna.. You can't make R1a and R1b out of the Near East, because they are linked to ANE, which is not in the farmers. Plus, those hunters had no basal Eurasian. They were pure HG's.

Chad Rohlfsen said...

Farming was in NE Iran before 6500BCE... those hunters didn't come from Iran at 6000BCE like you said. Coming from East and staying North of the Caspian is the only logical one.

Nirjhar007 said...

Chad, Just Relax it will take time for you to understand;) no hurry....

postneo said...

@chad
"You can't make R1a and R1b out of the Near East, because they are linked to ANE, which is not in the farmers."

And how do you know this ? You understand these farmers are not the first farmers of the levant right?

Chad Rohlfsen said...

Or the paternal lineage of the farmers... J2, G2, R2? was lost as it entered the steppes.. Just as G2a was overtaken by I1 and I2a going north in Europe.

Chad Rohlfsen said...

postneo.. Neolithic Europeans had no ANE, so the original farmers had none.. R1b and R1a hunters had ANE and no farmer ancestry. That is how.

Mike Thomas said...

Mate I'm European and I don't care where IE came from ; if it's centrala asia - that's cool; if it's the urals that's also cool. To me ; someone who initially accepted the kurgan hypothesis as a nube; the more I actually learned about archaeology and linguistics , the more problematic ALL models became . Despite their differences they all worked within the same forced, contrived approaches. I can Under stand that amateurish bloggers wish to be descended from male warriors blah blah; but for reputed academics to be beating a biased and contrived drum to me is criminal .

Now this study, in my mind at least; only proves what I had been saying all along :
1) a clear southern impulse toward the steppe from the Armeno-Iran region
2) presence of R1a in Europe which has nothing to do with the yamnaya and kurgan-related cultures (Mesolithic Karelia)
3) no evidence for yamnaya -> Central Asia

What remains to be clarified with more samples is how far R1a had already existed in Europe and where modern west european R1b came from exactly

Chad Rohlfsen said...

Yes, but putting R1a and R1b as only entries to the steppes with IE from Iran is gone. That was Nirjhar's big claim. That is not the case.

Nirjhar007 said...

Chad R1b and And R1a had different routes....

Nirjhar007 said...

Also J2.

postneo said...

@Chad
"Neolithic Europeans had no ANE so the original farmers had none.."

Sure but I am not talking about them. I am saying that HGs and farmers from Iran did. Thats the source for ANE which spread to the Steppe and also south Asia.

Mike Thomas said...

Yeah. To repeat ad neauseum , we need a DNA from the Balkans, Anatolia and central Asia.Of course more from elsewhere would be great too !

But I also re-iterate, we should not expect to find a nice "genetic trail", and furthermore , conflate that with the spread of a language. Genetically, populations back then were rather mixed as they are (moreseo) today. Language and genetics do not always co-occur and spread in same axes. For us to understand the expansion of IE, we need to explore theoretical sociolinguistics, and stop hanging our hats on R1a and ANE.

postneo said...

I am not talking about IE just ADNA. I hate mixing things up

Mike Thomas said...

I was responding to Chad, and generally, PN. Not directed at your comment.

Maybe Iran - central Asia did have high ANE. But ANE might also have already been present throughout Eurasia without needing further input from Iran. Whatever the case, I think it is clear that many here have been a bit presumptuous and overconfident with what they expected to see.

postneo said...

@chad
"Yes, but putting R1a and R1b as only entries to the steppes with IE from Iran is gone."

forget IE please. Its of little interest when talking about old HGs why how is the former ruled out?

Nirjhar007 said...

@MT
''For us to understand the expansion of IE, we need to explore theoretical sociolinguistics, and stop hanging our hats on R1a and ANE.''
Yes but if something is a fact then it will correlate with multiple fields of studies like if it works from Archaeological POV then it Should also agree with data from Linguistics and also Genetics BUT yes the main basal onus on Ind-European history is indeed linguistics and the evident data obtained from Archaeotexts etc....

Chad Rohlfsen said...

Iran as the source for EHG? what are you talking about? It was in Scandinavia and Siberia too. Why do you think it's all from Iran? MA-1 was in Siberia... an R.. That is completely illogical.

postneo said...

@chad
"Iran as the source for EHG?"
My question is mainly about R1a and R1b

Your answers seems to be because of MA-1.

if MA-1 was the source of all R lineages, then please explain the non existence of R lineages in east of MA-1 and lack of R2 west of MA-1.

As for ANE, it's probably correlated with R not just R1 and perhaps a bunch of other Y and MTDNA as well. this explains high incidence east of Iran. Iran ANE is probably reduces by levantine populations.

postneo said...

@Mike
" I think it is clear that many here have been a bit presumptuous and overconfident with what they expected to see"

No complaints about the paper if thats what you meant. Its the first of its kind. It makes cautionary statements where there are holes e.g. gap btw yamna and CW. Most other publications make sweeping assumptions on scanty data.

Oaie Porc said...

somebody make the yamnaya gedmatch run :D

Mike Thomas said...

No , PN; that wasn't a reference to the paper. I'm was monumental

capra internetensis said...

@Matt

The Baalberge R* sample (German, ~3600 BC) is low coverage, so few Y SNPs could be read. It had the mutations for P1 and for R (but the R might have been post-mortem damage to the DNA), and the ancestral state for R1a1a-M17 and R1b1a2a1a-L151.

So it could be a lot of different things, including Q or R2. It may well be R1b - it isn't L151, but that probably didn't even exist yet anyway.

Colin Welling said...

@david

Mike,

The Near Eastern mtDNA in the Yamnaya obviously does come from farmers' wives.

But the farmers appear to be missing, because R1b was already in the Samara before anyone from the Near East set foot there.

You clearly don't like this, but it's what the data show.


I don't think we should use uniparental dna to say which gender brought near eastern into the samara valley. Sure, its highly likely that the r1b in samaran yamnaya derived from steppe r1b of mesolithic samara and the near eastern like mtdna in yamnaya is surely just that. But this doesn't mean that only women contributed to the armenian like composition of yamnaya.

If a group of g2 men from south of the steppe, moved into the steppe and married local wives, who then had more girls, we could lose the ydna record of the contribution from those near east men. Expand some variety of this senario over many generations and you may get mismatch between how autosomal dna and uniparental dna can relate differently at times for men and women.

Actually, now that I think about it. If girls always moved, and guys only sometimes moved we could get a uniparental autosomal mismatch for just the men. In that case, women move into the steppe typically marrying a native steppe man. Men sometimes move into the steppe marrying a native steppe woman, and other times they settle and marry a fellow mixed steppe migrant. over time, the near eastern ydna gets wiped out even though it contributed to the aut dna.

Mike Thomas said...

Yes, it could have drifted out (something Y DNA is above all most liable to). And/ Or groups south of Samara - in the Armenia region- were already also R1b and ANE rich.

Time will tell.

Volodymyr Lutsyk said...

Interesting things that were not discussed here: 1)Which y-dna haplogroup was brought by the "Armenian" part of Yamnaya?
2)When did the major "Armenian" admixture take place? (My guess will be some time in the 5th millennium BC, given the fact that the older cultures were typically hunter-gatherers' ones ).
3)If Indo-Europeans had a strict patriarchal order and the "Armenian" newcomers brought new innovations (like bronze), how is it possible that r1a and r1b's found in Yamnaya belong to the local hunter-gatherers? Isn't it more natural to surmise that they were brought in by "Armenian" invaders?

Alberto said...

@Mike

"Certainly, that was my take. The seminal paper which was supposed to PROVE the Kurgan hypothesis appears to disprove it, yet the authors still claims it "proves" it. "

Yes. The data is there, but the authors had their hands tied by their previous paper. I guess they couldn't just come out and say their previous work was all wrong, so they avoid stating what the data states by itself.

@Nirj
Do celebrate. For some this will take some time to be accepted, but hopefully soon everyone will move on and look at the new opportunities instead of holding to old ideas.

mickeydodds1 said...

David,

This rather unexpected result from Reich is simply crying out for a carefully crafted series of responses from you.

Alberto said...

What does the Table 2 from the data supplements contain? Isn't it some genomic data?

Another observation from the admixture graphs: The "Armenian-like" population didn't seem to contain any of the Near Eastern Ancestry shown by EEF. It must have been roughly 100% West Asian.

This will translate into a mixture of ANE and ENF in K8, but still it should give some insane amounts of ANE (Kalash or Balochi are something like 65% West Asian).

The other option is that they did carry HG ancestry too, something like 70% West Asian, 30% WHG. And in this case it would mean a higher percentage of replacement of the EHG. So 50% is the lower bound, and in the Samara region.

ryukendo kendow said...

Reading the article very, very slowly, will prob take several days to digest everything.

First:
"Plotting the f4(Test, Ref1; B, C) and f4(Test, Ref2; B, C) statistics to detect admixture has an advantage
over the use of the f3(Test; Ref1, Ref2) statistic in that these statistics are not affected by post-admixture
drift in the admixed population, but rather rely on allele frequency correlations deep in the phylogeny."

SUCH GENIUS!!!!!!
These people are so smart I don't even. Consider that these f-stats are available to everyone, but these researchers keep finding more complex uses for them to extract ever more information.

That part of the SI is so good.

A few small musings, still reading:
@ Shaikorth
Thanks for pointing that out. Why is the east-west dimension on their PCA so distorted compared to David's PCAs?

@ Chad @ Krefter
Their PCA is highly distorted compared to Davidski's best. Usu ANE forces most W Asian variation into the left third of the PCA since most W Eurasians today have so little, but this one has all the eastern samples compressed together, which reduces estimates of dists among them. Probably because they ran so many ancient samples in the same PCA.

The f4 stat estimation is much more robust. Using their est of 52% Iraq Jew + 48% Karelian gives ANE:ENF:WHG at 26:44:30, and using 47% Armenian + 53% Karelian gives 28:38:34, right smack in the middle of my estimates.

But this is prob not important now, as we know the EHG are not simply ANE+WHG, but something altogether themselves.

ryukendo kendow said...

More first observations.

Also this:

"This hints that the Yamnaya population, while
visibly a much better ancestral source for the Corded Ware, may in fact not be the exact admixing
population. This also agrees with the results of N=3 discussed below and presented in Extended Data
Fig. 3, which estimates that the Corded Ware can be modeled as 29.1% Esperstedt, 9.4% Samara_HG,
and 61.5% Yamnaya, which suggests that the population of eastern migrants had a slightly higher
proportion of EHG ancestry in its makeup than the Yamnaya sample from Samara."

BAM. I always said that the CW likely had influence increase from HGs as they moved.

Also, f4 fits for Russian, Finn and mordovian increase greatly when Nganassan are added, but surprisingly also for hungarian and estonian. Those f4 stats are really powerful, they unveil Motala as admixed between EHG and WHG even when f3 stats don't, so the fact that they don't reveal that EHG is mixed between ANE and WHG is quite something. The D-stats that show that symmetrical relationship to NAms lends support to this too. This level of unexpected behaviour wrt NAms was kinda shocking to me.

ryukendo kendow said...



Lastly, no pop in europe is fitted as more than 50% Yamnaya, which is quite surprising coming from the point of pure arithmetic scores for ANE, but really shouldn't be when you understand what it means that EHG is not really a mixture of ANE and WHG, but in some ways something separate altogether. This immediately makes it clear that WHG survival is undeniable in much of Europe, with 30% levels in NE Euro in the fits and high levels accros a swathe even when EHG is present as a proxy for WHG rise through yamnaya influence--should EHG be a mixture of ANE+WHG, which it is not. The same occurs when Yamnaya is used as a sample instead of EHG. In fact there is a pattern of NE.Europe fitting as Yamnaya/EHG+WHG, with WHG peaking in the baltic and EHG peaking more in N Europe in general, and NW Europe fitting as Yamnaya/EHG+EEF. I wonder if we could try these fitted stats with WHG, EHG and Yamnaya together? Would even tell us even more about the source. The program for those 1000s of calcs would be a PITA to write though, hopefully the authors have a program they would release.

The fact that EHG is so high in Norwegian and Czech, and WHG considerably lower, while EHG is at the same level but WHG peaks in Balts, despite EHG being from the east and WHG from the west, makes me think that my conviction that autosomal contributions from the steppe and from Euro HGs can be distinguished, is confirmed. I made a prediction that E.Euro and N.Sea will be high, while Baltic would be low, in the IE samples, testifying to changes in E.Europe reflecting contris of native hunter-gatherers, and this result seems to reflect that a bit. Lets see how ADMIXTURE goes.

EHG would be absolutely incredible to explore in ADMIXTURE and D-stats. I really wonder what would happen to EHG in supervised runs, would the diff btw EHG and WHG finally account for the lack of WHG influence in C+S+W Asia? Maybe its all EHG? Or maybe 'ANE' in S+C+W Asia is closer to Mal'ta than to EHG? Or both? Or maybe europeans turn out to have no ancestry that is closer to Mal'ta than to EHG, but C+S asians do?

I really hope they release their calcs as a program!

Matt said...

RK: @ Shaikorth
Thanks for pointing that out. Why is the east-west dimension on their PCA so distorted compared to David's PCAs?


If you're interested in comparing the PCAs, I made an image document with all of (or most of) the PCA from Lazaridis 2014 and Haak 2015 side by side:

http://i.imgur.com/nRHmBur.jpg

(inverted colour scheme is because it makes it easier for me to distinguish points because there's greater contrast, flip it back in MS Paint or another piece of software if it annoys you).

If you look at my image file, biggest differences to me between plots seem to be that the introduction of modern far Southwest Asians (BedouinB) distorts the positions of all the samples by pushing Sardinians and Basques closer together and East Europeans and Caucasians further apart.

If you compare the plot in the bottom right hand (Laziridis 2014 S.10.6), without BedouinB to the plots with projected ancients and BedouinB included (Laz 1.b, Haak S5.2), the Levant->Caucasus cline is much more parallel to the Med->Russia cline in plots without BedouinB, where the European cline is more scrunched up. An in the projected plots, Basques and Sardinians are almost on top of each other.

This seems to be mitigated somewhat, but not totally, by the introduction of ancient European farmers, hunter gatherers and Yamnaya people.
The introduction of modern Bedouin seems to have a large effect (whether because they're drifted, which they are, or because related population have made a demographic impact, see expansion of Southwest Asian components in their ADMIXTURE).

ryukendo kendow said...

@ Matt
Wow those are eye-popping! Thanks.

In any case, the introduction of EHG as another pole independent from ANE-WHG-ENF probably spells the death of the blissful times when linear dists on pcas could tell us so much. The fact that they aren't in a straight line between ANE and WHG already rings the bell. Its gonna get so much harder.

Hopefully, we could do pcas of WHG and EHG alone, and then project modern-day variation onto that, to see if the f4 stats for europe are corroborated. I'm already concocting methods in my head lol. D-stats for S+C and W asians between EHG and Mal'ta? or plots of f-stats with x Mal'ta and y EHG? Or X EHG and y stuttgart. Hopefully that would show us the autosomal contribs into C+S Asia where Mal'ta vs. Loschbour failed spectacularly.

Davidski said...

Mickey,

I'll have more to say when I get my hands of the ancient genomes.

Alberto,

That's just a list of the SNPs they used. You can unzip it with 7-zip and then open it with something like Text Pad.

rk,

Those ~3,000 BC Yamnaya nomads from the Samara aren't ancestral to Europeans. You can see that in their Y-haplogroups, which are very Asian today.

So we'll only be able to accurately estimate Yamnaya admixture in Europe when we get more relevant samples from the Yamnaya horizon. That's basically what I was told by one of the authors.

Also, as far as I can see the EHG just look like 40/60 ANE/WHG mixtures, albeit with some ENA.

Matt,

After I removed the most SSA admixed BedouinB, as well as a couple of pairs that shared a lot of IBD, they stopped sticking out so much on my PCA.

Shaikorth said...

BedouinB samples don't seem to cause issues in this newest PCA because they don't define the 2nd dimension instead of neolithic farmers. If SSA or inbreeding in BedouinB caused issues, additional neolithic farmers were enough to overcome that.

RK, when it comes to introducing modern populations into their calculation for ancestry proportions we can note a couple of things. Adding Nganasans improved the fit most in the far Northeast, but also most Europeans in to some degree, Norwegians had similar change as Estonians. Adding Bedouins on the other hand showed no change in residuals for everyone except Maltese, Sicilians, Ashkenazi and Spanish, and unlike Nganasan, Bedouin component in most fits remained at 0%. This may indicate that extra Nganasan-like ancestry in small amounts may be widespread while Near Eastern ancestry beyond Yamnaya or Farmer ancestry is certainly restricted to few locales, but one has to be cautious with adding modern samples.

While small Nganasan-like mixture could explain various things like admixture signals for (x; Y, Dai) around Europe, they aren't the perfect proxy for anyone except perhaps Chuvashes. This is because they are admixed themselves (although they form a signature component at high K's) and have different types of eastern ancestry besides possible ANE/EHG mixtures. ADMIXTURE may not be perfectly informative but these results from the run in this paper do hint towards that too, pay attention to the component maximized in Dai:
http://s22.postimg.org/fkvitq0qo/admixt.jpg

Further, even through they are Uralic Samoyedic speakers, their paternal lineage is over 90% N-L666 which diverged from ancestors of N-L1026, almost all East European N, about 10k years ago, thousands of years ago before suggested Proto-Uralic or Proto-Indo-European timeframes.

Nirjhar007 said...

@David
''I'll have more to say when I get my hands of the ancient genomes.''
How long to Wait?

Davidski said...

Probably a few weeks.

Btw, I just noticed that Cochran actually made some useful points about this paper.

https://westhunt.wordpress.com/2015/02/12/origins/

Nirjhar007 said...

@David
Thanks and for that Comical Post as well!

postneo said...

@colin

"If a group of g2 men from south of the steppe"

The thing to realize is that it's just not g2

Just as Africa id the source for archaic humans to the Middle East, In the last 50,000 years The steppe and Europe Seem to lag behind West Asia Sc Asia. Lineages that die out in the source regions survive longer in the northern limits

Let's take for exams y HGs C, later I, G, J, R . It's the same for mtdna and Adna One difference - average mtdna mostly precedes the corresponding y of the source population.

Perhaps some r1 lineages bucked the trend during the last ice age we will Find out.

Matt said...

RK: In any case, the introduction of EHG as another pole independent from ANE-WHG-ENF probably spells the death of the blissful times when linear dists on pcas could tell us so much. The fact that they aren't in a straight line between ANE and WHG already rings the bell.

I think it's going to stay close, although I particularly have my doubts about the value in the future of using ANE rather than whoever it is we find in South-Central Asian dna along with EHG. They'll share some descent from ANE however using the modern admixed ANE descendant groups with additional drift rather than the ancient upper paleolithic population that was surely broken up and mixed in with other population fairly after that should give a better fit.

Not sure if ENF will still have a future as any part of a model long term or if we'll find that there were major migrations in West Asia that reconfigured the genetic landscape there, as it looks like there were major expansions of patrilineages and civilizations there during the late New Stone Age and early Bronze Age.

Shaikorth: RK, when it comes to introducing modern populations into their calculation for ancestry proportions we can note a couple of things. Adding Nganasans improved the fit most in the far Northeast, but also most Europeans in to some degree

Yeah, re: figs S9.26 and S9.27 looks like there is some improvement from adding Nganassan into the mix and then BedouinB, which improves to a lesser extent (BedouinB surely to some degree a proxy for African ancestry mediated by BedouinB as well, see improvement in the Spanish).

The least overall improvement from both Nganassan and BedouinB combined in Northwest Europe (English / Orcadian). Makes sense given these are in most instances going to reflect IBD since the last major migrations into Europe (or IBD with those Europeans who did experience minor migrations), and those would be least in the Northwest?

But then there's some strangeness there as well from that Nganasan signal, the Scots (also Northwestern) have similar improvement to residuals as the Lithuanians, the French South who otherwise look exactly like Basques on other measures have a similar improvement to Ukrainians. Also, a signal from Nganasan (if it actually has a meaning for most Europeans) might be picking up some signal from South Asia / Gedrosia, where people have some ENA ancestry - that's a distinction that would show up contrasting a Pathan or South Indian 4th population for example to a Nganasan.

Table S.9 puts in an interesting contrast. Residuals for North African and Turkish Jews, Cypriots are more improving for the Maltese, Sicilians and Ashkenazis compared to BedouinB. This is probably due to a lack of the African ancestry present in BedouinB, as you can see the candidates get worse residuals in linear order the more African they have. While Chuvash and Turkmen are more additive to Finns, Russians and Mordavians compared to Nganasan. Presumably again a lack of East Siberian affinity in C and T as the residual fits again get worse the more Japanese-like ancestry they would have, West-Central Asian post-Turkic populations fit better.

Matt said...

RK: Those f4 stats are really powerful, they unveil Motala as admixed between EHG and WHG even when f3 stats don't, so the fact that they don't reveal that EHG is mixed between ANE and WHG is quite something. The D-stats that show that symmetrical relationship to NAms lends support to this too. This level of unexpected behaviour wrt NAms was kinda shocking to me.

page 85 - Fig. S8.6 shows Karelia_HG as a 2-way mixture of ANE and WHG, but this may be due to the fact that the model of Fig. S8.1 was developed before the EHG were sampled. We do not want to reify ANE and WHG simply on the basis of their having been described earlier in the literature, so we also explored models in which EHG are not admixed. This was driven, in part, by our observation that the statistic f4(Karelia_HG, MA1; Karitiana, Chimp) = -0.00014 (Z=-0.2) which appears to be show a perfect symmetry in the relationship of (Karelia_HG, MA1) with Native Americans.

Woah, hold up, symmetry to Native Americans, between EHG and "ANE"? That's not what *should* happen if ANE is a partial ancestor of EHG and WHG and ANE have differences in drift shared with ANE.
What would be nice to see if the same symmetry to Native Americans for ANE and EHG is also found for EHG and ANE to Kalash. That would be *interesting* as a beginning of exploration of this topic (and I don't doubt they've already done it in the lab, its just not here).
One of their observations proposed for exploration (I haven't read through these tree models) is first, Karelia_HG share more alleles with Native Americans due to sharing additional common genetic drift (and thus Native Americans should share more alleles with Karelia_HG than with MA1). Surprised by that, as I wouldn't have thought that f3s should be sensitive to how late a population branches off a shared tree with another population? I guess they are.

Davidski: Also, as far as I can see the EHG just look like 40/60 ANE/WHG mixtures, albeit with some ENA.

ENA admixture in EHG would make life simpler given the above. If that's the case, though, then the paper should find that they deviate to some degree in a direction of sharing additional drift with East Asians / Onge compared to WHG or MA1. Is there a stat like that in there?

Arch Hades said...

On the ADMIXTURE analysis, K = 16..the aqua blue component in the Yamna is the Near Eastern component?

I'm just curious why it's so distinct from the early European Neolithic farming component.



"So we'll only be able to accurately estimate Yamnaya admixture in Europe when we get more relevant samples from the Yamnaya horizon. That's basically what I was told by one of the authors"

I think the reference # [9] is low, but what's wrong with these samples?

ryukendo kendow said...

@ Shaikorth
I understand that Nganassan contains large portions of ENA ancestry, which is why i thought it wwas very surprising its inffluence extended into the baltic area, and elsewhere into Europe also as you say.

@ Davidski
I'm not sure about that. The evidence seems to suggest otherwise.

To put the f4 plots in the plainest language for the benefit of more ppl here, first pop A and pop B havee their f4 stats to 15 outgroups calculated, say pop 1, 2, 3, 4, .... 15. So the distances between A and B and the other 15 are all known.

Then the test pop, C, is added. C is a mix of A and B if C has allele freq directly intermediate between A and B. A consequence is, if the distance between A and C increases when measured w.r.t some pop, say pop 1 of the 15, then the distance between B and C decrease, and vice versa. And if C is perfectly between A and B then the correlation is a perfect -1.

So where Yamnaya is similar to Karelian w.r.t the 15 outgroups, Yamnaya is not with Iraqi Jew, and where Yamnaya is similar to Iraqi Jew w.r.t. outgroups, Yamnaya is not with Karelian, and when one shrinks the other grows and vice versa in near perfectly correlated manner, indicating that Yamnaya is pretty close to being a mix of Karelian and Iraqi Jew.

However, where EHG is similar to loschbour w.r.t outgroups, EHG is also similar to Mal'ta w.r.t outgroups, and where EHG is dissimilar to Loschbour, EHG is also dissimilar with Mal'ta.

Also, EHG is equally related with native americans as Mal'ta. And the best-fitting tree shows EHG forming a clade with Mal'ta with no gene flow from WHG to EHG; the higher amt of allele charing between Karelian and Mal'ta are explained by EHG and Mal'ta in a clade, and Mal'ta having some introgression from some basal west Eurasian.

Perhaps the caveat they applied for their f4 stat applies in this case. Which is, there might be too little difference between Loschbour, Malta and their relationships with the outgroups such that they are not really disitnguishable, aka Malta and loschour are too similar w.r.t. the rest of the outgroups. But then why would it work for EHG and WHG? And also, they include Amerindians and Siberians in their outgroup, which would probably be the pop that discriminates strongly between WHG and ANE, so I'm not sure if changing the outgroups would change anything.

In any case, we should see if they have a software package, or we could try modelling on our own. The fact that Amerindian is unable to suss out EHG as admix between ANE and WHG makes me think that nothing will.

ryukendo kendow said...

@ Matt
What they mean is that EHG branches off from Mal'ta to the exclusion of WHG, while Mal'ta has introgression from some outgroup, so EHG ends up closer to WHG despite having more similarity with Mal'ta than WHG does.

Agree that the symmetry w.r.t. the Amerindians is a clincher for this one.

Shaikorth said...

Matt, assuming that ADMIXTURE result is informative, I doubt the "Nganasan" is South Asian-related. Any ENA in Pathans or Kalash seems to be mainly covered by other components than the "red" one, same for Burushos to a greater extent.

Chuvashes and related populations would reduce residuals the most among the samples they have, but if one did a "ENL+WHG+Yamnaya+Chuvash" model I'm quite certain that "Chuvash" would show up all around Europe too, and to a much greater degree than Nganasans because most of their ancestry is European and would cover Yamnaya, WHG and other older affinities. Additionally, in that ADMIXTURE at K=7/8 Chuvashes still have a noticeable Dai-related component like Nganasans which doesn't appear in any other European samples. This kind of distinction also shows up in the Yunusbayev Turkic migration paper's ADMIXTURE run and probably reflects Turkic ancestry. So while they are the best fit, they are not perfect.

There are a few modern populations that might reduce residuals even more than Chuvashes but are not included in the Human Origins dataset, Udmurts come to mind as unlike Chuvashes and Maris they lacked "Turkic" component in the Yunusbayev run and perhaps Saamis as well. But these are also mainly European and probably would show up everywhere hiding older ancestry. The best solution would be ancient DNA. The Bronze Age foragers of Kola Peninsula's Bolshoy Oleni Ostrov site perhaps, at least mtDNA-wise they look more eastern than mesolithic EHG's.

Re. ENA affinities of EHG, they didn't really test it. But as I said before, all Europeans except maybe some Volga-Ural populations like Tatars and Chuvash are more related to MA-1 than to Karitiana so EHG's symmetric relationship may suggest some ENA.

ryukendo kendow said...

@ Shaikorth @ Matt
I don't think ENA exists in EHG, as if it did all the trees with Onge in them would be rejected.

Matt said...

@ Shaikorth assuming that ADMIXTURE result is informative, I doubt the "Nganasan" is South Asian-related.

Assuming the red component is informative its only in the Europeans you'd expect it to be. You kind of have to assume its uninformative to an extent (only limitedly informative) if you're even going further with looking for residuals, at which point why not check out South-Central Asian groups? They didn't because most Europeans except for the exceptions were closely fit, but if you were trying to improve their fit...

@Ryu I don't think ENA exists in EHG, as if it did all the trees with Onge in them would be rejected.

That's my thinking to, however would have been clearer if there had just been a stat to that effect.

Simon_W said...

Two further remarks, I don't know if it has already been mentioned, but: The finding of R1a1 in a Karelian EHG also largely disproves the recently suggested theory that R1a originated in West Asia, as the dude had no West Asian component in the ADMIXTURE analysis.

Secondly, Karsdorf_LN looks curious; a female with dubious cultural affiliation, apparently contemporaneous with Baalberge (C14 dates pending), long before the Corded Ware horizon, carrying the West Asian late Neolithic / EBA haplogroup T1, but autosomally she was predominantly Yamnaya-like, deviating a bit towards WHG, and having almost zero EEF ancestry! What was she doing there in the EEF community, and how did she get there? She was buried inside the eastern gate of a burial complex from the Baalberge culture.

Shaikorth said...

The S-C Asian issue could be tested with a "ENL+WHG+EHG+Pathan" fit and then adding Nganasans, if that causes zero changes in most North Europeans we could say Pathans are a better source. However I don't think that will turn out to be the case. The red component is in most Europeans, EHG has more than anyone except the Chuvash and Saami. In S-C Asians it's not that high, Pathans, Kalash and Czech are about equal. There should be much more if these were to be proximate sources. Burushos have Dai-related extra and Hazaras I expect have even more.

EHG's possible ENA affinities should be also checked with modern Siberians and Dai. Does it have more than can be explained with MA-1?

Gaspar said...

@ Simon

The karsdorf area you mentioned was not a she, it was a male with T1a who had H1 as his MtDna. The other 4 males where ALL G2a2a

Davidski said...

That Karsdorf burial with high eastern admixture is described as "unusual" and the date is unknown.

truth said...

Roy King,

His R1b is basal also to the Levant/African clade V88, and given that Samara HG has exactly the same haplogroup, im more inclined to think he was some kind of dispersed HG that came from the East.

Chad Rohlfsen said...

Matt,
That is pretty interesting stuff, wrt EHG. I think that WHG is going to go bye-bye pretty soon too.
If I'm reading that right, then R1b and R1a were "pure" EHG, but of course, that too could be a mix of a couple UP types. Does it look like they're throwing out WHG and ANE? Kind of like we can throw out Baltic/West Asian, etc, as those are nothing more than mixes of these earlier groups, which are mixtures of earlier groups, and on and on....

I can't wait for some papers on UP phylogeny. It sure is needed. It seems like the more papers we have, the more questions there are to answer.

truth said...

Roy King,

The R1b of this early neolithic is basal also to the V88 clade, considering he has exactly the same haplogroup as the Samara Hunter, we can conclude he was also descendant of Eastern foragers, they probably were very dispersed and minoritary, all the way from Iberia to Siberia.

Chad Rohlfsen said...

They have Oetzi as like 5% Yamnaya.... hmm

Roy King said...

@Truth,
The Spanish R1b EN sample was not tested for V88/PF6279 so it is unknown whether he is derived for V88. We only know he is P25/M415 and not on the M269 and M73 branches and not V35 or V69. He could even be M18, like the Sardinians and the Lebanese.
If he is derived for V88, as I said above, he could have arrived via the Cardial Neolithic from North Africa as well.

Kristiina said...

As a linguist I do not have a mathematical approach to this issue, but as for your discussion on ANE/ENA, my interpreation is that the Carelian hunter-gatherer has some ancient Siberian admixture that it shares with Native Americans, in particular Chipewayan and Pima Indians. It is also noteworthy that his mtDNA is C1 which is a typical Native American halogroup. By contrast, Bolshoy Oleni Ostrov individuals that were not tested could possess more recent ENA which they share with modern Northeast Asians and which in Europe comes from the Butovo-type cultures that I mentioned earlier. It is a pity that we do not know their yDNA. In addition, there is also a more southern ENA which is spread to the West in particular via the Silk Road and the Turkic migrations.

Chad Rohlfsen said...

Roy, he isn't from Africa. We either have an EHG expansion into Europe, which is unlikely at this point, or a very rare EHG that wandered into West Asia early enough to catch one wave of Cardial migrants.

Mike Thomas said...

@ Dave, RK, et al

I understand that you want to analyse the genomes yourself; fair enough.

But lets take this back to a more fundamental level:

1) Do you actually think that the values obtained from the D-statistis, F-3 scores, etc, and the resultin estimates of admixture are actually translatable into a *real* population history . ?

Ie do you actually think if the results suggest the 3/4 take over by a Yamnaya population of German Neolithic groups, this means 3 out of 4 men got replaced ?

2) Are you now more willing to consider my suggestions. I know I'd never analysed the data like you guys can so only go by what you calculate, but nevertheless my explanations that the obvious eastward shift in central European groups during the LN - EB needs other considerations aside from "massive immigration from Yamnaya groups " ? Especially in light of the fact that for this shift we can infer other source populations, from SHG to EHG to Neolithich Highland west Asians (NOT Levantines - the part source populations of EEF); in addition to the need to consider "extra-migration' epiphenomena like selection & drift in what was obvious a population slump during the LN.

3) I've hear Chad's hypotheses about R1b, but what are yours now on the 'trail' of European R1b as well as R1a .

Krefter said...

We have people today who obviously have tinkles of ANE because of their Y DNA(R1b-U152 in Sardinians, from Italian admixture), but come out as 0% ANE in tests.

So, it is defintly possible that's why R1b1* was in Neolithic Spain. But the R(P1+, R+) from Neolithic Germany might be revealing this isn't a fluke?

Krefter said...

Chad,

Ancient Y DNA has thrown around many surprises.

Although Spain is far away from the modern distribution of R1b-V88, that doesn't mean that Neolithic Spaniard could not have had V88. You should be more open-minded about R1s history.

Chad Rohlfsen said...

Krefter, you're not paying attention. I said it's more likely an EHG by ydna was in west Asia and picked up in a farmer wave. If you look at bellbeaker blogger, you'll see I mention the v88 or prev88 possibility.

Roy King said...

@Chad,
R1b-V88 could easily be the marker that characterizes this sample. We need to account how R1b gets to North Africa from, say, Iran. Clearly, IMO, P297 likely originates in Central Asia/far east Russia. But its parent-M343-must have originated between the areas we find V88--Levant/North Africa/Sardinia--and Central Asia. I was arguing that the Cardial sample which autosomally resembles modern Sardinians, may have hopped along the coast from the Levant ultimately into Italy and Spain. I don't understand your counterargument to this proposition,

PersonaMan said...

Is V88 accompanied by any other typically European/West Asian y-lineages in Africa? If so that might tell us something about what 'group' V88 traveled with, if any. Ie if there were a handful of typically European Neolithic markers in V88's hot spots as well, that might be useful?

capra internetensis said...

Alberto:

Samara is not on the border of nowhere. Samara is on the way to the Uralic copper mines.

Fanty said...

I recall that companies like ftDNA used to claim years ago, that 7-10 generations admixing with locals completely exchanges the autosomal DNA to local one.

Of course, that might only hold true if its very few migrants and the genpool of the target population isnt conterminated by the migrants DNA, that flows back to the migrants, but there is only a washing out/exchanging effect with no backflow.

Roy King said...

@PersonaMan,
Yes, there is L91 in Morocco. L91 is the Y marker that characterizes "Iceman". Pre-L91 was found in Haak et al in the LBK context.

Krefter said...

A Russian poster from Eupedia gave R1b-Z2105 percentages in some modern Volga people. His source is Trofimova 2015

"36.2% Burzyan Bashkirs
21.2% Udmurts
8.0% Komi
6.8% Erzya and Moksha
3.8% Besermyan
2.3% Chuvash
0% Mari
0% Kazan Tatars,
0% Bashkirian Tatars,"

If we knew most of the M269 in the Volga was Z2105 before Haak 2015, the Z2105 from Samara Yamna would not be such a big surprise.

Eastern Yamna likely had much more L23 and Z2105 than any modern pops, maybe something like 80%(Like R1b-L11 in Ireland). Consdiering Bell beaker who was ~50% Yamna-like had mostly R1b(3/3) and specifically Z2105's brother L51, I think it's likely L23 itself spread out of the Russian steppe.

Add to this uniform Y DNA R in almost all pre-historic ANE-rich Y DNA samples from North Eurasia, who's to say R1 itself didn't originate in north Eurasia.

R1b-M73, R1b-L23, R1a-S224, R1a-L664, all probably orignate in north Eurasia. R1a1* and R1b1* popping up in the 2 only Mesolithic samples from Russia, isn't random. There's a connection.

Something like 30% of modern west Asians, Europeans, south Asians, and central Asians trace their father lines to the Russian steppe. That's incredible!!!

Chad Rohlfsen said...

rk,
That Basal West Eurasian into MA-1 could back Eske's lab stating some Kostenki is found in MA1. What do you think?

Alberto said...

That article from Greg Cochran I think that shows clearly one of the main confusing ideas that came out with the paper (or the teasers before it): That the population that entered the steppe were Near Eastern Farmers.

They were not. They were Ancient North Eurasians. Genetically they were not related to any basal Eurasian (NE) population. They are much closer to European populations. And y-dna R was native to them.

They might have been somehow "Armenian-like" because they moved to West Asia first, and probably picked up NE admixture there. Armenians have been living for 5000+ years in the Near East, so they've picked a lot more since then.

EHG had a good amount of ANE, either because of contact with these ANE population somewhere around Kazakhstan or because of a common origin. Not clear yet until we get an ancient genome from these ANE population before they started moving.

I wonder why they didn't write clearly in the paper the amount of ANE of the Yamnaya samples. My guess is that some of them might be 50% ANE, if not more.

And a thought: we have a population that from the late Neolithic and during the Bronze Age to Iron Age caused a very significant impact in every population between India and Ireland. During this time, we have a family of languages spreading from India to Ireland. Is there any other possible population that can be a second candidate for being responsible for the spread of these languages throughout these territories during this time frame?

Chad Rohlfsen said...

Krefter,
Are you including v-88 into that?

Nirjhar007 said...

@Krefter
''Something like 30% of modern west Asians, Europeans, south Asians, and central Asians trace their father lines to the Russian steppe. That's incredible!!!''
Incredible indeed! since most of them if not all Came from West Asia! and that is certainly the case of Yamnaya R1b which is mostly post 4000 BC and the H-G R1b and even R1a is of POST West Asian Intrusion period starting from before 6000 BC....

Mike Thomas said...

@ Krefter

"Something like 30% of modern west Asians, Europeans, south Asians, and central Asians trace their father lines to the Russian steppe. That's incredible!!!"


Not remarking on the figures ; but I'm querying the geography. Life in the steppe was too volatile : it's known arcaheologicLly and now we see genetic proof. If say if there was a refugium of R-derived ANE rich people's ; it'd have been further north- the forest steppe and southern forest zone .!

Fanty said...

"EHG had a good amount of ANE, either because of contact with these ANE population somewhere around Kazakhstan or because of a common origin. Not clear yet until we get an ancient genome from these ANE population before they started moving.

I wonder why they didn't write clearly in the paper the amount of ANE of the Yamnaya samples. My guess is that some of them might be 50% ANE, if not more."

Where would 50% ANE come from?

The article claims the Yamna are best decribed as a 50/50 Mix between those 2 EHG, who are like 60%WHG + 40% ANE
And MODERN DAY Armenians.

Do modern day Armenians have 60% ANE to arrive at 50% ANE for Yamna?

Krefter said...

Nirijh,

The EHG samples had no Basal Eurasian, no signs of west Asian-specific ancestry. If lets say the R1 in EHG(remember MA-1 had R*) is from west Asia, L23 and Z93 in Asia is still most likely from the Russian steppe.

The uniformity and diversity of R1 in pre-historic north Eurasia is incredible.

Till Haak 2015 I thought R1a and R1b most likely came out of west Asia, but the evidence isn't supporting this anymore.

R1a1* and R1b1* in two random Mesolithic Russian samples speaks for itself.

Earl Snerd said...

So far R1b1* seems to be oldest in the Near East. This haplogroup had to be at the Dead Sea by 6200 BC in order to diverge into R1b1c-V88, and to be carried into North and Saharan Africa with Chadic people herding sheep and goats. Plus the diversity of R1b1a2 in general dates back to 6000 BC in Anatolia. The diversity results of R1b1a2 places it in the Western Mediterranean at c.4800 BC, which is very close to the 5100 BC in which R1b1* was found in Spain. The genetic study done on modern Iranians shows that most of the R1b lineages are found in the Northern portion of the country. Since R1b1 must have been at the Dead Sea before arriving in Anatolia the R1b people were most likely the earliest pastoralists in the Zagros Mountains. They must have had a significant innovation in order for their Y-haplogroup to have a noticeable presence on three continents.

Earl Snerd said...

By the way I do believe that R1* must have originated on the steppes but that it may have expanded into Iran, and diverged into R1b in Iran either during the early Holocene or the Bolling-Allerod. The R1b in the Yamnaya must have come with the Near Easterners. I think this scenario explains ancient DNA and the modern.

Mike Thomas said...

@ Earl

"must have had a significant innovation in order for their Y-haplogroup to have a noticeable presence on three continents"

I don't think this is necessarily the case. Don't think special properties and propensities are prerequisites . More to do with blind luck


Chad Rohlfsen said...

Earl,
There is so much wrong with that, I don't know where to begin.

Mike Thomas said...

Earl
"By the way I do believe that R1* must have originated on the steppes but that it may have expanded into Iran,..."

I'd prefer to think of Northern - central Iran as part of one central-northern eurasian geo-ecotone. This might help make better explanations instead of the wrong assumptions we've often been making

Earl Snerd said...

@ Mike Thomas

"I'd prefer to think of Northern - central Iran as part of one central-northern eurasian geo-ecotone. This might help make better explanations instead of the wrong assumptions we've often been making."

That's what I was getting at. It's just a similar lineage going around in a circle.

Earl Snerd said...

@ Chad Rohlfsen

"There is so much wrong with that, I don't know where to begin."

Well I put it all together from the latest genetic and archaeological studies. Grugni et al. 2012. Zilhao 2001. Balaresque et al. 2010. Gonzalez et al. 2013. The R1b1* people at the Dead Sea must have been there by the 8.2 kiloyear event that's when nomadic pastoralism become more significant and the earliest site with domestic ovicaprids show up at Sodmein Cave in Egypt at 5700 BC. And 5400 BC is the age for R-V88 in Chadic people. How does this not make sense?

Marnie said...

@Mike Thomas

"I'd prefer to think of Northern - central Iran as part of one central-northern eurasian geo-ecotone. This might help make better explanations instead of the wrong assumptions we've often been making."

Thanks, and I agree with you.

A good friend of mine is Armenian and is very well connected in the Californian-French-German-Canadian Armenian community. She has literally hundreds of references on Armenian history and has been instrumental in getting many ethnographic accounts published. The Armenian ethnographic record is very well documented.

This record supports the idea of a pasturalist people, with a southern zone in Armenia-Caucasus-Eastern Anatolia, and a northern zone on the Russian Steppe.

Many Armenian celebrations are similar to those that you would see in parts of Northern Iran. Some of their symbolism and mythology is shared with Russians, Ukrainians, and Lithuanians.

I can't help but thinking that some of the time spent by people on this blog, staring at noise, might better be served by reading some of the plentiful references on Armenian ethnography.

That being said, the last vestige of the Armenian migration to the Steppe probably occurs at or before the time of the Pazyryk Carpet:

http://nazmiyalantiquerugs.com/blog/2012/06/pazyryk-carpet-oldest-rug-the-world/

So Armenians have been mostly in Eastern Anatolia and current day Armenia for the last three thousand years.

Like many people from the "Near East" or "West Asia", you would not know they were not from Western Europe, unless they told you.

Armenians are not particularly endogamous, and in fact are an open, friendly people, so it is not surprising to me that in terms of ADMIXTURE analysis, or other autosomal statistics, they look diverse and somewhat like other people from the Caucasus, Anatolia, Iran and Lebanon, while still retaining a signature from the Steppe.

Earl Snerd said...

@ Mike Thomas

"I don't think this is necessarily the case. Don't think special properties and propensities are prerequisites . More to do with blind luck"

If these people were the first to domesticate animals and or expanding as a result of pastoralism they would definetly spread somewhat faster than farmers or hunters, and their lineage also. But it does seem that they were the ones that introduced pastoralism into north Africa first. It just seems a like it's a bit more than blind luck to me. It's the same with haplogroup G and how widespread it seems as a result of farming. It seems more likely too that R1b1 was spreading out along two routes a Mediterranean and north through the Caucasus.

Mike Thomas said...

Sure Earl
You might be right if indeed we find a "genetic trail" in the future, based in aDNA

Marnie said...

Ulrich Shürmann's books on Caucasian and Armenian rugs are a good way to get a grasp of the Armenian/Caucasian pasturalist worldview of the last several thousand years.

Grey said...

@capra

"Samara is not on the border of nowhere. Samara is on the way to the Uralic copper mines."

Quite.

And if wetlands lead to relatively high density HG populations then wetlands around the Black and Caspian Seas and along the rivers draining into it like the Volga then you might be looking at quite a high density (sedentary?) forager population who happened to also be living on or near one of the biggest and most widespread copper fields on the planet.

https://www.awesomestories.com/images/user/c424f925e7.gif

http://en.wikipedia.org/wiki/Kargaly

Earl Snerd said...

The Yamnaya people seem evenly mixed between an indigenous group and people coming from the Near East. Since R1a is found as far north as Karelia during the Mesolithic this must be the haplogroup of the indigenous north Asian population, but with such a significant admixture with Near Eastern people their respective haplogroup must be present as well and since R1b1 is spread from Russia to Spain in two different cultures one certainly from the Mediteranean (Cardial) at 5100 BC it's the only haplogroup associated with the Near East. It is much more likely that R1b1 is Near Eastern in origin. Also the Samarra person was Neolithic those domestic animals had to be coming from the south but not from EEF because there is no sign of contact with EEF. And if you consider that R1b1 had to be in southern Israel by 6200 BC it makes a much stronger case for a Near Eastern origin for R1b1.

PersonaMan said...

Doesn't leave much room geographically for a population in the near east that remained unmixed with the Neolithic populations moving into Europe.

Grey said...

So is it settled now that there were (at least) two sets of farmers/pastoralists?

.

Copper artisans could spread very widely along neolithic trade networks as a minority almost everywhere except in the region where started and where there was some kind of fluke event.

.

"The fact that EHG is so high in Norwegian and Czech, and WHG considerably lower, while EHG is at the same level but WHG peaks in Balts, despite EHG being from the east and WHG from the west"

mountain vs maritime?

Mike Thomas said...

Yes it does ; theoretically, As I've higlighted before if the EEF came from The east Medit littoral; as far as Northeast africa ; and not the Caucasus-Zargos region.
"West Asia" is a big, diverse place
I might be wrong tho

Grey said...

@Earl

"It is much more likely that R1b1 is Near Eastern in origin."

I'd agree it always seemed that way before.

"Also the Samara person was Neolithic those domestic animals had to be coming from the south"

But maybe they traded domesticated horses for them - or stole them using the horses.

"but not from EEF because there is no sign of contact with EEF."

That to me seems like one of the most important bits.

Grey said...

Speaking of copper artisans another angle on that might be the very rapid spread of wheeled vehicles once first invented.

Now I think of it I wonder if that very rapid spread might have been artisans rather than just the idea spreading?

Mike Thomas said...

Lol grey again with ur horses , raids , stealing

PS 1 + 1 = 2

Chad Rohlfsen said...

Earl,
That Samara hunter R1b had no Basal Eurasian. It was not from the Near East, and was the oldest R1b. It was pure EHG.

Chad Rohlfsen said...

The more likely story is that a few R1b HG's spread into the South Caucasus as L278, then turned into V88 there, with other P25 derived lineages staying north. V88 does not spread with IE, Bell Beaker, etc. It seems the two were separate.

Grey said...

"V88 does not spread with IE, Bell Beaker, etc. It seems the two were separate."

Yes different directions.

Grey said...

Some fun quotes from


http://en.wikipedia.org/wiki/Plains_Indians#The_Horse


"In 1592, however, Juan de Onate brought 7,000 head of livestock with him when he came north to establish a colony in New Mexico. His horse herd included mares as well as stallions
...
By 1659, the Navajo from northwestern New Mexico were raiding the Spanish colonies to steal horses."

So at most 70-ish years from first getting them to raiding settled populations.

.

"Pueblo Indians learned about horses by working on the ranches of the Spanish colonists. The Spanish attempted to keep knowledge of riding away from Indians, but the Indians learned and some fled their servitude to Spanish masters—and took the horses with them."

.

"By 1664, the Apaches of the Great Plains were trading captives from other tribes to the Spanish for horses."

"The real beginning of the horse culture of the plains began with the expulsion of the Spanish from New Mexico in 1680 when the victorious Pueblo Indians captured thousands of horses and other livestock."

.

"By 1770, that Plains Indians culture was mature, consisting of mounted buffalo-hunting nomads"

Fully mature after 170 years from first getting horses.

.

This next bit imo may speak to the R1a/R1b angle

"It was the Comanche, coming to the attention of the Spanish in New Mexico in 1706, who first realized the potential of the horse. As pure nomads, hunters, and pastoralists, well supplied with horses, they swept most of the mixed-economy Apaches from the plains"

full pastoralist
> half-pastoralist
>> settled farmer

.

"Formerly egalitarian societies became more divided by wealth with a negative impact on the role of women. Rich men took several wives and captives (slaves) to manage their possessions, especially horses"

Mike Thomas said...

So Chad do u think all R1 originated somewhere in NE europe - NW Eurasia before beginning to differentiate and then expand

Earl Snerd said...

@ Chad Rohlfsen

Some group from the Near East bringing domestic animals would have come through the Caucasus before the Samara culture, they were already Neolithic which means that they acquired the animals and R1b not long after R1b reached Anatolia (6000 BC)from the Middle East, which means the Near Eastern DNA was slowly trickling in at first. By the time the Samara culture had developed, the Near Eastern genetic component would already have disappeared keeping only the R1b haplogroup. The Samara still did not have the most basal lineage of R1b the Cardial did. Also, it would have taken generations for people to go from the Middle East to Spain so the Samara person is only a bit older because he didn't have to travel as far. They both would have left the same area with a similar expansion event arriving, in Spain last only because of the distance. Because of that the difference in age is meaningless. What's left is the original culture of R1b people? the Cardial Culture goes back to 5800 BC in the Near East (similar in age to R1b diversity in Anatolia) the Samara only back to 5500 BC. R-V88 is actually found in the Mediterranean (R-M18). I don't think R1b had anything to do with Indo-European at all. It is obviously a steppe language. R1b could have been a part of the later Kassite language.

Earl Snerd said...

@ Chad Rohlfsen

Also, I wasn't implying that R-V88 had anything to do with Indo-European and its low frequency in the Mediterranean would also be the result of the Cardial Culture but did not take much a part in that since R-V88 originated further south in the Near East.

«Oldest ‹Older   201 – 400 of 438   Newer› Newest»