search this blog

Thursday, February 12, 2015

Eastern Europe as a bifurcation hotspot for Y-hg R1


The main angle of the recently released epic manuscript Haak et al. 2015 is that ancient DNA supports the steppe origin of at least some of Europe's Indo-European languages. That's certainly a move in the right direction, so that we can eventually do away with the Anatolian hypothesis, which was always a failed proposition.

But it's clear that the authors are holding back. They've obviously decided to be very cautious until they've looked at more ancient DNA, particularly from the Near East, Central Asia and India, before backing fully any one Proto-Indo-European (PIE) urheimat model.

That's understandable, considering how much opposition there is still to the steppe hypothesis, even though it does by and large have the support of historical linguists, which is what really counts. Nevertheless, my feeling is that Haak et al. are underselling their data, particularly the stuff from Eastern Europe.

I'm of the opinion that the steppe or Kurgan PIE model works just fine, and also not surprised by the ancient DNA evidence pointing to a massive expansion of people from the western steppe during the Late Neolithic/Early Bronze Age. So for me, the really big news in this paper is that the only two Eastern European forager samples belong to basal lineages of Y-chromosome haplogroups R1a and R1b. What this suggests, Id' say, is that ancient Eastern Europe was a key bifurcation region for R1.

Remarkably, it's possible to basically lay out the history and phylogeny of R1a in Europe using just three R1a samples from the paper. This can't be a coincidence.

- Mesolithic Hunter-Gatherer from Karelia: R1a (xM198)

- Late Neolithic Corded Ware pastoralist from Germany: R1a (M198, M417, xZ282)

- Late Bronze Age Urnfielder from Germany: R1a (M198, M417, Z282, Z280)

What we can see there is the progression from a basal R1a in pre-Neolithic Northeastern Europe to a derived R1a in late prehistoric Central Europe. The derived R1a is actually R1a1a1b1a2, which is by far the most common subclade of R1a in Europe today, and closely related to the Asian and Indo-Iranian-specific R1a1a1b2.

Interestingly, all seven of the Yamnaya males sampled by Haak et al., mostly from the Samara Valley, belong to R1b-M269, the most common subclade of R1b today. However, five belong to the West Asian-specific R1b-Z1203, but none to the West European-specific R1b-M412. Also, all nine Yamnaya samples show Near Eastern admixture, described in the paper as Armenian-like.

Does this perhaps mean that the Proto-Indo-Europeans (and thus Yamnaya) originated in the Near East, as per the Armenian Plateau hypothesis?

I doubt it. The aforementioned Eastern European R1b forager is also from the Samara Valley, and he clearly lacks Near Eastern admixture. So what are the chances that a Near Eastern population with a frequency of R1b-M269 of around 100% moved into an area of Eastern Europe where a more basal R1b was already present, and in fact in a population with no Near Eastern ancestry? Very slim, I'd say.

So how did the Yamnaya herders acquire their Near Eastern admixture? The answer is obvious if we look at their mtDNA haplogroups. These include H, T and W, all of which might have come to Eastern Europe from the Near East.

Of course this doesn't mean that the Eastern European steppe was overrun by Near Eastern Amazons. It's generally accepted that during the Neolithic the steppe was settled by farmers from the Near East, just like much of the rest of Europe, and I'd say that it was mostly the women from these groups who were incorporated into the later pastoralist societies of the steppe. The men, who probably belonged to Near Eastern haplogroups like G or T, might have been killed or marginalized in some way, so that their reproductive success was seriously hampered.

This is not a far fetched scenario. Typical hunter-gatherer Y-haplogroups like I2 and C6 have already been recorded alongside Near Eastern-specific mtDNA lineages at several Neolithic sites in Western and Central Europe. The social mechanisms for this might have been different there than on the steppe, but in any case, it seems that European hunter-gatherer males shacking up with farm girls of largely Near Eastern ancestry was not an unusual occurrence back in the day.

Now, if Eastern Europe was indeed a bifurcation hotspot for R1, then a large proportion, or even the majority of R1a and R1b in Eurasia today, might well be of Eastern European origin. If so, there should be some support for this in genome-wide DNA of present-day Asians, and indeed I think there is.

Below are a couple of principal component analyses (PCA). The first is from Haak et al. and the second from my own West Eurasia K8 analysis (see here). Unfortunately, I don't yet have access to the Yamnaya genomes, but I think it's petty easy to guesstimate where they will land on my plot when I run them in the K8. I marked this spot with an X.



Note that most of the Near Eastern and Caucasian populations are clearly shifted east towards ANE, and also up towards Europe. Moreover, I'd say many of these groups are specifically pushing up towards the Volga-Ural samples and thus the Yamnaya herders.

There's really no other way to explain this outcome. Quite simply, the vast majority of West Asians have relatively recent (post-Neolithic?) ancestry from the Ural or Kazakh steppe, which manifests itself as a west to east cline on PCA, running from the southern Levant to the north Caucasus. This result is easily reproduced on any decent PCA with West Eurasian populations, and can be seen on the Haak et al. plot.

I'm yet to find solid evidence that Indo-European speakers from the Near East, like Armenians, Kurds and Iranians, don't harbor fairly significant ancestry from this northeastern source.

For instance, unlike many people, I don't find unsupervised ADMIXTURE analyses very convincing when they show these groups to be entirely of Near Eastern ancestry. That's because when ADMIXTURE creates a modern Near Eastern/West Asian cluster, it usually lumps within it all of the ancient ancestral components that are today ubiquitous in the Near East. In other words, the steppe admixture which shows up amongst most West Asians on the PCA above is classified as native to the Near East, even though this is unlikely to be true.

See also...

High female mobility in Bronze Age Europe

Ust'-Ishim belongs to K-M526

594 comments:

«Oldest   ‹Older   201 – 400 of 594   Newer›   Newest»
Davidski said...

capra,

I certainly can't agree that this is the only decent argument. A better argument, or in fact two arguments, are the R1a/R1b foragers from Eastern Europe, the latter actually from Samara.

You must be used to some remarkable coincidences if you think the Samara Yamnaya R1b can't be from Samara foragers.

rk,

OK, but if I run the Karelia_HG and it comes out something like ~57% WHG, 38% ANE, and 5% East Eurasian, which I'm half expecting it to, I'll probably lose all faith in formal stats.

For me, that failed ADMIXTUREGRAPH model where Karelia_HG is 40% ANE, but an Amerindian shifted ANE rather than just ANE, gels too well with its mtDNA. I reckon if they had some ancient non-ANE Arctic ENA sample there to provide a mixture edge, instead of the Onge, that model would probably fit.

Unknown said...

But doesn't only giving it a WHG and ANE option to link up with go against seeing if it's different? It's forced to pick the closest thing(s). Maybe I'm wrong.

Shaikorth said...

"Here are some PCA you might find interesting. This one is the standard Human Origins West Eurasia design, but with Loschbour, Stuttgart and Ust'-Ishim added.


IMO, the Euros are too compressed, so I removed some of the Bedouins with inflated Sub-Saharan admixture and high IBD sharing with each other.


The Euros are now less compressed. It looks much better, don't you think?""

Yeah in this case. Same samples define the dimensions in both PCA's so informativeness should be the same, but not getting clustered up makes it easier for the reader.

Could you do the same PCA with your own dataset using these samples except with Loschbour, Stuttgart and Ust-Ishim in place of synthetic ANE:

https://drive.google.com/file/d/0B9o3EYTdM8lQUE9zUVpjRnUtMFU/edit?pli=1

Since this has more European groups and it would be interesting to see if North Africans behave similar to BedouinB to any degree.

Ryukendo K said...

@ Davidski

Maybe, maybe not.

If Onge suffices to produce an ENA edge into Karitiana which spit at least >17 kya from onge and prob much longer, then the fact that an ENA pop might have produced an edge into Karelian is probably a moot point, especially considering the fact that the same pop 'C' breaking off from Onge is likely to have given to both Karitiana and Karelian, instead of Onge itself, and ADMIXTOOLS is capable of inferring that.

All the same if the above happens, you should probably lose faith in ADMIXTURE rather than in formal stats.

ADMIXTURE really sucks at this kinda thing, because of the way it deals with variation. Australians always score as Papuans + East Asian despite this being manifestly untrue. Anything that is outside variation is gonna score in exotics, and if EHG is not just ANE+WHG it will score exotically in e.g. African? Indian? Asian? etc. So I am not inclined to trust it as much. But anyway the Karelians do not score in any ENA component that does not also contain ANE in the authors' ADMIXTURE, which is exactly the same behaviour as Mal'ta. So perhaps not too.

Davidski said...

Shaikorth,

https://drive.google.com/file/d/0B9o3EYTdM8lQaHhna1ZsSnNJem8/view?usp=sharing

rk,

Running the Karitiana is like running high coverage genomes. And we do have other lines of evidence apart from genome-wide DNA that the Karelian HG might have some sort of Arctic East Eurasian admixture.

I haven't yet seen what these 390K SNP capture samples behave like, but if they're anything like low coverage genomes, then they're gonna need some TLC for their results to make sense.

Shaikorth said...

Thanks. North Africans took Dim2 from Bedouins, but not the Near Eastern/WHG differentiation in Dimension 1, despite high SSA in North Africa compared to Arabs.

Result of them having some WHG-like stuff lacking in SW Asia, as ADMIXTURE occasionally shows?

Unknown said...

Dave/ RK/ SS/ Chad

Is there anyway you can summarise and simplify what you chaps have been debating about / refining ; and its implications for the study findings (for those of us a little less specialized in computer hardware), please ?

Matt said...

@ David & Sein - There's a suggestion upthread from David that (if I'm describing it right) Arctic ENA could be divergent enough to admix into EHG and give a higher similarity to Native American without inflating any shared drift with Onge.

Seems like if this is true, it also applies to MA-1 (MA-1 could be admixed with this Arctic group), and then the ANE concept becomes even more uncertain / is scotched again by a different route.

Davidski said...

Well, what I ws saying was that the algorithm possibly gave Karelia_HG a slightly eastern shifted 40% of ANE, with cryptic Amerindian ENA admixture, to compensate for its Arctic ENA or maybe ENA-like ancestry. You can see that in Figure S8.6. This quote caught my eye:

"A different detail of Fig. S8.6 is that the ANE ancestry in Karelia_HG is derived from the branch of “Ancient_North_Eurasian” that goes into the Karitiana Native Americans, rather than the MA1 branch. This is plausible, given the fact that modern humans arrived in the Americas after the time of MA1 (~24,000 years ago), giving added opportunity for more genetic drift to accumulate in the “Ancient North Eurasian” population; this could be shared by the EHG hunter-gatherers of European Russia from the Holocene period."

There's something funny going on here, and I suspect it's linked to the rare, basal mtDNA carried by Karelia_HG. But I don't know enough about ADMIXTUREGRAPH to really suss out what's happening.

Anyway, the model failed, maybe because the Onge were a poor ENA surrogate in this instance? I mean, I'm just guessing really, but think about it, how much would the Onge have in common with ancient Arctic people?

Kristiina said...

”The distrib of I outside of europe is extremely poor, even in siberia Afaik. I don't expect aDNA to change this view. Kristina prob has smth to say abt this.”

Ryukendo, yes, I admit that before the paper was published, I thought that WHG of that Karelian HG could come from y DNA I and his ANE from Q (because of the Native American like mtDNA), but things often do not go as you expect. North European hunter-gatherers have plenty of I2, and I still think that it is not so illogical to assume that a possible extict branch of I could be found in Karelia. According to Wikipedia, hg I*-M170 has been found in 1/35 Saami. Moreover, history seems to be full of extinct y lines. However, it is true that the results were a big surprise to me. I am ready to admit that I did not expect to find a 100% hunter-gatherer R1b from Samara and a genetically very similar R1a hunter-gatherer from Oleni Ostrov (in Finnish Peurasaari)! I readily admit that I have to change my thinking as a result of all this, although there were probably also other lineages in Karelia and Samara that remain unknown to us.

Shaikorth said...

Onge are very different from even South Chinese populations, never mind Arctic ones and it should show in their relation to EHG. Here's a snapshot of their ADMIXTURE graph which may demonstrate.

http://s28.postimg.org/rqa2facbf/admixt.jpg

Onge at even lower K's bring to mind Melanesians and Papuans more than anything (and I assume global PCA's show something like that too), ENA seems kind of a wastebasket taxon in that regard. Perhaps comparable to BedouinB and Loschbour being treated as just two pops in a single clade or even more extreme.

Ryukendo K said...

@ Davidski @ Shaikorth @ matt

First, sorry, I was referring to ADMIXTUREGRAPH all along, not ADMIXTOOLS.

It is extremely unlikely that that is what happened, because ADMIXTUREGRAPH doesnt even deal with the original genomes if I'm not wrong, much less have any ability to give anyone a 'portion of Karitiana with cryptic ENA ancestry'. ADMIXTUREGRAPH just look at all the f-stats between all the pops, in all the combinations between all the pops, and proposes a tree with drift in the branches and admix proportions that accounts for the distances/shared drift, and the -ve and +ve stats, and the Z scores between all the genomes you include.

In that sense ADMIXTUREGRAPH is a much more robust algorithm than ADMIXTURE, as the differences in allele freqs are 'real' and completely measured as opposed to being forced into a product of components produced from modern pops, and the ghost pops proposed by ADMIXTUREGRAPH are very likely real pops. The problem is that ADMIXTUREGRAPH cannot give you the resulting allele freqs of the ghost pops as it doesn't even deal with the genomes, just stats, while ADMIXTURE can fudge up something from the variation we include in the dataset to give us a 'Baltic', 'East Euro' and 'Pure ANE' genomes, etc. Also, ADMIXTUREGRAPH can only deal with a few genomes, at least a few of which must be unadmixed. Otherwise, total chaos, like in TREEMIX.

Therefore, if a tree produced by ADMIXTUREGRAPH fails, it just fails. It has explicitly failed to account for the dists between the genomes, and this is very clear. When ADMIXTURE fails or gives weird results its extremely unclear what's happpening. Hell, its not even clear whats happening when ADMIXTURE succeeds, lol. Thats why we have to refer to corroboration by formal stats.

Do you agree that East Eurasian in Arctic forms a clade with East Eurasian in NAms, which then forms a clade with Dai, which then forms a clade with Onge? That all these form a clade w.r.t. ANE-WHG? This is what the stats suggest, we can go back and look at the paper where WHG and ANE were first defined to clarify the arguments again.

If the above is true, then it necessarily follows that any contribution from an ENA pop will result in increased shared drift with Onge, as that ENA pop that contributed has shared drift up to the common ancestor between Onge and itself, while the west eurasian population only has shared drift up to the common ancestor of all crown Eurasian. f-stats will not fail to notice this, and neither will admixturegraph. It is impossible for a pop that derives from ENA to drift to a position where it is equi between ENA and West Eurasian, , aka impossible for a pure arctic ENA to drift until its signal of shared drift with Onge to the exclusion of West Eurasian is obscured, as no population drifts into any other, or in the direction of any other.

Matt said...

Davidski: Anyway, the model failed, maybe because the Onge were a poor ENA surrogate in this instance? I mean, I'm just guessing really, but think about it, how much would the Onge have in common with ancient Arctic people?

I thought I understood Admixturegraph OK, maybe I don't know it as well as I think, this is all IIUC:

That position you've spotted *should* only reflect closeness between the "ANE-WHG" side ancestry, unless it also inflates Onge-ENA like mixture at least a little.

A third Arctic population that contributed to Native Americans / Arctic and EHG but not Onge or MA1 could inflate EHG's closeness to Americans if EHG is both WHG-ANE admixed, the same distance to Onge as MA-1 and as close to Americans as MA-1. If so though, Ryu's objections about how unlikely it is that all this is perfectly balanced ring true to me, but maybe there's enough noise in the stats that the balance doesn't need to be perfect and just close enough to break statistical levels, I don't know.

Re: Shaikorth's, Onge having admixture from outside the ENA clade was exhaustively tested in their first models. Populations in East Eurasia can be very divergent and very admixed between very divergent subpopulations, the only thing that matters for their relationship in these phylogenic trees though, to West Eurasians is whether they are admixed with any population that split off before an ENA tree split off from West Eurasians.

It might be questionable whether there ever was an ENA tree which split off from West Eurasian clades (the basal and ANE-WHG clades), however there isn't the great evidence for that at the moment, maybe it will show up with Upper Paleolithic East Eurasian samples. There seems like there is some behaviour. At the moment it looks like East Eurasia mainly shows shallow branch offs than West Eurasia, from an ENA clade with less admixture between them, rather than deeper branch offs in the tree.

Davidski said...

There are no fool proof methods in this context. We can only be reasonably sure of something if there are multiple lines of evidence pointing to it.

I'm fully expecting Karelia_HG to show some Arctic ancestry when I run it myself in different ways, and I'll be really shocked if it doesn't.

Kurti said...

"So for me, the really big news in this paper is that the only two Eastern European forager samples belong to basal lineages of Y-chromosome haplogroups R1a and R1b."


Am I missing something or since when is M198 "basal" lineage of R1a?

The R1b's were not basal either. This samples just prove that Russia was a "refugium" for some R1a/R1b lineages since the mesolithic nothing more and nothing less.

Alberto said...

I'm trying to get my head around this idea of taking Armenian wives. Usually, when a group of men go to a new territory (to take it), they might kill the men and mate with the women. Because those are the women available. This I understand.

But EHG from the steppe were in their own homeland. Why did they go somewhere else to take wives? Did EHG women suddenly become sterile for some reason? Or the men suddenly just stopped liking them?

Maybe EHG had high libido, but they could just go to Armenia, have sex and come back. Why to bring the women back as wives?

Or maybe as someone suggested, older, richer men took many local wives, so younger ones needed to steal wives from somewhere else. But why not just go to the nearby village 5 kms. away instead of going 1000 km. to Armenia and back?

The whole idea of stealing wives has potential, I admit, but it still needs a bit of polishing before we can sell it.

Another doubt: Were any EHG found in any Kurgan? Or did EHG get the idea of building Kurgans coincidentally at the same time they got this idea of stealing Armenian wives?

Maybe these Armenian wives actually agreed to marry with the EHG if they promised to bury them in Kurgans? That could explain all and the whole concept could be ready to fly.

Kurti said...

Some people should not become too ridiculous because of their emotional state.

None of the R1a/R1b samples in Samarra or Karelia were basal.

The chances are even higher for the Iberian Neolithic R1b to be basal than the Samarra samples.

Davidski said...

Kurti,

The R1a and R1b from Mesolithic Russia are basal to most of the modern European and Asian R1a and R1b.

And my argument is that either these or other closely related basal lineages from Eastern Europe are ancestral to the vast majority of modern R1a and R1b in the world.

Alberto,

I don't know about stealing women, but the same pattern of Mesolithic Y-hg and Neolithic mtDNA has also shown up in Western and Central Europe. I haven't yet seen any detailed explanation for it.

Shaikorth said...

The Lazaridis et al successful models could fit Onge having Basal Eurasian, and one of the succesful models without Basal Eurasian gave Onge significant ANE. The famous one they ran with as an example had them unadmixed. But would that model look like it does if it didn't have just Karitiana and Onge but something else, like Han or Ami, even Oroqen or Nganasan? They also used ADMIXTURE to get their 15 World-Foci outgroups, from which they at least once dropped Karitiana because they admitted it was admixed. We know it's not the only one, Ulchis, Eskimos and Chukchis for starters have significant ANE. So inclusion there doesn't imply that the only difference between Onge and, say, Ami is drift.

The paper contains admission that their modeling of East Eurasians wasn't that thorough because it's not the study's focus. Once the EHG samples are publicized, we should see the truth from f-stat comparisons between it, MA-1, Loschbour and various East Eurasians.

Unknown said...

@ Alberto

"I'm trying to get my head around this idea of taking Armenian wives."

Im not sure you can get your head around a concept which is utter BS. :)

Autosomal, mt DNA & archaeological data show the obvious southern intrusion into Yamnaya.

What is stumping people (stumping is perhaps a wrong word) is the Y-DNA. As Ive mentioned hundreds of times, this weird patterning is something found many other places in many other periods. There is something about Y-DNA and the way it behaves. Sure, patrilocality has a role.

But what David hasn;t proven is that the newer R1b-Z103 in Yamnaya didn't come from the south. Considering that R1b* was found as far as Spain, the (high) possibility remains that R1b* did not necessarily originate in the Russian steppe. Just because a basal-ish clade is found in the Samara HG, it doens;t mean that the new sub-clades of R1b which invaded south Russia could not have originated amongst the Irano-Armenian pastoralists which moved into Russia. What we're then seeing is repeated movements from the 'west Eurasian homeland' west (spain; as evidence here by an earlier R1b) and north (Sth Russia, slightly later R1bs). Anthropologists have long known Europe was a population sink, which received multiple impulses from west Asia.

Unknown said...

@Marnie
Hi Marnie,
"Just because a culture is patrilocal, does not mean they engaged in wife stealing from enemy tribes, cannibalism, endless warfare, or massive exogamy."
I couldn't agree with you more fully! Also, the presence of Neolithic mtDNA and HG-Y chromosomes in Europe suggest that perhaps Neolithic females may have chosen to marry/adopt Mesolithic males into a Neolithic context.

Matt said...

Shaikorth: The Lazaridis et al successful models could fit Onge having Basal Eurasian, and one of the succesful models without Basal Eurasian gave Onge significant ANE. The famous one they ran with as an example had them unadmixed.

Shaikorth, I get the impression you have repeated this a few times, but this is either not true or you are omitting the reasoning for why they went with the model they went with.

The Basal->Onge model was only successful when they did not integrate MA1, in the Loschbour, Stuttgart, Karitiana, Onge models, p102-p103 of the supplement's text and 103-105 for the figures.

This was violated when MA-1 was introduced, p104-105 text and table.
Now, it could be that something is up with MA-1 and it has a more complex ancestry / is contaminated and so models like Basal East Eurasian->Onge are back on the table and NA could be unadmixed, and if that is the case I'm fine with that; however within Laziridis they did not choose the model they chose of West Eurasian->Native American not arbitrarily but because it was the most parsimonious for those groups (MA1, Loschbour, Stuttgart, Kartiana, Onge). It's irresponsible with the truth to make it sound like an arbitrary choice - for at least one reason, making it sound like an arbitrary choice just encourages the normal idiots who whinge about bias where there is none.

Shaikorth said...

Matt, yes they had a reason (MA-1) for their model and I never had a beef with the conclusion of ANE admixture in Native Americans. Still we come to the question of how much better the poor MA-1 sequence is compared to a ghost population, which they have no problems using as the Basal Eurasian idea and additional ghostpops they add to make Kostenki14 and Ust-Ishim fit the model shows. It's unfortunate that they did not make a successful model without the Basal Eurasian assumption for MA-1.

To restate the original point I think the main issue with those EHG trees is simply just Onge, and not something like Han, being alone with Karitiana as the ENA representative. Once the EHG sample is released it'll be easy to check if I'm wrong in that.

Marnie said...

@Roy

"I couldn't agree with you more fully! Also, the presence of Neolithic mtDNA and HG-Y chromosomes in Europe suggest that perhaps Neolithic females may have chosen to marry/adopt Mesolithic males into a Neolithic context."

Thanks, Roy. The marrying adopting females male relatives probably would have had to have played along with the bargain.

Anyway, if you read Native American history, you come across many accounts of both men and women marrying/being adopted into patrilocal tribes.

Regarding the Armenian discussion, and the somewhat ridiculous scenarios, Armenians have very diverse ydna, not just R1b. So obviously, both men and women "married out".

Ryukendo K said...

@ Shaikorth @ David

I agree that the mtDNA kinda suggests ENA ancestry.

However David, that model that you refer to is precisely the one that the authors described as unparsimonious, because it relies on two different processes, 1) that Karelian shares more drift with ANE in Karitiana than in Mal'ta, and 2) that Karelian contains WHG, to completely balance, such that Karelian is exactly the same dist from Karitiana as Mal'ta is. That it fails even under such a scenario just adds to the implausibility.

Also, I would actually be very surprised if Karelian turned out to have ENA ancestry, because this is immediately obvious in D stats even if its not in F stats, and the authors make no attempt anywhere in the paper to model EHG as admixed with ENA, and the possbility is not even mentioned. This possibility is probably excluded in the detailed calcs in the main paper, which would have a whole panel of ENA pops such as Han, Ami, Papuan etc other than Onge.

@ David
David, I think we will have to rely more and more on formal testing such as ADMIXTUREGRAPH of f-stats, D stats, and the f4 admixture proportion stats (I'll call this 'Haak modelling' from now on lol), in the future, because I agree with Chad that more branches are gonna appear with aDNA until ADMIXTURE is not useful except as a sanity check after the tree has already been constructed. I think if we want to keep abreast of the researchers, we have to invade their methods lol. Otherwise we'll be left with the crumbs left after asteroids of this kind, and at other times we're just stargazing.

Last of all, and this is directed to everyone, I would like to caution against a 'fuzzy' or 'flexible' understanding of what we are talking about at this point. We don't want to make the kind of mistakes that some people make, e.g. 'this pop pulls in this kind of tendency', 'that pop is opposed to that kind of tendency'. The phenomena we are exploring are very precisely defined mathematically, so even if the phenomena are statistical, they have a very precise kind of reality and obey clear rules.

Thats part of the reason why formal modelling of population dists and admix proportions is better than ADMIXTURE I think, because this gets at the heart of the matter directly, instead of ADMIXTURE, which lets us see through the glass darkly at best.

Marnie said...

@Chad

"Yeah, our "Gravettian" or the continuum also put forward by Eske."

I doubt that Eske is the only person to have noticed or proposed this.

You should give credit where credit is due, and that lies firmly with archaeologists like:

J. G. Enloe, D. G. Drucker, A. Bridault, C. Cupillard, Alexander Verpoorte, S. Hartz, T. Terberger, M. Zhilin, R. Miller and L. G. Straus.

Furthermore, I proposed something like this on my blog when the Lazarides 2013 data first came out (which I know you guys all read):

http://www.linearpopulationmodel.blogspot.com/2014/01/mesolithic-western-european-hunter.html

And again on December 20th, 2015:

http://linearpopulationmodel.blogspot.com/2014/12/the-archaeological-record-in-northern.html

In fact, this is exactly why, last December, I insisted that there was no "mass migration" into Europe from the Steppe during the Neolithic. I knew there were earlier waves before that, including a signature left from a Gravettian continuum, but also subsequent waves of movement during the Mesolthic and Neolithic.

Anyway, if I see the notion of a Gravettian continuum not properly attributed, especially after all the BS I've put up with on this blog for the last three month, sh*t is going to hit the fan.

Marnie said...

I would add, to the above list of important archaeologists, M. Germonpre.

Nick Patterson (Broad) said...

(From Nick Patterson)
You guys work hard! I have a job
that (maybe??) one of you would like to take on.
ADMIXTUREGRAPH is only
semi-public because it uses a
commercial optimizer (NAG). It should be possible to replace
NAG
with (say) the GNU scientific library. Anyone want to try?
No money, but appropriate credit!

Qualifications:
Knowledge of C/Unix and programming experience.
email me
nickp@broadinstitute.org
if interested.

Shaikorth said...

"Also, I would actually be very surprised if Karelian turned out to have ENA ancestry, because this is immediately obvious in D stats even if its not in F stats, and the authors make no attempt anywhere in the paper to model EHG as admixed with ENA, and the possbility is not even mentioned. This possibility is probably excluded in the detailed calcs in the main paper, which would have a whole panel of ENA pops such as Han, Ami, Papuan etc other than Onge."

They brought forward the possibility of East Asian-Loschbour admixture event as an alternative for Kostenki14 having Basal Eurasian. If they don't reject something as big as that that, why would omission mean they've ruled out a similar event for EHG?

"The hypothesis of Basal Eurasian ancestry in Kostenki14 needs to be further tested, as the negative D(Mbuti, Han; Loschbour, Kostenki14) statistic could also reflect gene flow between Han and Loschbour. This a priori plausible, as these populations are much younger
than Kostenki14 and may share intra-Eurasian genetic drift that Kostenki14 lacks because of its age."

Marnie said...

@ryu

Admixture, if run up to high enough k values, can be quite informative.

Treemix, if *not* run up to high enough k values, is a dog's breakfast, because it over constrained and forces tree like splits on populations with only soft splits.

f and D statistics smear out low level non-linear effects.

What about ncMCE?

http://www.nature.com/srep/2015/150130/srep08140/full/srep08140.html

How come nobody has tried this?

capra internetensis said...

@Davidski

Of course the Samara Yamnaya R1b could have come from the Samara foragers. That ought to go without saying. (I know some people are in serious denial about this, but I'm not one of them.)

The point is that it is also could have come from somewhere else.

Our sample distribution (if you can dignify it with such a term) does not allow us to narrow down the probabilities here beyond "maybe".

Marnie said...

@capra

You on a sunny day:

https://c2.staticflickr.com/6/5182/5860804547_d15ec071b5.jpg

?

"Our sample distribution (if you can dignify it with such a term) does not allow us to narrow down the probabilities here beyond "maybe"."

Thank you, capra. I would concur.

postneo said...

@kurti
"None of the R1a/R1b samples in Samarra or Karelia were basal.The chances are even higher for the Iberian Neolithic R1b to be basal than the Samarra samples."

Basal R1b has been sloshing around a triangle btw Spain, Samara and Bhutan. Its of course found in Anatolian and Iran. derived R1b seems to be missing from Central and northern Europe till migrations from the Steppes.

The question is how many snp mutations separate the Samara HG r1b and the Yamnaya ones.

postneo said...

Davidski's hypothesis is simply this.

All near "near eastern" contribution to Yamnaya are farmer wives. Since autosomally such a one sided contribution may not be enough, he proposes the notion of steppe raiders/chaperones for the wives.

These raiders are responsible for all basal and derived R1 and perhaps even R2 lineages anywhere outside the Steppes.

I think such a hypothesis has sufficient wiggle room and cannot be disproved. It cannot be proven either.

Krefter said...

In Haak 2015 they label the Samara Hg as R1b1a in that table with all the mtDNA-Y DNA results, but in when they describe the key SNPs it was positive for they don't mention P297.

Did anyone see somewhere in the paper where it mentions he was R1b1a-P297? This would be big news and makes it more likely he was ancestral to late L23.

Fanty said...

I dont know how people can put up ideas like that here because of 6 yamna guys dugg up from a single graveyard.

If these would be 6 random Yamnas from all over the huge Yamna area, I would say: WOW...

But all 6 are from one site (Samara), isnt it? So it doesnt tell shit.

It could be like checking for I1 in Germanic speakers of 2015.

6 our of 6 I1 from a site in Sweden is likely. 6 ou5 of 6 all over the Germanic speaking areal of Europe today is not.

Krefter said...

Fanty,

If I remember correctly all the Yamna males were unrelated. They didn't have the same father, grandfather, great grandfather, etc., etc. They had as separate paternal lineages as 7 random Swedish dudes. I guess there is a chance they were distant relatives.

Autosomally CWC is very similar to Yamna, even though they lived all the way west in Germany. mtDNA wise all Catacomb and Yamna from Bulgaria to the Volga were very similar. Yamna could have been diverse, but probably less diverse than modern Italy or France.

So, if R1b-L23 was very popular in Samara it probably was in other parts of the European steppe.

The R1b in Samara Yamna is a very big deal.

capra internetensis said...

@Krefter

I haven't seen anything to suggest he was P297+. I have also heard people say that one of the Yamnaya guys is P297*. He is actually P297(xL51) only, so there is nothing to stop him being Z2103 like the others. There is one definite L23*(xL51, Z2103).


@Fanty

It's not from one graveyard. 4 are from one relatively small area, but the others are hundreds of km away. One site is in the southeastern Urals, close to the border with Kazakhstan, and the far away ones are still Z2103, while the L23* is from the central site.

Chad said...

If SNP dating is correct, then we are going to see a switch from L51 to L11 between Sredny Stog and Cotafeni.

Chad said...

I think the EHG thing, if anything, would lend more favor to the Steppe hypothesis. It gets rid of that lack of WHG in South Asia argument. Time will tell though. Older samples are needed.

Fanty said...

I wont buy it, before we have one Yamna from Ukraine and a Yamna from almost the border to Georgia aswell with the same Turkish R1b

Fanty said...

Dont get me wrong.
What I really mean is, the forum "war" going on, if these R1b guys are the EHG or the Farmer side of Yamna.

If there is either all EHG guys exterminated and Yamna is 100% Farmer Y-DNA or if all the Farmer guys are exterminated and Yamna is 100% EHG Y-DNA....

I just think its a bit early.

And of course highly ideologic. Thats why it becomes so heated up, close to use of physical violence. ;)

Matt said...

Impressed by Nick Patterson's offer, hope someone takes him up on it. Sounds like a cool opportunity or experience.

Chad: I think the EHG thing, if anything, would lend more favor to the Steppe hypothesis. It gets rid of that lack of WHG in South Asia argument. Time will tell though. Older samples are needed.

I found this section of the paper interesting:

p46 "Modern Armenians have a signal of admixture from the Yamnaya, as when we test f3-statistics of the form f3(Armenian; Yamnaya, X) we find the lowest Z-score for f3(Armenian; Yamnaya, BedouinB) = -0.00296 (Z=-7.1).

However, the lowest Z-score of statistics of the form f3(Armenian; X, Y) involves the (X, Y) =
(LBK_EN, Sindhi) pair (value -0.00575, Z=-15.3), so the signal of admixture from the Yamnaya is not the strongest one for Armenians.

Moreover, as shown in SI 7, the Yamnaya have a negative f3-statistic with (X, Y) = (Karelia_HG, Armenian). A negative statistic for both Armenians and Yamnaya with each other as a reference population may suggest that a third (unsampled) population admixed into both the Yamnaya and to Armenians. The question of directionality can only be furthered elucidated by the study of additional ancient samples from the Caucasus, Near East and the steppe"


This is very interesting to me - the Armenians get a stronger signal from an early farmer population and an Indian population heavy in what looks like a "West Asian Neolithic" component, rather to what is supposedly our best ref for "ENF" (BedouinB) plus either Yamnaya or EHG. The Armenians (and their cohorts like the Georgians, Iraqi Jews) had their strongest f3 stat with Gujarati and Stuttgart in Laziridis 2014, and it seems at least the Armenians still have roughly the same strongest stat, even now we have EHG and Yamnaya.

Doesn't rule out, of course, any spread of y-dna haplos and language into from the steppe (and there is arguably a certain steppe like signature present in the ADMIXTURE for West Asian and South Asian Indo-European speakers). Still, worth keeping in mind.

Strengthens my feeling that it is still possible Early Neolithic Europeans may actually still end up being almost unadmixed Levantine(?) farmers, and that BedouinB plus Yamnaya / EHG admixture may not end up approximating the ancestral history of any people in West Asia so well.

There were a number of populations whose strongest f3 stat was Stuttgart plus MA1 in Laziridis 2014, wonder if any of them have switched to Stuttgart plus EHG?

Marnie said...

For me, it's one thing for members of the public to be discussing the "origin" of a particular haplogroup, but quite another for researchers and those who fund research on genetic ancestry going on the forums (anonymously) to promote the idea that there is likely to be a definitive easy to locate definitive male ancestor.

Worse, this idea is being popularized in journals like Nature, long before there is sufficient sampling coverage to make assertions like the "mass migration" from the Steppe assertion in the Haak 2015 paper.

And the only ideological attachments I have are to scientific honesty, wise use of research dollars, and a reasonable effort to not have scientific research manipulated by special interests.

Many scientists and scientifically trained people complain about people who reject the theory of evolution. But who can blame them when "scientists" themselves are falling back on there own ideologies and biases, tossing objectivity to the wind?

postneo said...

"Strengthens my feeling that it is still possible Early Neolithic Europeans may actually still end up being almost unadmixed Levantine(?) farmers, and that BedouinB plus Yamnaya / EHG admixture may not end up approximating the ancestral history of any people in West Asia so well."

beats me why bedouin are being used as proxies for levantine farmers. They have not practiced agriculture and would have a very different trajectory. early SE European are more pristine
levantine farmers.

Similarly EHG may carry unadmixed HG components from Iran.

Shaikorth said...

They used Bedouins in S9.27 because ancient samples + BedouinB is a better model (smaller resnorm) for Maltese, Spanish, Sicilians and Ashkenazi jews than just one based on Yamnaya/EHG and Neolithic farmers. Bulgarians, Greeks and Albanians on the other hand could be explained with just ancient genomes and got 0% BedouinB and no resnorm reduction when it was added. It's probably not a perfect proxy for all extra Middle Eastern or North African ancestry involved, but they don't have any ancient genomes from the Near East.

Marnie said...

@Mike Thomas

Yes, I agree, it's a generous offer of Patterson. Based on some of his comments I've seen quoted, he seems like a thoughtful person.

I'd even be interested in his proposal, if it weren't for the fact that I've got my hands full at the moment designing next generation cell phone technology.

I agree that there's a lot of important results in the Haak paper. My objection is more with the title and the overselling. And I'm also speaking more generally, not just about the Haak paper.

Davidski said...

Let's get back to discussing the specifics of the paper and data, instead of Kurgan conspiracy theories, or I'll have to start deleting comments.

Unknown said...

@ Fanty

" before we have one Yamna from Ukraine and a Yamna from almost the border to Georgia"

It is possible that further sampling from Yamnaya at large, and the western region in particular, will reveal a native R1a substratum with infiltration of West Asian R1b

Davidski said...

Mike, don't you mean the infiltration of a native R1a/R1b substratum with West Asian R1b?

See that's why your theory is a fail. It's just not very parsimonious.

Grey said...

@postneo @Shaikorth

"beats me why bedouin are being used as proxies for levantine farmer"

If there was an assumption that there was only one source of farmers when in fact there were two (or more) this might have lead the model to be skewed so as to fit the hardest to fit element.

(As Shaikorth's point seems to suggest if I'm understanding it right.)

So maybe a mostly maritime Levantine wave (hence Maltese, Spanish, Sicilians)(maybe due to being somehow sealed off from large chunks of the middle/near east at the time (maybe by the fertile crescent being a giant swamp at the time?) and a separate near eastern wave (until after the fertile swamp was drained by farmers).

With possibly some extra confusion over EEF/WHG because they had shared ancestry in the past.

https://magnaaura.files.wordpress.com/2015/01/720px-aurignacian_culture_map-en-svg.png

Unknown said...

Dave,
Im not saying its parasiminous. Im saying its possible and hasnt been excluded

R1b was already widespread before 5000 BC , yes ?

We haven't excluded that Z103 wasn;t native to north-western Central Asia; have we ? No.

That's all Im saying. Granted, it might be more likely that Z103 simply evolved in Samara from a parental clade.

But then again, who is to day we are to draw a sharp line dividing the areas north and south of the Caspian ?

Moreover, although far too early to tell, it might be that R1a and R1b have differential distributions, with R1a representing an outer arc around a central R1b "core".

Grey said...

@Alberto

"I'm trying to get my head around this idea of taking Armenian wives."

General point: I reckon if people read what happened historically when some plains hunters got horses then eventually the most likely option will sink in.

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

.

However specifically,

"Armenian"

The generally accepted idea a few days ago was a farmer intrusion onto the steppe.

If so then the intrusive farmer population would at that time have been *adjacent* to the steppe population north of the Black Sea - possibly including Kartvel speakers.

So it wouldn't be at all surprising if - if both were patrilocal - for mtdna to swap between the farmers and the steppe people north of the Black Sea.

If at some later date there was a tussle and the farmers retreated the most likely place for them to retreat to would be the regions they originally intruded from.

So any steppe mtdna and/or loan words in Georgia or Armenia?

Davidski said...

Ponto,

I just saw your question.

All Southern Europeans carry WHG, and even many Near Easterners and Central Asians do.

But the WHG that you see in the K8 results is soaked up, in some cases completely, by the EEF component in Haak et al.

The reason this happens is because the K8 doesn't have an EEF component, but rather an ENF component, which lacks WHG. However, this is very difficult to model, and we probably won't see it in scientific literature until someone sequences the genome of an early Neolithic farmer from the Near East who has 0% WHG.

Unknown said...

Dave
Can u clarify & explain the implications if what you guys were discussing about EHG-WHG at length above for the present study ?
If you're soon doing a new post that don't worry

PF said...

"The reason this happens is because the K8 doesn't have an EEF component, but rather an ENF component, which lacks WHG. However, this is very difficult to model, and we probably won't see it in scientific literature until someone sequences the genome of an early Neolithic farmer from the Near East who has 0% WHG."

What do you think this will look like Davidski? Most close to Yemenite Jew?

Davidski said...

Mike,

We were discussing whether EHG was part WHG, ANE and ENA, or WHG part EHG and something as yet unsampled.

My opinion is that all of these possibilities might turn out to be correct, and in fact, MA-1 (Mal'ta boy) might also be mixed, between EHG and a somewhat more extreme ANE, like the one in my K8 test.

This actually makes good sense, considering the patchy sampling of aDNA across space and time in north Eurasia, and the high mobility of hunter-gatherers on the north Eurasian mammoth steppe.

Nevertheless, this doesn't change anything about the fact that the ANE signal, which is very robust if fleshed out correctly, only appeared in Europe west of Ukraine and south of Scandinavia well after the middle Neolithic, or even Copper Age in most places.

Davidski said...

PF,

Yes, it'll probably be very close to Yemenite Jews, some Bedouins and some Saudis, those with less Sub-Saharan admix and no ANE.

Krefter said...

Davidski when do you think the Haak 2015 genomes will be available? Can't wait to see a Yamna admixture test.

Grey said...

Pure guess really mostly based on physical geography but I think there will turn out to be three sets of farmers/herders in the west:

1) E1b related Levant/Egypt set
2) G&J related Caucasus/Zagros/Taurus set
3) R1b related herders from somewhere in the vicinity of Tien Shan

Ryukendo K said...

@ Grey
If you're trying to say that Hunters can make the leap to pastoralism straight, please note that that only occurred in the case of the Indians because all the complex technologies they used, including guns and firearms, were contracted out to Europeans. And actually, these indians did not have a pastoral economy, they had a hunter-gatherer economy with guns and horses, they did not rely on cows or goats or domesticated plants for the rest of their diet, and this is an extremely weird situation in human history which made their lifestyle highly unsustainable ecologically, resulting in the near extermination of the bison throughout much of its range. It was made possible only with firearms that also made them co-dependent with Europeans. In fact, so many aspects of their economy were in a co-dependent relationship with europeans, and in earlier periods the point still stands that there is no precedent for the movement of hunters straight to pastoralists, or the appearance of the nomadic pastoral cultural/tech package among HGs suddenly.

Pastoralists are not in a co-dependent relationship with agricultural pops, they had everything they need on the steppe, though the surplus they accumulated from raiding agris often forced them to structure their political hierarchies.

RMB that the PIEs used their own metallurgy, and a very advanced one too. There is no basis for a metallurgical economy among hunter gatherers.

The situation that occurred with the PIEs in fact is highly remniscient with the Hongshan-->Lower Xiadajian-->Upper Xiadajian transition. In the beginning the YDNA was all O and N, but later as the economy pastoralised it became dominated by YHap C. Then the Hap C pop entered the steppe and the rest is history.

It is pretty clear in ADMIXTURE that if an altaic component is not allowed to form or just doesn't form, all altaics differ from other Siberians and Uralic HGs by having a component that is assoc with agricultural societies further south like Han in addition to the 'Uralic' component, and that this level reaches half of pops like Yakuts, who dominate the altaic component were it allowed to form. I get this from an article that focused ADMIXTURE on a set of NE Asian, NAm, and Siberian pops, but I really cannot rmb the paper, someone pls help me. Was it the one with the Saqqaq?

Ryukendo K said...

About the Haps, they are from Chinese aDNA from Jilin Uni if I'm not wrong.

The full neol package never appeared as far north as the Samara AFAIK, so the appearance of so much farmer ancestry on the steppe suggests to me that the pastoralisation did not take place close to the Samara Valley, but that the same process that took place with the Hongshan took place further south amongst pops that were semi-sedent already.

Grey said...

@ryu

"If you're trying to say that Hunters can make the leap to pastoralism straight"

I'm saying there's a historical example of hunters getting horses and then taking a lot of already domesticated animals (and captives to herd them) from pre-existing farmers/pastoralists.

So yes and no. They make the leap by importing pastoralists.

Grey said...

@ryu

"the point still stands that there is no precedent for the movement of hunters straight to pastoralists"

Isn't it certain that farmers / pastoralists must have started with HGs at some point?

They'd probably need to be sedentary first though so the next question is what makes HGs sedentary.

I'd guess a particularly rich static food source.

So what kind of static food source makes HGs sedentary.

Seafood/wetlands maybe but aren't sheep / goats more of an upland thing.

So what kind of particularly rich static food source might be near a lot of wild sheep and goats.

Wild cereals are one possibility but my guess would be fruit trees: dates, figs, pears, apples etc.

Once it happened once it would likely spread very dramatically so wouldn't likely happen independently too many times.

Grey said...

just to be clear the second point just above is unrelated to the first.

Unknown said...

@ Davidski

Thanks for reply, appreciated.

But you say: "ANE only appears west of ukraine and South of Scandinavia after LN"; based on n=0 samples from the Visla-Carpatho-Dniester basin or indeed southeast Baltic (arguably an ANE refugium) ? ? :)

It also depends which exact part of Ukraine you're referring to ? I'd look to the forest-steppe; and not the open steppe so much.

In fact, Im ready to speculate what future studies will show, based on my archaeological reading:

* large scale depopulation in Late Neolithic central-northern Europe
* tempered somewhat, naturally, by regional variation.
* Eg we have survival of some groups eg the west Alpine (Swiss Lake region) - the SW most extent of future CWC territory.
* Mesolithic (& strong ) survival in
Baltic rim.
* TRB survivals here and there.
* and arguably most importantly Late Tripolye groups in the Moldavo-Ukrainian forest steppe.
* These groups 'fused' and formed the new CWC network.
* The ANE rich, Yamnaya -similar, but not Yamnaya derived groups were perhaps those residual LN groups in Ukraine forest-steppe, which also moved into the open steppe (to form west Yamnaya and related groups), as well as some late Polish TRB groups

Eduardo Pinto said...

"Wild cereals are one possibility but my guess would be fruit trees: dates, figs, pears, apples etc."


@Grey


Sedentarism is a consequence of a change in the mode of production, i.e from food gathering to farming and what you're suggesting is not a change in the mode of production, but simply seasonal or geographical over-abundance.
Plus, Food gatherer societies or HG societies with a focus on food gathering are matricentric which in this case it would exclude PIE with their patriarchal structure.

Davidski said...

I doubt that the southeastern Baltic was an ANE refugium. Although it's possible that the forgers around the Baltic were significantly EHG and thus around 20% ANE, like the Gotland foragers.

Ryukendo K said...

@ Grey
Maybe. But that was a marginal method they had for getting food. And it was the cowboys who had the horses first, not the indians. And we all know what the result of that was. The cowboys are still here, but where are the indians now?

I'm gonna post this here cos some of the speculations here are getting really ahistorical.

Pastoralism is a highly complex and specific cultural package that requires no less a suite of cultural and tech and social adaptations from a HG lifestyle than agriculture. Its less like trade, or music, or sculpture, and more like agriculture, or writing, or realistic art--it needs to be innovated into existence, and there's nothing 'natural' about it that suggests it would simply arise once the bare essentials, e.g. domestic animals plus humans, are present.

This is why it makes sense to ask questions like Whence Pastoralism in West Africa? or Multiregional origin of Pastoralism in Eurasia?, because its one of those things, like the alphabet or plant domestication or furnace metallurgy, that spread out from several rare foci instead of arising spontaneously wherever HGs met neols. And that is how archaeos conceive of nomadic pastoralism as an econ package too.

Pastoralism doesn't appear in a single generation, it takes time for people to figure out things like, how to control the animals at night? How to herd to prevent them from running away? Where to go for each season? What kind of dwelling is the best, and how to build it? How to watch for wolves? and all these things have to be fixed before it becomes sustainable as a basis for your economy. In the meantime you need some other basis that allows enough change to occur in the behaviours w.r.t. domestic animals without disrupting anything.

Ryukendo K said...


Even more important, every soc has some role or behavviour that it esteems, and these are very slow to change. We know that uncontacted HGs are absolutely baffled when they come into contact with neols, they can't for the life of them figure out why they work so hard? Same for urban capitalists like us, why we sent our kids for stupid drudgery evvery day? Why we decorate ourselves then show off in a loud place on saturday night? But we know, because our culture already made the decision for us, it tells the farmers that productivve and reliable farmers get wives in neols, and ppl who have skills and keep appearances up get social status for us, and these interact with human instincts to produce behaviours that keep the society sustainable, which causes those cultures to spread. The fact that those indians put the herdders in the position of slaves is highly telling, and that they didn't esteem the role over the few centuries they had horses is very telling, and the fact that they failed to produce a cultural package that produced the behaviours that demonstrate the full package of nomadic pastoralism as an economy is very telling.

I think its highly likely we're gonna find R1b, half EHG and partially or even fully neolithicised, semi-sedent peoples if we sample densely. This phenomenon has happened many times before. So many times w.r.t the peoples around mesopotamia, who tended assimilate partially at the margins of the settled pops and then conquered mesopotamia in ever-widening circles, that Ibn Khaldun noticed a pattern by the medieval period already.

There might even be fully EHG people with a HG lifestyle who also semi-pastoralised like the Indians, but they probably went the way of the indians as well, before the partially neol ones.

Ryukendo K said...

@ Grey
Lastly, I highly doubt that sedent HGs existed in Siberia, because the temperate places where they have existed before, such as Jomon Japan and the pacific NW, and some parts of the S Am cone, had extreme levels of rainfall, so high it approaches the level of tropical countries, and were also next to the sea near areas of mixed currents and upwelling. Ecol succession in Japan takes place in ~100 years (!)--thats how productive the area was, which was used by archaeo in Japan to explain why HG japan was so densely populated. And kelp forests existed in both japan and NW Pac and Chile. This is of course not the case in Siberia.

Grey said...

@Eduardo Pinto

"Sedentarism is a consequence of a change in the mode of production, i.e from food gathering to farming and what you're suggesting is not a change in the mode of production, but simply seasonal or geographical over-abundance."

I'm suggesting if a group of HGs has access to a particularly rich static food source they'll fight for it and that will tend to make them semi-sedentary sticking to a range around the source.

Anyway it's not a big deal but if i was an archaeologist looking for sites where farming/pastoralism started I'd look for pre agricultural settlements whose old names meant something like valley of: dates, figs, apples, pears etc.

For example Jericho

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

"Jericho is described in the Hebrew Bible as the "City of Palm Trees""



Grey said...

@ryu

"I highly doubt that sedent HGs existed in Siberia"

I was thinking hte southern tip of Kazakhstan near the southern tip of Davidski's ANE map

https://drive.google.com/file/d/0B9o3EYTdM8lQOS1YODhvOWpWNzg/view


(this is totally separate to the Samara thing btw - prob shouldn't have mentioned them at the same time)

Marnie said...

The plains buffalo went (nearly) extinct for the following reasons:

American Plains Indians found a willing European market for buffalo hides and increased their production of hides to meet European demand.

Buffalo were hunted not just by Plains Indians, but by Europeans/Americans, for their commercial value.

European/American access to buffalo herds increased dramatically with the building of the railways.

The range of the buffalo was reduced with the westward expansion of American and Canadian settlers.

After the Battle of Little Bighorn, in 1876, the Lakota Sioux, Cheyenne, and Arapaho tribes took refuge on the Canadian side, in the traditional territory of the Blackfoot, thereby doubling the population in Blackfoot territory, which was one of the last refuges of the buffalo.

It is thought that this doubling of the American Plains Indians population in Blackfoot territory, due to actions by the American government to kill the Sioux, forcing them north, was a primary contributor to the final reduction of the buffalo in Blackfoot territory, in 1879 and 1880.

The adoption of the horse by plains group was only a secondary or tertiary factor in the extinction of the buffalo.

And I agree with "ryu" that the situation of American Plains Indian hunter-gatherers trying to meet the demands of a European market for hides does not represent the situation on the Steppe 5,000 or 10,000 years ago.

Grey said...

@ryu

"And it was the cowboys who had the horses first, not the indians. And we all know what the result of that was. The cowboys are still here, but where are the indians now?"

Sure but what if it was the LBK who colonized America?

I think they'd have been blocked on the great plains same as they were blocked on the steppe.

Anyway, two arguments are jumbled up here so best leave it till another time :)

Grey said...

"LBK farmers" meant to be "neolithic farmers"

Ryukendo K said...

Last of all, there are a lot of comments of the kind that 'CT people were rich and smart' or 'This or that society decided to do this cos its favourable', when its pretty clear that from the point of view of cultural evolution that rich and smarts doesn't do *** for you or your society when the military or ecological circumstances change such that the cultural and economic adaptations of your society produce behaviours that are maladapted. Also, its pretty clear that all kinds of cultures produce behaviour that is extremely counterproductive to their soc, but this never changes or only does so when forced, because anyone who tries to do so is socially penalised, thus cult changes so slowly.

If America suddenly has a zombie apocalypse, the first place to be depop would be all the rich and smart and liberal areas, and thats not gonna change even after years.

So it doesn't just happen that a person decides, 'Oh, I'm gonna try to live off of animal products in a unsedentary economy only!' and then 'POOF'--pastoralism!

A lot of these changes, e.g. neolithicisation, or urbanisation, or indistrialisation, or trade, occurred due to factors that were outside of any persons' control, and its unclear what 'control' over these processes societies have, or what that even means.

Marnie said...

@Grey

"I'm saying there's a historical example of hunters getting horses and then taking a lot of already domesticated animals ( from pre-existing farmers/pastoralists."and captives to herd them)

This is not the case.

Once they acquired horses, Native Americans mostly bred their own horses. They were not in a constant state of stealing horses from "farmers" or keeping "and captives to herd them."

See "Native American Horse Breeds":
http://www.nmai.si.edu/exhibitions/horsenation/breeds.html

Eduardo Pinto said...

@Grey


"I'm suggesting if a group of HGs has access to a particularly rich static food source they'll fight for it and that will tend to make them semi-sedentary sticking to a range around the source."


Well, first of all the practice of pastoralism is linked - for the most part - with scarce, flat, semi-arid places not with abundant ones, hence the mobility plan, the want for land and the possession of cattle.

Secondly, the emergence of farming is linked with the existence of multiple closeby ecosystems - as i.e. in the levant - and the creation of relations of production between it, an economic structure so to say, last but not the least, a climate change would also be needed to force HG into farming, otherwise it would be pointless - i.e Younger dryas - The Pontic-Caspian steppe however is not a very diverse place in the terms above exposed, and farming could have not developed there on its own but only through transmission.

Marnie said...

@ryu

"A lot of these changes, e.g. neolithicisation, or urbanisation, or indistrialisation, or trade, occurred due to factors that were outside of any persons' control, and its unclear what 'control' over these processes societies have, or what that even means."

Thank you.

Grey said...

@Marnie

"Once they acquired horses, Native Americans mostly bred their own horses."

So once they obtained domesticated animals they learned how to breed them quite easily then.

Grey said...

@Eduardo Pinto

"Well, first of all the practice of pastoralism is linked - for the most part - with scarce, flat, semi-arid places not with abundant ones"

Which would be a reason for sticking close to a particularly valuable food source like a valley of fruit trees.

.

"The Pontic-Caspian steppe however is not a very diverse place in the terms above exposed, and farming could have not developed there on its own but only through transmission."

I agree with that. Two separate arguments have got jumbled.

Marnie said...

@Grey,

"So once they obtained domesticated animals they learned how to breed them quite easily then."

In the case of horses, that seems to have been the case for many Plains Indian groups, but not universally so.

In the case of cattle, Plains Indians weren't interested in switching from buffalo to cattle.

In the end, with the demise of the buffalo, a few groups did make the transition to farming cattle, but it wasn't easy.

Here you go:

http://plainshumanities.unl.edu/encyclopedia/doc/egp.na.014

Eduardo Pinto said...

@Grey

The pastoralist mindset doesn't work that way.
Pastoralists would not settle for such a place, because for them the horizon is the limit and their thirst for power and land and cattle is unquenchable.
The line of the horizon on a desert or steppe is terribly powerful on man, it leads him to gaze the infinit, to contemplate abstraction and to seek wealth and power.

Ryukendo K said...

@ Grey

"So once they obtained domesticated animals they learned how to breed them quite easily then."

So why didn't they?

Grey said...

@Eduardo Pinto

They wouldn't be pastoralists yet though.

If becoming semi-sedentary led a population of HGs to eventually domesticate some sheep or whatever then yes I agree.

Grey said...

@ryu

"So why didn't they?"

Marnie's post suggested they did.

http://www.nmai.si.edu/exhibitions/horsenation/breeds.html

Grey said...

@Marnie

"In the end, with the demise of the buffalo, a few groups did make the transition to farming cattle, but it wasn't easy.

Here you go:

http://plainshumanities.unl.edu/encyclopedia/doc/egp.na.014"

ty

Marnie said...

@Grey

welcome!

Grey said...

http://www.nmai.si.edu/exhibitions/horsenation/breeds.html

off-topic but a nice bit of symmetry in the horse breeding link with the Nez Perce breeding their Apaloosa breed with what might be the oldest domesticated breed from Turkmenistan.

Marnie said...

Grey,

I don't know if this would be helpful in answering the question of farming vs. pastoralism on the Steppe, but I read the great book a few years ago:

"The Nomads of South Siberia, the Pastoral Economies of Tuva"

by Sevyan Vainshtein

I don't think Tuvans can be simply classified as hunter-gatherers or pasturalists.

Ryukendo K said...

@ Grey
So I suppose the agriculturalists in the steppe taught the HGs-with-horses how? As part of a concerted effort by their tribal authorities, no doubt.

I believe the evidence is pretty clear on this one. Its very unlikely that the yamnaya up north got so much farmer ancestry locally from ppl close by, they just weren't there. There was probably a pop in the s steppe that initiated the process of neolithicisation, then pastoralisation + metallurgy, just like the situation with the altaics spinning off from neol NE China, or the Arabs spinning off into the peninsula from the Mediterranean world, or the Semites from the levant, the tibetans from the sichuan foothills, and so on.

Marnie said...

@Grey

"off-topic but a nice bit of symmetry in the horse breeding link with the Nez Perce breeding their Apaloosa breed with what might be the oldest domesticated breed from Turkmenistan."

yeah. funny coincidence.

Shaikorth said...

"I doubt that the southeastern Baltic was an ANE refugium. Although it's possible that the forgers around the Baltic were significantly EHG and thus around 20% ANE, like the Gotland foragers."


The Haak fits point towards it being a WHG refugium in the sense that there's extra WHG over Yamnaya/EHG and Neolithic packages. Although the fits with Nganasan added have smaller residuals around Northern Europe and they say Orcadians have a bit more WHG surplus than for example Lithuanians, this could just be a problem caused by modern siberians being an imperfect proxy as I said before. Especially as Norwegians and Scots which should be close to parent populations for Orcadians show no such extreme WHG surplus in them.




"I get this from an article that focused ADMIXTURE on a set of NE Asian, NAm, and Siberian pops, but I really cannot rmb the paper, someone pls help me. Was it the one with the Saqqaq? "

RK,

http://biorxiv.org/content/biorxiv/suppl/2014/08/13/005850.DC2/005850-1.pdf

The "East Asian farmer" component you mention appears at K=3. It's in all Altaics including Yakuts, although the "pan-Altaic" component at K=9 forms around Buryats and Mongols and not Yakuts which makes sense due to sampling in the run and also historically.

Ryukendo K said...

@ Shaikorth
Thanks!

@ Grey
The component appears at K3 and persists in all Altaics to the exclusion of Siberian HGs until the Altaic component forms, and is at 50% in Mongols, who have the other 50% from a component that maxes in Siberian HGs, in a situation remniscient of Yamnaya.

Krefter said...

It's interesting how diverse Late Neolithic/Bronze age Germany was. Add to this BR1 and BR2 from Hungary who were much less Yamna-like than bronze age Germans.

This makes sense because the samples are from differnt eras and cultures, and "eastern" ancestry had just arrived and was gradually admixing with natives. All except maybe Bell beaker appear to have more Yamna-like ancestry than any modern Euros.

>Corded ware obviously has the most EHG/Yamna ancestry.

>Unetice has significantly more WHG-affinity than the others.

>Karsdorf is very non-ENF. Similar to Unetice, but with a higher EHG/WHG.

>Alberstadt_LN, Benz_LN, and Halberstadt_LBA appear to have little(relatively speaking) WHG like CWC and significantly less Yamna but more than modern Euros.

Overtime proportions of EEF, Yamna, and WHG/SHG ancestry must have became blended and more smooth.

It'll be interesting with more ancient genomes to see when/how this happened. Even up to the Iron age we might find "weirdos" in France or wherever with something like 70% Yamna.

Anyways Haak 2015 suggests that LN/BA-type or other "eastern"-heavy pops replaced most of the previous pops of northern Europe. They also made a big impact on southwest Europe, as Basque and other Iberians can fit as over 50% LN/BA.

This leaves room for Iberian R1b-L11 to be descended of LN/BAs, as they are as about as similar to them as to EEFs.

Unknown said...

Dave, RK, anyone

Looking at the graph of of Yamnaya components in modern populations, why is it that it peaks in Scandinavia, Baltic, and even Scots over Ukrainians and Bulgarians, who should have absorbed the very first impact of the "Yamnaya invasion" ?

Balaji said...

Actually the Haak paper provides evidence that the bifurcation of R1 happening in Asia and not in Europe. I am referring to the R1b early neolithic individual from Spain. His ancestors are very unlikely to have been hunter gatherers from far-Eastern Europe. Instead they must have migrated there from Asia. There were neolithic communities stretching all the way from Southern Europe to the Middle East and Central Asia. His ancestors must have been craftsmen whose skills were in demand in these communities and who migrated to the west.

The R1a and R1b hunter-gatherers also likely had ancestors from Central Asia and is part of their 40% ANE ancestry.

The interaction between the “near-Eastern” ancestors of the Yanmaya and their EHG ancestors probably happened in the Steppe East of the Caspian Sea. The near-Easterners would have found these high latitudes not favorable for cultivation of their crops but quite suitable for grazing their cattle. They moved into the territory of the EHG. Hunter-gatherers typically live in small bands and at low population densities. The same land can support a higher population of pastoralists. The interaction would have been somewhat analogous to that between Cowboys and Indians in the American West and we know how that turned out! But the Near-Easterners did not enjoy the same degree of technological advantage as the Cowboys. They must have had to entice the EHG to allow them into lands by giving them some Kim Kardashian-like trophy wives and some of their cattle. In this new merged population which became the Yamnaya, eventually most of the male lineages that won out would be from the Near-Easterners. This is because they would have been the more successful herders with more cattle and therefore more wives. Probably all nine Yamnaya individuals have their paternal ancestry in the Near-East.

The Haak paper does not give an estimate for the ANE of the Near-Eastern ancestors of the Yamnaya but looking at the PCA, the ANE appears to me to be about the same as that of the EHG or 40%. So finding R1b among them should not be surprising.

While the Yamnaya and the Corded Ware people spread Indo-European languages in Europe, the language family itself likely originated in Asia.

It appears that R1a was not too common in Eastern Europe in the early days and became prominent only in historical times with the Slavic expansion,

Shaikorth said...

Besides what David said, modern Ukrainians and Bulgarians haven't lived there unchanged since Yamnaya times. There have been many population invasions to Bulgaria since Bronze Age, and Ukrainians are mostly Slavic migrants who came in after Russia took the region from tatars, in previous times the Ukrainian plain has been inhabited by huns, scythians etc.

Their model also suggests there's excess WHG in many Eastern Europeans beyond what Yamnaya can explain which may go along with the idea of Southern Baltic being a WHG refugium, the remnants of which spread with Slavic expansion. Mordovians who are fst-wise closest to Yamnaya, do have high levels of it according to the least-residuals fit:
http://oi57.tinypic.com/bvgja.jpg

Davidski said...

Balaji,

The R1b Neolithic sample isn't relevant. R1b is an old haplogroup, so the Neolithic farmer might have had ancestry from Eastern European mammoth hunters, or a very early EHG incursion into the Near East, like with pressure blade technology from Siberia.

What really counts is that we now have ancient R*, R1a and R1b samples from the mammoth steppe and nearby. Modern samples from the Near East can't compete with that.

Btw, the Yamnaya R1b isn't really from R1a country. R1b is still common in that part of Russia, while R1a more to the west.

Mike,

Those Yamnaya estimates seem to be a little off. Oetzi scores a bit of Yamnaya, which is very unlikely. So whatever is causing Oetzi to show a bit of this admixture might be inflating it among the Norwegians and Scots.

Anonymous said...

@ Balaji
Tanais steppe scythians had also already R1a and Corded ware in eastern Europe was also rich in R1a so R1a in eastern europe predates the slavic expansion. As long as there are no more samples from yamna core regions we can make no conclusions about the dominant haplogroups or subclades there

Unknown said...

Sure, David, those estimates, you can argue, are a little off. But nevertheless, there should be a decreasing gradient with IBD as one goes to NW. This is not the case at all. Just doesn't quite add up, for me.

Matt said...

Re: why Yamnaya does not peak in Ukraine and Bulgaria, one thing I noticed in the f3 stats was:

For the Late Neolithic f3 stats (Dinka;Corded_Ware_LN;Y), page 66, measuring sharing with Corded Ware, shared f3 is higher for LBK_EN (German LBK) than for Ukrainians, Hungarians, Russians or Croatians and also higher for what seems like all the Northwest Europeans than LBK.

So whether Yamnaya is measured by its improvement to the residuals or via direct shared drift with Corded Ware (as this sole mediator of Yamnaya into Europe), it seems like this shows the same pattern of peaking in Lithuania and Norway.

Other interesting f3 stats that popped out at me:

- Looking at the figure S7.4 (page 63) - the Karsdorf_LN sample, the weird one which is hard to place, shares a lot more drift via f3 with MA1 than any other West Eurasian, almost close by the Karitiana or Motala HG (although it does have a large error bar on its f3) while its a lot more of a typical Late Neolithic Corded Ware like sample when it comes to sharing with Samara or Karelia HG. Odd.

- With the Bell Beakers (Dinka;Bell_Beaker_LN;Y), the Spain_EN and Spain_MN signals also get a bump pretty high up the list. Specifically Spain- seems like that might support BB as a signal out of Iberia?

- Karelia also seems to share more drift with Native Americans than Samara, which doesn't have any NA on its list of top shared drift.... but then MA1 is also higher on the Karelia_HG shared drift list as well, so that doesn't mean an ENA signal.

- Under (Dinka;[WHG];Y) Middle Neolithic populations in Spain shared as much drift with Loschbour and La Brana as Lithuanians do today, SwedenSkoglund_MN more than Estonian, etc.

- Despite replacement for LBK_EN (German LBK Early Neolithic), under (Dinka;LBK_EN;Y) Unetice shares more drift with LBK_EN than Tuscans, Corded Ware more drift with LBK_EN than English or Spanish and Bell_Beaker_LN with LBK_EN more drift than Spanish_North or French_South. That's not what my intuition would've been.

- That pattern seems to reduce a lot when the LBK sample from Hungary is compared, nor the Starcevo or Spanish_EN samples. However, the Bell_Beaker and Unetice samples seem to retain high levels of similarity to them relative to present day North Europeans.

- Similarily the f3s (Dinka;Spain_MN;Y) and (Dinka;Esperstedt_MN;Y) are higher for Corded_Ware_LN has higher shared drift with than any modern North European populations.

Maybe some of these stats have to do with the German provenance of many of the samples?

Alternatively, if real, could reflect that further population movements or absorption involving EHG / others have effected Eastern Europe (or some parts of Eastern Europe) since Yamnaya and Corded Ware?

Davidski said...

Yeah the reason it doesn't add up is because of the large-scale migrations we've been talking about here.

Unknown said...

but the large scale migrations (eg Slavs to Ukraine and Bulgaria should only have increased Yamnaya type compnents; not decreased. The Slavic migrations were an 'internal-colonization' of a population whose ancestors were arguably direct descendents of CWC and north-western Yamnaya groups.

Unknown said...

Thanks Matt & Dave

Shaikorth said...

Slavic migrations form just a small part of modern Bulgarian makeup.

For Ukrainians the explanation is more complicated, it could be that something's just wrong with their method, or Western Yamnaya samples would give higher Yamnaya to Balto-Slavic pops or Slavic expansion spread extra WHG from South Baltic which makes them off-balance compared to Norwegians in their relation to Yamnaya.

Davidski said...

OK, well let's say that most of the R1b in Western Europe came from Yamnaya groups from the present Russo-Kazakh border, while most R1a in Eastern Europe came from Yamnaya groups in Ukraine. That would screw up the isolation-by-distance model a bit, wouldn't it?

Shaikorth said...

Matt, Samara_HG has less SNP's available for the analysis than Karelia_HG according to page 69 and that may have an effect to drift stats.

There is also Yamnaya continuity in Russia according to the model with Nganasan and BedouinB added, it improves resnorm for both Mordovians and Norwegians and gives more Yamnaya to the former, as it does for Lithuanians.

RE: Yamnaya R1b, wait for more samples. Certainly an issue with potential.

Unknown said...

Dave that scenario is implausible unless those R1b chaps from Kazakh Region flew. :)
I'll be proven right - R1a came from a more broad EE groups ; not just western yamnayans (which in any case were themselves derivatives and not conquerors );
I'm not prepared to spectate on where WE R1b came from though

Davidski said...

Mike, something remarkable happened on the steppes during the late Neolithic. It's time to acknowledge that.

Unknown said...

Yes. It was colonized and given life by Communities of the Cucuteni-Tripolye culture. :)

Unknown said...

and communities from south of the Caucasus.

Ryukendo K said...

@ Shaikorth @ Matt @ David @ Mike
I'm pretty sure that EHG survival would actually drive up Yamnaya stats for a variety of Northeast europeans in Haak modelling, since the only source of EHG is gonna be from Yamnaya whether or not it was actually IE. Which is why I always thought replicating their Haak modeling with EHG, WHG and Yamnaya all together would result in much better results, even though the residuals are not gonna improve by much.

To give you an example, the f3 stats for Karelia HG have Saami, Lithuanian, Estonian, Finnish, as the top present-day pops. The Samara HG have Lithuanian, Estonian, Finnish, as tops.

The IE pops howevver have West Asian admix that would be expected to distinguish between EHG survival vs contribs from Yamnaya. So CW has Lithuanian, Icelandic, Norwegian, Orcadian, as tops, and finnish drops all the way. Unetice has Lithuanian, Norwegian, Icelandic; finnish is close to bottom. LBA has Lithuanian, English, Icelandic, Norwegian; and finnish isn't even high enough to appear on the list.

Let me remind you that this is also the pattern demonstrated by Gedrosia in Europe, as well as the West Asian component in Yamnaya ADMIXTURE in the paper itself.

So I don't really think there is inflation going on for NW europeans, since all the IE samples show it. If there is any it should be for NE Europeans, because in the Haak modelling Finnish consistently score higher Yamnaya than any other pop, and this is of course untrue, and is probably an artifact of EHG survival and the fact that all of it has to be attributed to Yamnaya in the model that models europeans with only 3 pops, west asian portion disregarded. I think they should have incl. both EHG and Yamnaya in a four-way. Its not so much the method as the way they used it.

Shaikorth said...

RK, EHG inflating Yamnaya is a plausible explanation, especially considering WHG+EN+EHG reduces residuals for Finnish sample more than WHG+EN+Yamnaya and if Yamnaya can be linked with West Asian/East Med components. However such inflation could happen in Norway too given that they may have extra shared population history with Saamis compared to Czech, Ukrainians etc.

Vadim Verenich had enough Saami samples to include in MDLP tests.
http://magnusducatus.blogspot.com/2012/09/the-component-maps-of-mdlp-world22.html

"North-Euro-Mesolithic" is the Saami signature component.

Extra EHG survival could be further verified with something as simple as this PCA, if NE-euros stretches further along dimension 1 when Loschbour is replaced with Karelian HG.

https://drive.google.com/file/d/0B9o3EYTdM8lQaHhna1ZsSnNJem8/view?pli=1

postneo said...

@Balaji
Kim Kardashian or Mayawati ?

Heres a better explanation. Two adjacent populations one large the other small. Mostly endogamous but with some incidences of exogamy.

The MTDNA of both spread first. The big difference is the pastoral, high density populations carry diseases and immunity which can prove catastrophic to the HGs. So the agro-pastoral MTDNA quickly overwhelms the HG. The HG mtdna in the farming population is negligible and similarly lost.

Later the Y HGs follow the the women folk and remaining traces of HGs are finally absorbed into the agro-pastoral fold.

HIgh population and animal domestication equates to a sophisticated immune system and higher population densities.

Now theres the question of HG resurgence. Its basically their immune system catching up and them learning agro-pastoralism. Therein perhaps lies some hope for Davidski hypotheses.

I agree though that the Iberian sample does support an asian bifurcation of R1.





postneo said...

The only way for the Steppe R1 HG lineages to spread into asia is to have participated in the HG resurgence. and then spread south or perhaps east from Yamnaya.

But Iberian sample is too early for that. Archeology and ADNA from Armenia does not seem to support this.

Alberto said...

@David

"The reason this happens is because the K8 doesn't have an EEF component, but rather an ENF component, which lacks WHG. However, this is very difficult to model, and we probably won't see it in scientific literature until someone sequences the genome of an early Neolithic farmer from the Near East who has 0% WHG."

Thinking about this ENF component, and the West Asian one, I have a question:

Don't you think that the West Asian component is somehow related to ANE one? I mean, both peak in Caucasus and Pakistan, both are lower (among Europeans) in Sardinians and in Basques, both are absent in North West Africans and in EEF...

And we know that ANE is closer (genetically) to WHG then to Basal Eurasian (ENF). So the best way to isolate this Basal Eurasian component wouldn't it be to use something that you're sure lacks not only ANE, but also West Asian?

I'm thinking about this because once you get your hands on the Yamnaya samples (that seem to have a lot of West Asian, but none EEF) I'm afraid they will show quite higher ENF than they should and quite lower ANE than they should. This is because I think your ENF uses some population high in West Asian. Bedouins and Saudis in k7b show up to 30% West Asian:

http://dodecad.blogspot.com/2012/01/k12b-and-k7b-calculators.html

Also this could be the reason why Fst distance between you ANE and ENF is lower than between ANE and WHG. Yes, I know you said Fst distance is not 100% reliable, but I guess it's better when it matches the expectations than when it doesn't (and note than in k7b, Fst between West Asian and Southern is almost double than between West Asian and Atlantic_Baltic, as it should).

So here's the question: Would it be possible to isolate this ENF from sources than surely lack West Asian? Something like using Ötzi and masking the WHG segments? (or including more EEF, except maybe Stuttgart that seems to show West Asian for some reason, maybe contamination). There are other possibilities, like Sardinians (they still have a bit of West Asian, but much lower than Bedouins), or even NW Africans (but here not only WHG should be masked, but also Sub-Saharan).

Any thoughts about this? Thanks!

Matt said...

R Kendow:

I'm pretty sure that EHG survival would actually drive up Yamnaya stats for a variety of Northeast europeans in Haak modelling

I'd expect greater EHG contribution would inflate similarity for NE Europeans with American, Arctic and Siberian outgroups and lessen it with Europeans and West Asians, so some form of greater Neolithic component would need to balance, and it maybe that would be Yamnaya, in the absence of a direct measure of West Asian?

Way I understand "Haak modelling" is that it sort of combines sharing stats in the ingroup from the donor populations to produce the correct outgroup relationships when ingroup population specific proportions are used.

I think they should have incl. both EHG and Yamnaya in a four-way. Its not so much the method as the way they used it.

Not sure why they didn't include EN+WHG+Yamnaya+EHG models before including Nganasan, as EHG would probably result in a some further fit, esp. since adding Nganasan to the fit outside the identified outlier may be representing increased EHG ancestry, which would also inflate the outgroup relationships Nganasan does, with a slightly different pattern. Perhaps some technical issue?

Chad said...

R1a will not be dominant in SW Yamnaya, but I will bet that it is West of Samara, on into Belarus.

If R1a Corded was from SW Yamnaya it should have a good amount of MNE/Balkan ancestry, but it doesn't. Bell Beaker does though!! And it has too much Yamnaya to be derived from Corded Ware. That would make BB over 2/3 Corded without sharing much in uniparental markers... not likely.

I will put money on L51 being found around Ukraine. New SNP dating would put it in the Sredny Stog timeframe. It should be there.... Those Kemi Oba Kurgans end up west of the Carpathians, while Corded was never found south of the Carpathians. Those Kurgan groups show influence on Early Mako, Vucedol, and Bell Beaker. This group is distinct from Corded Ware, and as mixed Balkan elements. It doesn't fit Corded, but Bell Beaker R1b.

Ryukendo K said...

@ Matt
Edited my comment, the previous logic is incorrect.

If I'm not worng, the 15 outgroups should fade into the background; they're only there to give us a sense of what the Yamnaya, WHG, EEF 'are', in the sense of how are these three differentially related to the outgroups? Then the pops are fitted between these so that something like the differential rel. of e.g. Lithuanian to the outgroups is replicated using contrib from each in their respective weights, aka lithuanian is a weighed avg of the 'positions' of the rest, as described in the paper.

If we think about this carefully, increased EHG should shift pops closer to Yamnaya strongly to the exclusion of the others, because the shared drift btw Yamnaya and EHG occurred in the branch between the EHG in Yamnaya and EHG in Karelia, aka they are pretty much an extremely solid clade w.r.t. all the outgroups. This should outweigh any kind of increased drift on the branch btw Karelia and loschbour, or through that Karelia and E neols, even though the Yamnaya are admixed. So EHG admix should increase the rel dist similarity strongly with Yamnaya first, and WHG and EEF only weakly/incidentally compared to yamnaya.

Because the pops are shifted closer to yamnaya in terms of the rel dists to the rest of the 15, the contribution of Yamnaya will have to increase to account for the rel dists of the modeled pop.

Even if it also shifts them closer to WHG and through that EEF, the increased arithmetic contrib of Yamnaya simply precludes the increase of the rest.

Davidski said...

R1a should be in Dnieper-Donets.

http://en.wikipedia.org/wiki/Dnieper%E2%80%93Donets_culture

But R1b somehow made it into late Yamnaya near the Balkans, and then moved into the Carpathian Basin.

Btw, Alberto, the West Asian component is just ANE + ENF + local drift in the Caucasus.

Alberto said...

@David

"Btw, Alberto, the West Asian component is just ANE + ENF + local drift in the Caucasus."

So you think that if you isolated ENF from "West Asian"-free populations (like EEF) it wouldn't make any difference? Who knows, but never mind then.

Davidski said...

I'd just get ENF I think.

Unknown said...

Yes, the yfull snp and str have the L23 MRCA at 6400YBP. That makes it highly illogical to put L51 in a branch that develops far from Z2103. There is no migration from West Asia to Western Europe in the archaeological record during that period. The split of L23 happened in the east. Our L23* Yamnaya is more evidence, along with Beaker aDNA.

Chad said...

I'd bet on here being where Corded began to take shape.

http://en.wikipedia.org/wiki/Fatyanovo%E2%80%93Balanovo_culture#mediaviewer/File:Fatyanovo-culture.jpg

Unknown said...

Dave what are ur estimates of ane in Armenians; Pakistani; Indians and Iranians ?

pnuadha said...

@chad

R1a will not be dominant in SW Yamnaya, but I will bet that it is West of Samara, on into Belarus.

If R1a Corded was from SW Yamnaya it should have a good amount of MNE/Balkan ancestry, but it doesn't. Bell Beaker does though!! And it has too much Yamnaya to be derived from Corded Ware. That would make BB over 2/3 Corded without sharing much in uniparental markers... not likely.

I will put money on L51 being found around Ukraine. New SNP dating would put it in the Sredny Stog timeframe. It should be there.... Those Kemi Oba Kurgans end up west of the Carpathians, while Corded was never found south of the Carpathians. Those Kurgan groups show influence on Early Mako, Vucedol, and Bell Beaker. This group is distinct from Corded Ware, and as mixed Balkan elements. It doesn't fit Corded, but Bell Beaker R1b.


not trying to stir the pot but why do you see a need to place r1a in part of the yamnaya? Do you not think that IE related r1a (in Europe, India, Persia, and Tarim) could all be derived from Corded Ware? If you think that, then what IE related r1a migration can't be derived from corded ware? The tarim mummies lack samaran r1b so I find it less likely that tarim mummies actually derive from yamnaya.

I get that the corded ware autosomal dna is so close to yamnaya dna but isn't it reasonable to assume yamnaya's northwest neighbors would have a similar composition before some hypothetical migration?

Chad said...

I think that R1a looks like it formed an arc around R1b. I don't think R1a outside of Europe came from Corded, but Corded is just one group of R1a people. How R1a fits into Yamnaya and IE is something that needs to be figured out with more sampling. Just like the case for L51.

Chad said...

You guys need to remember that these R1b Yamnaya are just elite men. I would bet that if flat graves were found and tested that we would find those "Neolithic" haplogroups like G2a, with other R1b's and maybe a couple R1a. There was more than 7 males living between 3300-3100BCE, in the region. Trying to say that Z2103 is from West Asia, because it's dominant there, is as ridiculous as saying that L51 started in the Atlantic fringe.

Krefter said...

@Davidski,
"So whatever is causing Oetzi to show a bit of this admixture might be inflating it among the Norwegians and Scots."

No non-Yamna ANE-heavy component also screws results.

Also, it's totally possible that Norwegians and Scots can be fit as more Yamna than Balts. You might find this when you do your own analysis. More ANE doesn't mean more Yamna.

Krefter said...

"So CW has Lithuanian, Icelandic, Norwegian, Orcadian, as tops, and finnish drops all the way. Unetice has Lithuanian, Norwegian, Icelandic; finnish is close to bottom. LBA has Lithuanian, English, Icelandic, Norwegian; and finnish isn't even high enough to appear on the list."

I thoroughly took nots of the whole paper and never remember where drift stats or F-statistics were used to compare Unetice and LBA to modern Euros.

Where did you find that?

Unknown said...

Colin
It'll be no secret that I agree with u
In which case , then, the entire fragile hiuse of cards that is the Kurgan hypothesis falls apart .
Quote clearly ; whilst we are definitely seeing migrations ; they are not from an Yamnaya epicentre ; but more regionally confined : R1b west steppe<-> west Asia ; R1a Z280 within EE itself (internal migrations ); R1a-Z93 in Central asia - Tarim - Andronovo

Unknown said...

@ Chad

"I think that R1a looks like it formed an arc around R1b. I don't think R1a outside of Europe came from Corded, but Corded is just one group of R1a people. How R1a fits into Yamnaya and IE is something that needs to be figured out with more sampling. Just like the case f.."

Glad ur taking my revered advice on board :)

Chad said...

Mike,
We just don't know yet. We could sample 5 guys from 50 miles south of Samara, and all are R1a. Then sample 50 miles south of that, and get all R1b again. I think the only inference that is feeling pretty safe is that L51 will be there. Sampling is the key. Someday, we will have 50-100 of them, and it will be pretty clear what happened.

Unknown said...

Of course

Krefter said...

Anyone going to discuss Sup info 7-10 which is the heart of the Haak 2015?!

You're just going in circles with the Y DNA. And the ADMIXTURE part is pretty much useless, and I don't see much reason to waste time looking through it.

Chad said...

The significant statistic was f4(Loschbour, HungaryGamba_HG; LaBrana1, Chimp) = 0.00429 (Z=5.6) which suggests some structure within the WHG population, as LaBrana1 shares more alleles with Loschbour than with HungaryGamba_HG. This statistic continues to be significant when limiting to transversions (Z=3.9).

Hmm.. I want to look for an EHG and Outgroup cline possibility in WHG.

Krefter said...

@Chad,

I agree Loschbour being part EHG is possible. The trees in Figure S8.7-9 show this is possible.

I'm surprised EHG doesn't fit very nicely as MA1+Loschbour(S8.1, S9.14c, d) and that Karelia_HG is as related to Karitiana as MA-1 is. Something contrary to the WHG-ANE model is happening or we have bad references.

When Motala12 was fit as EHG+WHG and La Brana-1 was the WHG reference Motala-12 scored 14% more EHG(S9.14a-b). This could be because Loschbour has EHG.

I wish Haak 2015 would have tested whether Karitiana are more related to Loschbour or La Brana-1. If they're more related to Loschbour that suggests Loschbour has EHG-type ancestry.

The ANE-WHG model not being perfect isn't going to change much about the origins of Europeans because....

>Loschbour, MA-1, and EHG all appear to lack any form of basal and east Eurasian and trace 100% of their blood to the same "West Eurasian" branch. So, ANE-WHG ancestry vs ENF or basal won't change.

>Now we can use EHG and Yamna as "eastern" references instead of MA-1, and they'll show similar patterns as MA-1.

pnuadha said...

@Mike

We might agree that Eastern Europe received its, now IE related, r1a more locally rather than the steppe but after that our theories seem to diverge.

First off, I think its pretty clear that the closely related forms of r1b-L23 found in modern West Europe, Turkey, and Altai derive from yamnaya migrations that carried the earliest IE languages into the region. That is, a single cultural horizon carrying the common language and dna, which connected regions like Ireland and China, which otherwise share little to no connects. Id say that fits with what I imagine as the kurgan hypothesis.

I also see a similar type of origin for the IE related r1a found in Europe, Central Asia, and India having been derived from one or a few East European cultures, closely after the emergence of yamnaya. I actually think these cultures were not the yamnaya themselves but that not really the point. Poland and India connecting their ydna and language back to an East European Cultural groups is the point.

Overall, I agree that there is more complexity then there being one single homogenous group PIE riders spreading their language and dna in a consistent proportion. But I don't think anyone with much intelligence would have thought that. I also agree that there was differentiation within the (likely PIE) yamnaya realm in terms of language and versions of similar ydna. But I supposed these things before the recent paper and the findings of ANE. Again, that basically seems like the kurgan hypothesis to me.

The special things about these recent papers is not only that they confirm IE related migrations from eastern europe but they also confirmed that these migrations were quick, far reaching, and massive in terms of how much they changed the autosomal landscape of europe. Before these papers we could only guess that actual migrations carried PIE derived languages from the PIE homeland and we could only guess how direct the impact of these migrations were on lands distant from the PIE homeland. Now we know that the migrations were strong and far reaching and not some slow dissemination process that went through a bunch of local filtering, i.e. ireland.

Krefter said...

@Chad,

Laz 2014 showed that Moatal12 is much closer to Loschbour than to La Brana-1. I think the same may be true for K01(Hungary) but I haven't seen any tests to prove it.

If La brana-1 is significantly closer to Loschbour than to K01 that's big news, because Loschbour is about as close to both of them. This may mean Loschbour is a mix of something K01-like and La Brana-1 like.

EHG may be the reason for this diversity within WHG.

Chad said...

I think they all have EHG, but I think Loschbour may have more than the other two. I guess we'll see after the genomes are released.

Krefter said...

@Chad,

How do we work this out with Y DNA though? Obviously WHG was mostly I and C-V20 and Q-R can be associated with ANE-EHG affinity. If Mesolithic central-west Euros have EHG they probably had some P.

However MA-1, EHG, and WHG are admixed or not admixed we can pretty confidentially they form a clade when compared to east Asian.

Unknown said...

Ruki

Thanks for ur comments

But "europeans in Haak modelling, since the only source of EHG is gonna be from Yamnaya.." I don't think so. Not to harp on , but baltic mesoliths were excellently adapted to their niche and obviously out-survived farmer newcomers. I must insist; unless proven wrong in future studies, in their role in repopulating eba Europe; at least partly. We cannot exclude this just because they allegedly weren't pateilocal and warrior ethotic "IEs".

I'd love u or Dave to try to show us that 4 way model u suggested ; including EHG separate to yamnaya ?

Grey said...

If ancient human populations mapped onto biogeographical regions (which seems likely to me) then possible R1a / R1b distributions might be hinted at on maps like this especially if the imagine the boreal zone was a lot further south in those days i.e. a (roughly) north / south distribution.

http://www.eea.europa.eu/data-and-maps/figures/biogeographical-regions-in-europe-1/map_2-1_biogeographical-regions.eps/image_original

.

@MT
Looking at the graph of of Yamnaya components in modern populations, why is it that it peaks in Scandinavia, Baltic, and even Scots over Ukrainians and Bulgarians, who should have absorbed the very first impact of the "Yamnaya invasion" ?

http://besuccess.com/wp-content/uploads/2014/04/Big-Fish-Eat-Little-Fish.jpg

.

The good argument against the hunter -> pastoralist via raiding idea is that hunters like being hunters and there's lots of example of hunters rejecting any kind of transition.

The thing is though, if you think on it, if the hunters become mounted hunters and then steal domesticated animals from farmers and captives to herd them then the hunters themselves never turn into pastoralists.

The hunters turn into a mounted warrior elite who spend all day hunting on horse back while other people grow/herd their food for them.

Mounted warrior elites who spend all day hunting while other people do all the work - I'm sure that rings a bell somewhere.


Alberto said...

@Krefter

"Loschbour, MA-1, and EHG all appear to lack any form of basal and east Eurasian and trace 100% of their blood to the same "West Eurasian" branch."

Absolutely.

Notice (Figure S7.4(c), page 63) that Samara_HG is closer to Yamnaya and Loschbour than to CW (even if CW has more HG than Yamnaya). This is precisely because CW also has some Basal Eurasian, that pushes it away from a pure EHG. Apparently Yamnaya has less Basal Eurasian, which means that the Armenian-like pop didn't bring much basal with them.

And also thanks for encouraging more discussion about heart of the paper. So much y-dna debate when that data is inconclusive is not that productive IMHO.

Chad said...

I think that ydna will be a mix. I think we will see F,C,I, and maybe Q, and R. I'm not too focused on that, as some probably pre-date the Basal West Eurasian break.

I read something a while back about new lithic technology coming from around the Aegean around 8000BCE.. So, we may have two out groups, with EHG. EHG itself is probably two to three things.

It will probably be nuts. We may need a K6-8, just for the West Eurasian part.

Chad said...

KO1's R3 mtdna is another hint, if you ask me.

Unknown said...

Colin and Dave

It seems that for you close enough might be good enough; for me the devil is in the detail.;)
It's not good enough to point to broad genetic and cultural similarities and say "they're all interrelated IE groups". Simply ; that grossly underestimates the sociolinguistic complexity and diversity of Copper Age EE/ west Eurasia; and falls back to 1930s pots= people theory which has been proven wrong time and again. Even if you insist in arguing through such archaic approaches; then nevertheless CWC has big differences to yamnaya

It's like saying modern Germans, Spanish; French and Irish and even Basques are one and the same people ; recently derived from a proto -Western European population which lived 1000 years ago; jus because they're all R1b and share in a catholic-dervived ; western culture .

Perhaps too much has been made of ANE. No doubt big shifts were happening in at least some regions; but ANE underplays far older population events in central-western Eurasia which have nothing directly to do with, and far ante-date , PIE.

The big shifts in EE might certainly have something t do with PIE; but maybe not ; especially if the shift might have different proximate source populations (R1a vs R1b; which must have been long diverged).


Ultimately I guess definite proof will lie in studies of Mycenaeans. We KNOW they spoke IE. IF we have clear evidence of arrival of Steppic DNA; then the matter ends. But I suspect this is NOT what the data will show :)

pnuadha said...

It's like saying modern Germans, Spanish; French and Irish and even Basques are one and the same people ; recently derived from a proto -Western European population which lived 1000 years ago; jus because they're all R1b and share in a catholic-dervived ; western culture .

Ive never been that crazy about the details, they can leave you lost and skrew up your predictions. For example, I only know a tiny fraction of what other avid posters know when it comes to dna and culture. Thats actually putting it really mildly. My knowledge of the details when it comes to the ydna cladistics and cultural practices are quite shallow. But that does not matter. I have still made a far better prediction on r1b and yamnaya than most of those who know far more than me, even those who favored r1b to be in the western yamnaya.

I think its all about knowing what to look for and knowing how to put it together.

My talk on the spread of IE is more about predictions and knowing what produced the end result with is modern languages and our modern dna. Ive maintained that corded ware likely had genetic and linguistic diversity. But all I said in the previous post is that CW, not necessarily all of it, was a likely launching point for r1a and IE in the east. See, so I'm not really trying to describe the nature of the CW, I'm more so trying to figure out the basic role they played in the IE problem.

pnuadha said...

Mike, now the real question is...

Who the H is Dave?

pnuadha said...

It's like saying modern Germans, Spanish; French and Irish and even Basques are one and the same people ; recently derived from a proto -Western European population which lived 1000 years ago; jus because they're all R1b and share in a catholic-dervived ; western culture .

Its certainly true that a culture does not dictate the dna, and certainly the dna does not dictate a culture. I would have been much more favorable to your strong cautioning of associating certain sects of a culture with a migration. However, it is truly amazing how much archeological cultures can mirror genetics. The most shoking realization I had on this issue is when swedish farmers who had been living next to hunter gatherers for many hundreds of years were very distinct genetically, i.e. they did not mix much. That proved to me that culture, as we define it from an archeological perspective, can often be viewed as a genetic barrier and in the right context a genetic marker. Everything we have seen since then is a remarkable correlation between culture and genes.

I am NOT saying that the two are synonymous, but in the right context they overlap enough to the point that you can reference either when it comes to a migration.

Unknown said...

Dave= Davidski
We discussed similar issue earlier

Fanty said...

"and falls back to 1930s pots= people theory which has been proven wrong time and again."

Hmm...
And I thought we are already back to that.

It felt like:

1930: "Pots are people, because people who had different pots had also different skulls."

1990: "Pots are NOT people. Its all learned technology only and even if the skulls are differently shaped its the new lifestyle that did change those skulls but no migration!"

2011: "Pots ARE people, we checked their genes... sorry 1990 :P"

Unknown said...

Colin
I too am surprised
I must admit
The homogeneity of Central european farmers; the HG continuum; and the massive BA shift are startling and fascinating - I have to admit
But we need to figure the details
It might well be that the Samara samples were flukishly all R1b; and there is in fact both R1a and -b derived lineages in both; and simply show differential spread due to drift acting on Y-DNA.
But we might get more curve balls.
But to me it's striking that we already see clearly differentiated zones -Z280 EE, R1b Central eurasia-central/west Asia , z93 Central-eastern steppe -Central Asia

Unknown said...

Ok Fanty
Pots don't ALWAYS = people, then

Chad said...

David,
Check out Pg 23. It's actually Bell Beaker that is more Yamnaya than Unetice. Bell Beaker is only West of them due to more MNE ancestry. That wasn't expected. Sweet! That is the resurgence in local ancestry after Yamnaya, they were talking about.

Davidski said...

I'll believe it when I've tested the genomes myself.

Chad said...

Kostenki kind of looks like something that is 2/3 MA-1, 1/3 Ust-Ishim. I'm not sure what to make of that. I'll have to go over other supplemental stuff.

Chad said...

You'll probably have to make a Yamnaya component for that. An EHG one might be misleading.

Davidski said...

The first thing I'll do is run all the main samples with the K8. I reckon the K8 ratios and PCA will sort everything out.

The most Samara Yamnaya-like will be the Erzya, Moksha and Chuvashs from the Volga-Ural IMO.

After that, for a bit of fun, I'll play around with EHG and Yamnaya components.

Unknown said...

Yeah. Some work on WHG and EEF is in order.

Chad said...

I've got some more ideas. I'll send an e-mail.

Unknown said...

So Dave yiure going to try to flesh out separate EHG components for reichs data ?
Also; how would u account for the finding that Usty and Kostenki hug modern Caucasuan / west asian populations ? To me this suggests preexisting ; considerable levels of ANE there

Chad said...

Everything I'm asking him to look for could take a couple weeks to tease out. It's going to involve f4s and all that. Hopefully, it works.

Helgenes50 said...

The Major Y-Chromosome Haplogroup
R1b-M269 in West-Europe, Subdivided by
the Three SNPs S21/U106, S145/L21 and
S28/U152, Shows a Clear Pattern of
Geographic Differentiation


http://www.scirp.org/Journal/PaperInformation.aspx?PaperID=53754

Davidski said...

Mike,

You're taking the PCA results too literally.

Neither Kostenki14 nor Ust'-Ishim can be described as partly ANE, although it's likely their close relatives gave rise to the ANE genetic signature about 25,000 years ago.

Krefter said...

@Davidski,
"The most Samara Yamnaya-like will be the Erzya, Moksha and Chuvashs from the Volga-Ural IMO."

I've thought this for months. Non-Slavic Russians will probably turn out to mostly be Yamna-types who admixed with central Asians and Siberians.

Ironically Saami the indegnoious butterflys of Europe will be very similar to the IEs who conquered Late Neolihic/bronze age Europe.

Helgenes50 said...

@David

I posted a link above

Are these results still valid ?

Davidski said...

Nope, looks outdated. The author seems to think that M269 is a Neolithic marker in Europe.

Unknown said...

R1b must have originated somewhere *south* of the Caspian. For if it is already in the Neolithic Spain, then an origin in Eastern Europe is impossible: M-343 definitely is West Asian, as is P-297.

The expansions of later groups like Z-2103 and and L-51* depend on where M-269 began, and that we don;t really know.

Whatever the case, R1b is of southern origin. The question which remains is from where and when exactly L-51 entered Europe and how it was related geographically to Z-2103.

Davidski said...

How do you know that the R1b carried by the Neolithic Spanish farmer isn't due to...

1) Eurasian forager admixture in Europe?

2) Eurasian forager admixture in West Asia?

The ancestral line to R1b-V88 probably entered West Asia with Central Asian forages very early, so it might have had time to get caught up in Neolithic movements to Europe.

It's very difficult to see the relevance of these sorts of basal stragglers far away from Eastern Europe, which looks like the main bifurcation zone for R1.

R-M417 and R-M269 must be the main Kurgan markers, and all we need for them to be native to Eastern Europe has already been located in Eastern Europe.

Unknown said...

To answer your questions, Dave, I think: R1b can not have arisen in eastern Europe.

First of all, simply look at the major three branches of R1b-M343:

- M335 is restricted to Anatolia
- V88 is restricted to the Levant & Africa.
- P-297, from which derive M-73 and M-269 are found both north and south of the Caspian.

And we have ancient upstream lineages found already in wide-spread regions: Samara and Spain; with wholly different autosomal make up.

So you tell me where R1b began ?

"Stragglers', and 'lost' foragers don't really cut it, especially given that what your proposing is impossible.

Davidski said...

R1b began on the Eurasian steppe, just like R before it.

Unknown said...

I can't disprove what your saying; but can only ask what groups specially adapted to the north eurasian forest foraging cultures were doing wandering South. Ie to completely different social, economic and cultural environments ? - chasing a mammoth ? :)

Unknown said...

Also: are you familiar with the phylogeny of Hg K ?

Davidski said...

I have no idea what if anything they were chasing.

https://drive.google.com/file/d/0B9o3EYTdM8lQb1JScVI1TXpxaTA/edit?usp=sharing

http://www.springer.com/social+sciences/anthropology+%26+archaeology/book/978-1-4614-2002-6

pnuadha said...

To answer your questions, Dave, I think: R1b can not have arisen in eastern Europe.

First of all, simply look at the major three branches of R1b-M343:

- M335 is restricted to Anatolia
- V88 is restricted to the Levant & Africa.
- P-297, from which derive M-73 and M-269 are found both north and south of the Caspian.

And we have ancient upstream lineages found already in wide-spread regions: Samara and Spain; with wholly different autosomal make up.

So you tell me where R1b began ?


Gotta be honest, this reasoning makes no sense. It's also surpassing to hear this type of language from you.

Simon_W said...

There was this Epi-Cardium sample from Catalonia: 5 times G2a and one E-V13. And there was Treilles, a post-Cardium sample from southern France with 20 times G2a and just two I2a1. The new Spanish EN sample from Aragon however is quite a bit in the hinterland of the initial Cardium settlement and a few centuries later. So assimilation of local foragers isn't unlikely, and this small sample has no G2a. It has one F, one I2a1b1 and this R1b1.

A lot depends on the origin of R1b1a-P297 in Yamnaya. The possibility has been raised that this was from the West Asian element in Yamnaya. Its predecessor R1b1 was present in an Iberian EEF, in an EHG from Samara, and quite possibly it was also present at an early date in West Asia, since R1b1b and R1b1c are typically West Asian. But I think the most obvious explanation is that R1b1a-P297 in Samara was descended from the R1b1 in Samara. Same place, and just one mutation away. Also, if the IE element had arrived on the steppe from West Asia, linked with R1b1a-P297, the Corded people should have R1b1a-P297 as well, otherwise they couldn't be IE too. The presence of R1a1 in a Karelian EHG pretty much refutes the recently proposed hypothesis that R1a originated in West Asia. The geographical distance alone makes this unlikely, but it's even more unlikely since that EHG has no detectable West Asian autosomal admixture. And so far no Corded male carrying R1b has been found. On the other hand, pre-proto-IE might have developped among EHG groups who were in mutual contact, then there wouldn't have to be one haplogroup dominating the other.

Some have suggested that the predominant R1b1a2a2 in Yamnaya came from West Asia. But its direct ancestor R1b1a2a was also present in Yamnaya, as was R1b1a. So the only unattested link is R1b1a2, which may easily have arised from the local R1b1a. Also I note that they don't have the typical West Asian R1b-L277, but its ancestor. And this is exactly what would be expected if IE languages in West Asia were derived from the steppe. Also it has been criticized that this Yamnaya sample doesn't have R1b-L51, which is typical for Europe. But it does have its direct ancestor, R1b-L23. Quite possibly L51 didn't yet exist back then, but it was just a mutation away.

Unknown said...
This comment has been removed by the author.
Unknown said...

Simon : Your reconstruction of M-297 is certainly valid; but I stated that it could have risen anywhere north or south of the Caspian, in fact, I'd argue that both regions have been linked. R1a, is certainly northern, as Ive stated previous times, in an arc since the Mesolithic.

But the overall phylogeny of R supports a relatively more southern origin. Im not talking about the Levant, or India; but central Asia.

Simon_W said...

Yes, central Asia is a serious possibility.

Also I'll admit that the R1b1 in the Iberian farmer might just as well have been descended from West Asian EEFs, it may have been rare but present in Cardium people, like E-V13. We don't know.

Simon_W said...

I've counted and calculated: The overall mtDNA evidence from Yamnaya known so far seems to be roughly 45% West Asian and 55% EHG, if we count all the H as EHG. So it's not really possible to deduce the origin of the yDNA from this evidence and the autosomal makeup. One side (EHG or West Asians) seems to have predominated the uniparental markers, but it's not deducible from the mtDNA which one.

Unknown said...

"Central Asia' refers to regions both north and south of the Black Sea - Caspian line.

Davidski said...

Online Table 1 is interesting. There's a whole bunch of Yamnaya and also Poltavka (supposedly proto-Indo-Iranian) samples that didn't make it into the paper, probably because they didn't offer enough markers. But I bet some of their Y-haplogroup results are known to the authors.

Davidski said...

The Poltavka remains would have to be R1a-Z93.

Richard Rocca said...

I don't know if you guys misssed it or not, but here is an L23(xL51) proxy map for R1b-Z2103 with the position of the R1b Yamnaya samples and the extent of Yamnaya. As you can see, plently of space left for R1a and R-L51 to found in Yamnaya...

http://www.r1b.org/imgs/Z2103_Yamnaya.png

Unknown said...

Excellent book on Mesolithic Russia and EE.

https://www.academia.edu/5601712/Dolukhanov_Sarson_and_Shukurov_Eds_The_East_European_Plain_on_the_Eve_of_Agriculture_BAR_Int._Ser._No._1964_2009_246_pages

Unknown said...

Excellent book on Mesolithic Russia and EE.

https://www.academia.edu/5601712/Dolukhanov_Sarson_and_Shukurov_Eds_The_East_European_Plain_on_the_Eve_of_Agriculture_BAR_Int._Ser._No._1964_2009_246_pages

Davidski said...

That's a nice map Richard.

Volodymyr Lutsyk said...

R1b forager comes from Elshanskaya culture. It was the first ceramic culture of the region. The spread of the culture is connected with the influx of new population from Central Asia.

Romulus the I2a L233+ Proto Balto-Slav, layer of Corded Ware Women said...

@Simon

The I2a1b found alongside the R1b1 in Spain is acutually I2a1b1a L161.1 which is in the present day found most commonly in the Celtic areas of Great Britain, kind of implies a spread with the Beakers.

Unknown said...

Mike,
M335 is more common in some Central Asians than Anatolia. We've posted than before. It probably came with Turks.

Unknown said...

As I closely look at Fig 2 and Fig S5.1 in the PC plots from Haak et al, I'm noticing that the Cardial Neolithic samples from Els Trocs in Spain cluster further "south" (low values for PC dim2) that the EN LBK and Starcevo samples. Could an unknown Mesolithic sample from the Central Mediterranean--like the Castelnovian horizon--be drawing these samples away from the LBK/Starcevo cluster? Another possibility is that the Cardial Neolithic is demographically much less dense than say LBK and thus is more influenced by drift. Also, note that the North Africans--Berbers, etc...Are on the admixtures spot on with the Sardinians except for some Sub-Saharan admixture. This result also might suggest a South Mediterranean route to the Cardial Neolithic and the R1b1 samples found at Els Trocs.

Nirjhar007 said...

@ Volodymyr Lutsyk
''R1b forager comes from Elshanskaya culture''
Can you give a link on that culture?
@Wesolowski
''There's a whole bunch of Yamnaya and also Poltavka (supposedly proto-Indo-Iranian) ''
IIR Poltavka is as possible as Uralic originating in India.

Unknown said...

Nirjhar, people say it's either local or from the east side of the Caspian.

Marnie said...

@Volodymyr Lutsyk

"R1b forager comes from Elshanskaya culture. It was the first ceramic culture of the region. The spread of the culture is connected with the influx of new population from Central Asia."

Thanks! I would be intrigued to know why you think this culture is associated with R1b hgs.

I found this great paper on the Elshanskaya culture:

"The problem of the neolithisation process chronology in Povolzhye"

Authors: Vybornov, Kulkova, Goslar, Possnert

From the paper, the earliest dates (from charcoal at the Kairshak III site) are 6950+/- 190 BP).

From pottery at the same site:
7780+/-90 BP).

Formation of the Elshankaya culture is placed in the second half of the 7th millennium BC, at earliest, according to this paper.

It seems that some researchers do not agree with these early dates.

Nirjhar007 said...

Can Be there is also the BMAC effect anyway it by no means can be called as IIR specially if some one studies their(IIR) culture practically.....
See You Guys in 9 hours.....

Krefter said...

"Can Be there is also the BMAC effect anyway it by no means can be called as IIR specially if some one studies their(IIR) culture practically.....
See You Guys in 9 hours....."

Americans have the day off today. Suckers!!!!

Marnie said...

Thinking about Volodymyr Lutsyk's comments, it's interesting to look at early pottery dates in the Balkans.

From this set of conference papers:

http://www.academia.edu/8629431/Tracing_pottery_making_recipes_in_the_Balkans_6th_4th_millennium_BC

See, for instance, this paper:

The first potters from Bulgaria (6200-6000 cal. BC): interpreting diversity.

So, at 8,200 BP, pottery in Bulgaria would pre-date the earliest pottery of the Elshankaya culture by at least 400 years.

And, as you should all know by now, the greatest haplogroup diversity for R1b is in Bulgaria, according to "Chad".

So the idea that R1b hgs definitively "orginates" with the Samara culture seems to have a bit of a problem.


«Oldest ‹Older   201 – 400 of 594   Newer› Newest»