search this blog

Friday, April 3, 2015

The teal people: did they actually exist, and if so, who were they?


The ADMIXTURE analysis in Haak et al. 2015 includes a series of intriguing teal colored components from K=16 to K=20 (see image here). The main reason I'm so intrigued by these components is because they generally make up over 40% of the genetic structure of the potentially Proto-Indo-European-speaking Yamnaya people.

But there's only so much one can learn by starring at a bar graph, so I thought I'd have a go at isolating the same signal with ADMIXTURE to study it in more detail. You can view the results of my experiment in the spreadsheet here.

I wasn't able to completely nail any one of the teal components from Haak et al., because I don't have access to all of the samples used in the paper (I'd have to sign a waiver to get them). Nevertheless, the signal looks basically the same.

Below is a bar graph based on the output featuring selected populations and ancient genomes from Europe and Asia. The Fst genetic distances between the nine components are available here.

Note that the teal component peaks in the Caucasus and the Hindu Kush, and generally shows a strong correlation with regions of relatively high MA1-related or Ancient North Eurasian (ANE) admixture. On the other hand, the orange component peaks among Early European Farmers (EEF), who basically lack ANE.

To learn about the structure of the three main West Eurasian components - blue, orange and teal - I made synthetic individuals from the P output to represent each of the components, and tested them with my K8 model. As expected, the teal component harbors a high level of ANE, while the orange component lacks it altogether. Refer to the spreadsheet here.

It's very likely that the teal and orange components from Haak et al. share these traits. I think this is more than obvious by looking at their frequencies across space and time in Eurasia.

I also analyzed the synthetic individuals with PCA based on their K8 ancestry proportions. The samples representing the orange component fall just south of the Stuttgart genome from Neolithic Germany, and this is basically where I expect Neolithic genomes from the Near East to cluster when they become available.

Interestingly, the samples representing the blue component are dead ringers for Scandinavian hunter-gatherers (SHG). However, I suspect this is something of a coincidence caused by the small number of Western European hunter-gatherer (WHG) and Eastern hunter-gatherer (EHG) genomes in the dataset. The algorithm probably doesn't have enough variation to latch onto to create both WHG and EHG components, and in the end settles for something in between, which just happens to resemble SHG.


But the fact that the orange and blue samples more or less pass for ancient populations leaves open the possibility that the same might be said for the teal samples.

So did the teal people actually exist, and if so, who were they?

My view at the moment is that a population very similar to the teal samples formed in Central Asia or the North Caucasus during the Neolithic as result of admixture between MA1-like and Near Eastern groups. This population, I believe, then expanded into the Pontic-Caspian steppe by the onset of the Eneolithic.

Were they perhaps the Proto-Indo-Europeans? Probably not. I'd say they were Neolithic farmers who eventually played a role in the formation of the Proto-Indo-Europeans. In any case, someone had to bring the Caucasian or Central Asian admixture to the steppe, and I have it on good authority that it was already present among the Khvalynsk population of the Eneolithic, albeit at a lower level than among the Yamnaya of the early Bronze Age.

Citation...

Haak et al., Massive migration from the steppe was a source for Indo-European languages in Europe, Nature, Advance online publication, doi:10.1038/nature14317

Update 16/11/2015: 'Fourth strand' of European ancestry originated with (Caucasus) hunter-gatherers isolated by Ice Age

640 comments:

«Oldest   ‹Older   201 – 400 of 640   Newer›   Newest»
Simon_W said...

Some Proto-Uralic words:

*kota, clearly related to PIE *kata, both denoting cabin.

*wete, denoting water, related to the PIE word *wodr.

*ki, signifying who, compare to Latin quis or PIE *kwis.

Compare the Finnish verbal endings to IE ones, here the verb laulaa (to sing):
Laulan, laulat, laulaa, laulamme, laulatte, laulavat.

Note: In Proto-Uralic the first person singular ending was -m, exactly like in PIE.

I think intimate similarities like these don't arise all of a sudden from contact between local HGs and a foreign intrusive herder population. They're much more likely to arise as a result of longstanding mutual contacts between small groups of HGs.

The Hurrian verbal system in contrast was much more complicated, and its endings had absolutely no similarity with IE endings:

E.g., for verbs with -i suffix:

-af / -au, -io, -ia, -ausa, -asso / -assu, -ia.

Some shared vocabulary may be possible, but that's about it.

So, I think this speaks in favour of a PIE homeland on the steppe.

But at the same time I think from a genetic perspective it would really make sense if the Greeks and Albanians were from West Asia, and the Armenians not from the Balkans. But how to reconcile these two positions?

It occured to me today, that quite possibly, the Kura-Araxes culture may have been at the root of the Albanian-Greek-Armenian branch of IE. The earliest nucleus of the Kura-Araxes culture was between Erzurum in eastern Anatolia, the Kura-Araxes lowlands, and the Caspian coast and nearby slopes of northeastern Azerbaijan and southeastern Dagestan. It's not compelling to associate this area with other linguistic groups, I'd say it's quite possible that the above said IE branch had spread from there. And ultimately from the southern steppe close to the Caucasus, and hence probably didn't have a lot of EHG ancestry from the beginning.

Gaspar said...

@simon,

yes, this is what I should have said,

firstly the west-asian in western Sicily is from the Phoenicians/Carthaginians and Anatolian "Greeks". Corinthian Greeks only settles in the eastern part of Sicily.

Corinthians came from Doric origins, Corinthians are the main Greek group that settled in Italy and where as far north as Ancona

Maju said...

@Karl: There isn't anything homogeneous that we can call "Subsaharan African" (regardless that I hate the term "subsaharan", so incorrect in all aspects): all Humankind (H. sapiens) is "Subsaharan African" for the first 100 Ka or so, about half the paleohistory of the species, and that includes many branches of which Khoisan and Pygmy are just two. From one of those early Human branches a sub-branch split that migrated to Asia (India, East Asia - I'll ignore the intermediate Arabian stage by the moment, as it probably left only very limited and local legacy) and some people of this Asian branch migrated West to the "Neanderlands" some 50 Ka ago. This branch of India and beyond is what Lazaridis et al. call "Eurasian" (and fairly so) and it is also today the most numerous branch of Humankind, existing not just in Eurasia but in Australasia, Native America and even having back-migrated to parts of Africa (North and The Horn in essence).

"Basal Eurasian" can be therefore anything between the branching of the Pygmy populations (second to diverge after the Khoisans) and the arrival to India and that includes mostly African branches, although it potentially also includes a branch that migrated to Arabia and Palestine and that is at the direct origin of the Eurasian branch. As I said above, the legacy of this "First Arabian" population, which shrank under the pressure of drought and Neanderthal expansion, as well as the backflow of (Eur-)Asians in the early UP, is probably too small to be detected unless you are very focused in your analysis (IMO it's not just zero but not too large either). In any case the methods used by Lazaridis (with the Mbuti as control group) could not discern between it and other early branches from Africa (which obviously survived much better). When the Dinka are used instead the results are contradictory but many do suggest a Dinka-like genetic influence in EEFs. This does not need to be Dinka but something Nubian-like or Paleo-Egyptian it is probably the case. The influence of the Nile cultures in Mesolithic Palestine (and also NW Africa) is well documented and has been suggested as being at the origin of Northern Afroasiatic language expansion (however in the case of Basque I see much more Nubian affinity than Afroasiatic, so it may have been ethnolinguistically complex). It is also a very logical and parismonious explanation for the spread of E1b-V13 to Europe.

Maju said...

@Krefter: indeed, "Basal Eurasian" or right away African admixture should always lower the ANE affinity in D-stats and such but, with the proper controls, this should be detected in Admixture or TreeMix analyses. One just have to run a supervised Admixture forcing pops. to fit, for example, Lochsbour, Yemeni Jews (who have no apparent recent African admixture) and Ma1. Or you can include also Dinka as another forced component, to control for (extra) NE African admixture.

I'm pretty sure that the ANE in West Asia will still be low (although slightly variable in clinal form because we can't expect perfect isolation of West Asians from Central Asians). What you cannot do is to run West Asians for Lochsbour vs Ma1 and expect the results to be meaningful: logically that exaggerates the amount of ANE because West Asians are not WHG and never were.

Maju said...

@Simon said: "The people who keep thinking that Basal Eurasian is closer to Subsaharans seem to believe that they stopped drifting/evolving as soon as they had left Africa".

Not at all. What happens is that anything genuinely "Basal Eurasian" (= "First Arabian") will always appear indistinct from something NE East African in a Mbuti vs Onge (or vs Dai) test, which is the test that Lazaridis et al. run.

It's a problem of test design: the result is necessarily unclear. In order to be reasonably sure, we'd have to use another more proto-Eurasian African control (for example the Dinka) instead of the Mbuti, who split apart quite some time before the OoA.

As defined by Lazaridis, the Dinka (and even the Yoruba) are in the range of "Basal Eurasian", there's nothing in that test that can discern them from some elusive "First Arabian" component.

In Skoglund's supp. materials there is complementary data and in some cases at least the Dinka do appear as donor of EEFs.

Alberto said...

Did someone figure out a way of measuring EHG admixture in Europe? I think it's quite a relevant matter, though not too easy to do because of the overlapping of EHG with WHG (for the HG part) and "West Asian" (for the ANE part).

Would qpAdm "just work" for this? Like trying to model European populations based on:

WHG
EHG
Armenian

So that for example Sardinians would be something like:

35% WHG
0% EHG
65% Armenian

And Estonian something like:

35% WHG
30% EHG
35% Armenian

Or would it require more complex methods (or an ancient "teal" genome instead of using Armenian)?

Davidski said...

Maju,

"What you cannot do is to run West Asians for Lochsbour vs Ma1 and expect the results to be meaningful: logically that exaggerates the amount of ANE because West Asians are not WHG and never were."

Then how come Starcevo_EN and LBK_EN don't show any ANE?

Are you suggesting their ancestors were pure Basal Eurasians when they landed in Europe, and all of their hunter-gatherer related ancestry is from Europe?

If not, then what are you suggesting? Did it just slip your mind that we had ancient genomes from Neolithic Europe?

Unknown said...

David
Pure speculation in my part; but I'd imagine (if they actually contributed) some south Balkan HGs would be almost pure Atlantic & west Med; and would sit between WHG and modern Spaniards.

Davidski said...

Modern Spaniards are shifted east because they have ANE, so continental European hunter-gatherers could not have clustered in a position between modern Spaniards and someone else, because they lacked ANE.

Moreover, there's no reason to think that any European hunter-gatherers had Near Eastern admixture until the Neolithic.

Maju said...

@David: "Then how come Starcevo_EN and LBK_EN don't show any ANE?"

If one of the references is EEF, they will cluster with EEF. ANE is not an absolute thing, but relative to comparisons.

"Are you suggesting their ancestors were pure Basal Eurasians when they landed in Europe, and all of their hunter-gatherer related ancestry is from Europe?"

No. I'm suggesting that they were something very similar to Palestinians when they landed in Europe. That includes some "Basal Eurasian" (African-like) original admixture and no or very low ANE (always depending with what you compare with, of course).

The West Asian (or "pre-European" or trans-Mediterranean) element in EEF is only a small part "Basal Eurasian" (African-like). That's something we know and that's more or less what we see still today in Palestinians and similar populations.

What I'm suggesting (and is very clear in Lazaridis' modelings) is that early West Eurasians split sequentially:

1st- ANE vs "core" West Eurasians and then these into:
2nd - West Asians vs Paleo-Europeans

That there are three basic meta-populations here, not just two. There is a West Asian meta-population that you guys are apparently ignoring and pretending it was just "Basal Eurasian", when that BE thing is just something extra and exotic and not the defining element of West Asians (neither ancient nor modern).

Paleolithic West Asians and Europeans were separated at least between Gravettian (or maybe even earlier, if you follow the European origin of Gravettian theory) and Neolithic, i.e. for some 25 millennia! Enough to make them two different populations, almost as different among them as either one relative to ANE.

Additionally West Asians had and still have some internal geographical structure, notably the Zagros vs Palestine (or Highlander vs Lowlander, N vs S) one, but also the extra African-like admixture which can get confused easily as "Basal Eurasian".

EEFs have (again per Lazaridis) a core that is West Asian (of mainline Eurasian origin, just as WHG or ANE but different from either), plus some exotic BE (African-like) and something WHG-like (UHG). The first two elements are still very apparent and defining in Palestinians (core and Bedouins), so they are roughly the same thing.

Krefter said...

Maju,
"Paleolithic West Asians and Europeans were separated at least between Gravettian (or maybe even earlier"

That looks like a high-estimate to me. I haven't done any archaeology research, but IMO it's common sense one way or another after 10,000s of years people in neighboring regions will mix.

No strong evidence in remains people in west Asia and Europe interacted before the Neolithic, doesn't mean they couldn't have IMO.

Davidski said...

Maju,

ANE is missing in WHG, EEF, and certain Bedouin and Saudi samples. These groups don't cluster together. So clustering together can't be used as an argument for lack of ANE.

EEF lack ANE despite being mostly of Near Eastern origin. That means their Near Eastern ancestors lacked ANE.

Palestinians don't lack ANE but they have some SSA, so Palestinians don't look like the Near Eastern ancestors of EEF. They're not even close IMO.

Unknown said...

@ David

"Moreover, there's no reason to think that any European hunter-gatherers had Near Eastern admixture until the Neolithic."

Well that depends on what you mean by "European".

My reasoning for my previous statement (a southern Balkan HG composed of mostly West Med and Atlantic components) is thus

* Starcevo, LBK from Hungary & northern Serbia have WHG admixture no more than 30%

* This 30% is attributable to Carpathian/ north Balkan HGs.

* south Balkan (ie "Greek') HGs must have been different.
(a) because of the isolation from north Balkan HGs (some 1000km of 'no mans land')
(b) Greek HGs show connections with Antalya (southern Turkey)

* so there was some kind of a Euro-Mediterrean HG group.

* this is tempered by the proviso that the Greek / sth Anatolian HGs actually contributed to EEF.

* Thu EEF can be more accurately modeled as WHG + GreekHG + Levantine (Natufian) farmers.

Davidski said...

All indications are that genomes of early Neolithic farmers from the northern Near East won't be very different from those of early Neolithic farmers from Europe.

Unknown said...

Dave
I'm not saying that near eastern farmers will/ should be different to early european ones
What i am saying is that there existed a yet different population in the near East- different from true Levantines and the later "West Asian" hybrids. Ie a euro-mediterranean one ( I guess similar to what Maju is saying).

Grey said...

@Maju

"What's the EVIDENCE behind this wild speculation?"

There's the EVIDENCE you yourself provide in the post *immediately* after the one quoted where you ask for evidence.


"B. Genetic data even: (1) clearly at least two distinct populations in West Asia (both autosomes- and Y-DNA-wise), one centered around the Zagros and the other around Palestine (again consistent with archaeology), (2) clear affinity of the EEF (Thessalian-derived Neo-Europeans) with the Palestinian rather than with the Zagros-Taurus-Caucasus core, (3) clear affinity of the West Asian component in Yamna with the Zagros component instead."


.

me
"3) The current idea is the Levant farmers were a mix of Basal Eurasian (BE) and WHG".

@Maju
"NO! That's a gross misinterpretation of Lazaridis' reconstructions."

Fair enough that was sloppily put on my part.

Let me try again as it's a very simple point but I seem to have a hard time getting it across clearly.

EEF in Europe is said to be a mixture of two components: WHG and to save quibbling, component X.

The important point is there are only two components which means logically - not evidentially, logically - there are only a limited number of options for the make-up of the source population in the source region of X.

They can (logically not evidentially) be 1) WHG, 2) WHG + X or 3) just X.

If we eliminate options 1) and 2) cos other evidence that only leaves (3), just X.

So far so obvious.

However if you add to that conclusion that a) farmers must have derived from pre-existing HGs and b) those HGs must have had a range then the previous conclusion leads to two other logical possibilities, either

1) the HGs that gave rise to the first farmers had a range that extended beyond the Levant

or

2) their range was limited to the Levant.

What does that mean?

1) If the range of HG population X extended beyond the Levant then the farmers with X dna moved into territory of HGs who also had X dna and the calculations for EEF aren't sound

or

2) if the range of HG X was restricted to the Levant (and/or maybe also down the coast to Egypt) then there must have been a significant barrier turning the Levant into it's own little island.

I'm easy with either.

Grey said...

@Simon_W

"The people who keep thinking that Basal Eurasian is closer to Subsaharans seem to believe that they stopped drifting/evolving as soon as they had left Africa."

Thing is either way it should be a distinctive signal and big clue.

Chad said...

Matt,

What is your email? I have some Dstats to show you.

Alberto said...

But I don't see what are you all discussing about. I think everyone agrees on the basic points:

- Ancient Near Easterners (from the Levant, or lowlands) were like EEF and lacked ANE (and maybe bits of European HG acquired in the Balkans).
- Ancient Middle Easterners from Iran (or highlands) were different, either becaouse of a more basal Eurasian or because they had ANE, or both.
- Modern Near Easterners (lowlands) are a mix of both.
- WHG had a stronger affinity with ancient Near Easterners (lowlands) than EHG did (and this is seen since Kostenki).
- EHG had a higher affinity with Iranians (or highlands) and South Asians than WHG did (and this is seen since MA-1).

I think all those things are clear and no one seems to argue against it.

Btw, I think the most Basal Eurasian component is probably the South Asian one. And not Onge, but ANI. The Fst distances show that the South Asian (light green) is the closest to Sub-Saharan. The oldest samples, especially Ust-Ishim, show a large chunk of it. And in Haak et al. Fst table, the closest Eurasian population to Yoruba is Sindhi.

So this is why I think that Iranians (highlands) had a more basal Eurasian admixture than near Easterners, even if the Near East is closer to Africa.

Unknown said...

Alberto

"The oldest samples, especially Ust-Ishim, show a large chunk of it."

As did mal'ta

Unknown said...

"Btw, I think the most Basal Eurasian component is probably the South Asian one. And not Onge, but ANI. The Fst distances show that the South Asian (light green) is the closest to Sub-Saharan. The oldest samples, especially Ust-Ishim, show a large chunk of it. And in Haak et al. Fst table, ...."

Well the centrality of South Asia has been known since the early studies in 1990s. . . in terms of early disperal patterns.

It's role in more recent prehistory requires elucidation

Unknown said...

Alberto

"But I don't see what are you all discussing about. I think everyone agrees on the basic points:"

Well; my question was that there might have been yet another type of forager group- an Aegean type. Differen to WHG and lowland Levantines

Alberto said...

Yes, the facts I stated are rather obvious and known, that's why I say I don't know what's this debate about.

Whether EEF took some HG ancestry in the Balkans? Not a big deal. Consensus is that they probably did get some, but not a lot.

If Balkans HG were pure WHG or had some Near Eastern Admixture? Again, not a big deal. It's possible they had some NE admixture due to proximity with Anatolia.

I don't see how those small details are relevant. Maybe I missed something.

Unknown said...

Yes you're right
It's small detail which doesn't change the overall impact of the Neolithicization process
But the fact it's small detail doesn't detract from it's significance

Alberto said...

@Mike

"my question was that there might have been yet another type of forager group- an Aegean type. Differen to WHG and lowland Levantines"

This would be important if those Aegean HG were different in that they had ANE, for example. But that's unlikely, since we don't see ANE anywhere in Europe before the Late Neolithic. Their difference could be if they were pure WHG or mixed with NE, but either way it won't make a difference.

Or what sort of difference would you expect that would be relevant in some way?

Unknown said...

No I'm not discussing anything about ANE

I'm positing that there existed a wholly different group of foragers - "South balkan"- Anatolian - west mediterranean one. I think there is good evidence from our data that they existed

Starcevo, hungarian LBK have good amounts of "Atlantic" but less of the "baltic", and virtually no North sea and EE components that WHGs otherwise had- even KO1 and la brana.

This huge chunk of Atlantic (and west med) in the Carpathian EEFs can't have all come from Levantine farmers , nor from WHG type groups either (obvously )

Alberto said...

Yes, I see your point. That has always bugged me too. Why WHG had a lot of Baltic (and even East Euro: More La-Brana, less KO1), but EEF were mostly Atlantic. I did argue long ago that they didn't look like a mix of NE and WHG.

It's very evident in K13, compare La-Brana and Loschbour with Otzi and Stuttgart. From 50% Baltic to 0% Baltic.

https://docs.google.com/spreadsheets/d/1Oz6P5-SVEJciPX1TciGe-zoqA5JtOGIMG7nh-rCOj0c/edit?usp=sharing

So yes, maybe those Aegean HG had this specific difference with WHG. But then where did WHG go? Otzi is already Chalcolithic and from Tyrol. Why he didn't pick up any Baltic either?

Unknown said...

I think WHG is a specific amalgam of post glacial processes which primarily concern North, Central and Western europe (+/-) east.

Italy and the balkans (and perhaps Southern Spain) had their own trajectories. The thing is these HGs were never prolific (especially the ones in Aegean and Adriatic europe) didn't contribute anything to WHGs. In fact they might not have contributed much apart from a small part their own regions. Certainly, the Adriatic ones might have become more or less extinct even before the arroval of Cardial- Inpressed ware

Davidski said...

The Atlantic component represents genetic drift along the Atlantic Facade, while the West Med component is basically Sardinian, and in fact a very specific type of Sardinian.

I'm not seeing any signals of as yet unsampled forager groups in these components. I'd say they're both mixtures of ancient Near Eastern and WHG allele frequencies, but the Atlantic also has some ANE influence.

Unknown said...

Perhaps Davo
But why would Atlantic compnent be so high in East-central european farmers ?
I think there something very basal about it

Davidski said...

Mike, have a look at the results above of Spain_EN and NE1 from Neolithic Hungary. They're almost identical.

I'd say Spain_EN is what the Atlantic component is mostly based on, and it's actually less basal than NE1 and Stuttgart.

Alberto said...

It could just be an artefact of admixture, yes. It's strange that in k13, where the WHG is split between North_Atlantic and Baltic, WHG are:

Loshbour: 50% Baltic
La-Brana: 49% Baltic
KO1: 57% Baltic

In contrast:
NE1: 0% Baltic
NE6: 0% Baltic
Stuttgart: 0% Baltic
Oetzi: 0% Baltic

I haven't seen K13 for Spain EN or MN, But the samples above show a clear pattern that looks strange. But I'm willing to see it as an artefact rather than reading much into it.

Maju said...

@Krefter said: "it's common sense one way or another after 10,000s of years people in neighboring regions will mix".

Why specifically Europe and West Asia and not either of them with (NW/NE) Africa or Siberia? Or in the case of West Asia with South Asia as well? Those are also "neighboring regions". In fact the land/strait borders with them were not smaller than between Europe and West Asia (just the Gallipoli-Istambul ancient isthmus and the Caucasus).

My impression from aDNA and also archaeology is that the bulk of admixtures happened in very specific sizable migration processes such as the various Neolithic flows, the Solutrean→Oranian genesis, etc. Therefore they can be tracked in the archaeological record (in many cases at least). Otherwise the contact zones were, especially in the Paleolithic, too small and low density to allow for major gene flow (rather an introgression-style drip at most).

"No strong evidence in remains people in west Asia and Europe interacted before the Neolithic, doesn't mean they couldn't have IMO".

But exactly the same for the people of West Asia and Siberia, the people of West Asia and India, the people of West Asia and Egypt, the people of West Asia and The Horn of Africa. Same for Europe mutatis mutandi.

Of all those possibilities I only detect (1) the Siberia-Eastern Europe intereaction (ANE in EHG/SHG), (2) the SW Europe-NW Africa interaction (mtDNA H and U6, Y-DNA E-M81, etc.) and (3) the West Asia-NW Africa interaction ("Basal Eurasian", E1b in the Eastern Mediterranean, J1/R1b in Africa). I do not see any clear West Asia-Europe interaction between the Early UP and the Neolithic. A reason may be that the population in Europe was then concentrated far away from West Asia (SW Europe), while the evidence in Asia also suggest specific areas rather looking to Africa (Palestine) or with diverse potentials (Zagros region).

Karl_K said...

@Maju

"When the Dinka are used instead the results are contradictory but many do suggest a Dinka-like genetic influence in EEFs. This does not need to be Dinka but something Nubian-like or Paleo-Egyptian it is probably the case. The influence of the Nile cultures in Mesolithic Palestine (and also NW Africa) is well documented"

Once again, you are misunderstanding the data.

The simplest way to understand this is probably by looking at Neanderthal ancestry.

All "Out Of Africa" populations have a similar (but not exactly the same) amount of Neanderthal ancestry. African populations, however, have a wide range of Neanderthal ancestry, but always less than non-Africans. The amount of Neanderthal ancestry in the African populations is always directly proportional to their similarity to the "Out-Of-Africa" populations.

This means that the Dinka and other North Africans are not intermediates between "Sub-Saharan-Africa" and "Out-Of-Africa" populations. In fact, they are the result of ancient admixture between populations of "Sub-Saharan-Africans" and "Out-Of-Africans".

Clearly the divide is not between Africa vs Elsewhere, but the Sahara is the clearest boundary to most of the differentiation of populations.

Basal Eurasians already appear to have had Neanderthal admixture, and that is very unlikely to have occurred within Africa.

All people with Y-haplogroup E also have Neanderthal admixture, which is unlikely to have occurred within Africa. The Tunisian Berbers are >85% haplogroup E, yet are clearly of pre-Neolithic Eurasian origin. Ethiopian's have earlier branches of haplogroup E, and they too have substantial "Out-Of-Africa" ancestry.

The most parsimonious explanation is that there was a major division of populations for tens of thousands of years, and then at a later time, nearly the entire population of North Africa was eliminated/diluted by migrations from Eurasia and Sub-Saharan-Africa.

The only semi-intermediate population we know about was this middle eastern Basal Eurasian population that survived because of their agriculture.

Maju said...

@Davidski: "ANE is missing in WHG, EEF, and certain Bedouin and Saudi groups. These groups don't cluster together. So clustering together can't be used as an argument for lack of ANE."

The form two clusters, with EEF being mixed between both (assumin UHG=WHG, what for this purpose is valid). Those two clusters represent the pre-Neolithic populations of SW or West Europe (Lochsbour) and of Palestine/Levant. They were both the most distant in geography from Central Asia/Siberia, so it makes sense they have "no ANE" - they will still cluster with ANE relative to Africans or East Asians, but not relative to other West Eurasians or at least some South Asians: there is not a yes/no logical port here but relative affinity, which depends on what you compare with (relative to chimpanzee even Khoisan are 100% ANE).

"Palestinians don't lack ANE"...

Because the bottom line is arbitrarily placed at other less admixed populations like Palestinian Bedouins (who are also Palestinians, not some people from the Empty Quarter of Arabia, as some maps wrongly depict them located at). Core Palestinians, being historically more cosmopolitan relative to their desert nomad cousins have absorbed some ancestry from the North carrying ANE. But they are still low enough to serve as reference (as I said before, you may decide to use Yemeni Jews or Palestinian Bedouins instead but you risk distortion on the side of excessive endogamy).

"they have some SSA"...

This can be the same as "Basal Eurasian". In fact Lazaridis' tests consider that possibility when comparing with Bedouins (their range of allowance may be a bit arbitrary though). It's unclear what part of African among Palestinians is BE or something more recent. If that is a problem for testing, I'm repeating: use Yemeni Jews, who lack that African-like thing (but risks missing the BA component, so IMO Palestinians are better precisely because they do have some African, which is roughly the same as BE). Just try with several populations, some may err on one side, some on the other but overall you'll get a good multi-perspective analysis and will be able to judge better.

Of course ideally we want some PPNA and PPNB samples from Palestine and Cyprus (among others). But we still do not have them.

Balaji said...

There has been much discussion of BEA lately. I believe it is possible to use Ust_Ishim to get a sense of how much BEA + African admixture there is in any population. This is because Ust_Ishim is equally related to all Eurasian components except BEA. The following D statistic should do this.: D(Chimp, Ust_Ishim;LBK_EN, Pop) will tell how much BEA, Pop has with respect to LBK_EN. This statistic should be positive for most Europeans and negative for most Near Easterners. D(Yoruba, Ust_Ishim;LBK_EN, Pop) should give an indications of how much African admixture Pop has.

Another statistic of interest would be D(Papuan, Ust_Ishim; LBK_EN, Pop). Here I am expecting that Papuan and Ust_Ishim will be equally related to BEA in LBK_EN and Pop so that this statistic will be negative in proportion to any ENA in Pop.

Davidski, I have created files for the calculation of the above D stats. Please try them when you get the chance.

https://drive.google.com/file/d/0B-aM8HL5oc9WZzRLeDZaek9USms/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WXy15WlJzdUtBUmc/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WakFZOVFkWDJVaTQ/view?usp=sharing

Davidski said...

No Maju, Palestinians carry ANE because their ancestors experienced gene flow from populations carrying ANE that moved into the Near East after the ancestors of early European farmers migrated from the Near East to Europe.

It's that easy.

Maju said...

@Alberto said: "I think everyone agrees on the basic points:"

Not really, let's see:

"- Ancient Near Easterners (from the Levant, or lowlands) were like EEF and lacked ANE (and maybe bits of European HG acquired in the Balkans)".

Not just "bits of UHG" but a sizable chunk (maybe 20-30%). That's pare of the European Neolithic founder effect in Thessaly. EEF is not the same as West Asian farmers therefore: those were similar to modern Palestinians, while EEF were similar to modern Sardinians.

Also Davidski thinks that both early West Asian farmers and UHG could be from Anatolia, a possibility that I don't take seriously.


"- Ancient Middle Easterners from Iran (or highlands) were different, either becaouse of a more basal Eurasian or because they had ANE, or both."

They were different because they were a different subpopulation, not because of ANE/BE. ANE/BE are interesting admixture elements but not the measure of all things genetic.

Let's put it this way: X's parents are Spaniard but has a Nigerian great-grandparent and Y's parents are Swedish but also has a Chinese great-grandparent. What differentiates them: the Nigeria-China axis or the Spain-Sweden one? Both do but the latter is more important in fact.

Same for the various ancient populations: not just minor admixture makes them apart but also intrinsic differences, which overall weight surely even more. Here people is emphasizing ANE/BE too much and often totally not seeing the forest of West Eurasian intrinsic diversity because of these two notorious but singular trees.

"- Modern Near Easterners (lowlands) are a mix of both."

Mostly not in the case of Palestinians, peninsular Arabs. Syrians and Iraqis yes but clinally so. My impression is that in spite of the time passed since then, much of the original duality persists. Hence Palestinians are somewhat admixed but not hugely admixed, they still look quite like the pre-EEF West Asian population. For testing it may be more useful to use some other populations with a less admixed profile, but as Davidski suggests there's a risk that there is extra admixture in the way of Africa in them (or also excess of differentially unique drift), so I'd try using Palestinians or a "zombie" based on Palestinians.

"- WHG had a stronger affinity with ancient Near Easterners (lowlands) than EHG did (and this is seen since Kostenki)."

If true, that's only because of extra ANE in EHG. Otherwise WHG-EHG make a single meta-population relative to West Asians. There's nothing specifically West Asian in WHG nor specifically WHG in ancient West Asians. And this can be tested, so prove me wrong.

"- EHG had a higher affinity with Iranians (or highlands) and South Asians than WHG did (and this is seen since MA-1)".

Even after you remove ANE? There are differences intrinsic to isolation/endogamy (drift) that have nothing to do with Paleo-Siberian or African admixture. Even without ANE, EHG is not quite the same as WHG and the paleo-population of Kurdistan was surely not the same as that of Palestine.

Maju said...

@Karl: do you have any data that supports that Dinka are an admixed population (other than by an irrelevant trickle)? I do not, probably they are purer Africans than the Maasai.

Anyhow, it doesn't matter if they carry some Eurasian admixture, as long as it is pre-Neolithic. Or do you imagine the "First Arabian" residue in Neolithic Levant unadmixed? No way! This is difficult to assess but if the carry just residual Eurasian admixture (and the Dinka are in that negligible range) they are good enough, particularly as TreeMix can apparently discern between donors and receptors.

However your Neanderthal observation may eventually be used to assess if the BE is from either side of the Red Sea: West of it it should have similar Neanderthal component as other Eurasians, East of it probably not. So if the BE component does not include Neanderthal admixture or includes it at low levels, then it's almost certainly African.

Maju said...

@Karl: "All people with Y-haplogroup E also have Neanderthal admixture"...

No way! What an absurd claim! Most Africans are E and their Neanderthal admixture is negligible. In fact Neanderthal admixture was originally tested using San and Yoruba and these are also E. Only one individual of the Yoruba sample appeared to have some very minor Neanderthal admixture (and hence also very minor Eurasian admixture).

Ethiopians and Tunisians are precisely not the kind of populations I'm talking about, but they cannot be used to claim that Dinka or Yoruba are Eurasian-admixed. That's totally manipulating things, I don't want to call you "cheater" but it's borderline, really: I'm truly spooked at the burning-nail quality of your argumentation.

Maju said...

@David: "Palestinians carry ANE because their ancestors experienced gene flow from populations carrying ANE that moved into the Near East after the ancestors of early European farmers migrated from the Near East to Europe".

Totally in agreement. They still carry it at so low levels that we can ignore them and use Palestinians as proxy for proto-EEF (prior to UHG admixture).

"No Maju"...

Actually we are in agreement in this bit at least, so "yes, Maju" and "yes, David".

Davidski said...

Balaji,

https://drive.google.com/file/d/0B9o3EYTdM8lQcWRGVEJXc1IwNlE/view?usp=sharing

Karl_K said...

@Maju

I am sorry. You are indeed correct on all counts. All Y-haplogroup E people do not have Neanderthal ancestry, and Dinka and Yoruba have very negligible Neanderthal ancestry.

So, I guess that means that Basal Eurasian is actually African-like after all...

Except that Basal Eurasians had the same amount of Neanderthal ancestry as the rest of the "Out-Of-African" populations. Otherwise Europeans with substantial Neolithic ancestry would have less Neanderthal, which they do not.

Unknown said...

@ Maju - "When the Dinka are used instead the results are contradictory but many do suggest a Dinka-like genetic influence in EEFs. This does not need to be Dinka but something Nubian-like or Paleo-Egyptian it is probably the case. The influence of the Nile cultures in Mesolithic Palestine (and also NW Africa) is well documented"

HLA data suggests a very strong relationship between Chad/Sudan and Corsica/Sardinia. This includes Podokwo and Uldeme [Chadic langage speakers in N. Cameroon]; and Nuba and Shaigiya of Sudan.

Unknown said...

Maju,

I've already posted the Dstats where West Asians prefer Loschbour, the same as early EEF. I guarantee you that early Anatolian farmers are going to be rather indistinguishable from EEF.

Unknown said...

Maju

When do you propose the Nile influence on levant occurred ?

From what I've read, Egypt was barely populated at the beginning of Holocene

Simon_W said...

Well, the Phoenician colonies on Sicily (main island) were Solus, Panormus, Eryx (though originally an Elymian foundation), Drepanum and Lilybaeum. And it may well be true that many people don't know the three indigenous peoples of Sicily, but I won't forget them.

Simon_W said...

The question's just: Where did the strong West Asian (highland type) admixture come from? The Sicani were presumably Sardinian-like, and the Siculi were allegedly were from central Italy, and linguistically related to other Italic tribes. And the Elymi were confined to the westernmost part of Sicily, albeit with more controversial roots. So presumably the Greeks were the main source of the West Asian (highland type) admixture.

Maju said...

"Except that Basal Eurasians had the same amount of Neanderthal ancestry as the rest of the "Out-Of-African" populations. Otherwise Europeans with substantial Neolithic ancestry would have less Neanderthal, which they do not".

How much would that change the account? Per Lazaridis, the BEA component in proto-EEF (before admixture with UHG) would be like 30%. After UHG admixture it'd be like 20%, which would mean a mere 20% decrease (at most) in the Neanderthal admixture rate, that's the difference between 2.4% (Sardinians, French) and 3%, which is what has been reported (in private testing) for some less EEF-like Europeans, including people from NE Europe and Turkey.

I would like to see clear data on this matter but you do have a point, at least potentially so. My contention is that it is not very clear.

Maju said...

@Mike: "When do you propose the Nile influence on levant occurred ?"

In the Mesolithic. I reproduce here a quote by G. Barker, taken from Wikipedia:

"(...) similarities in the respective archaeological records of the Natufian culture of the Levant and of contemporary foragers in coastal North Africa across the late Pleistocene and early Holocene boundary".

It may be even older than that, as old as the Kebaran (late Upper Paleolithic) but I would have to dig further to document this (the article on Natufian however offers suggestions in that line too).

The influence might have been even more intense in the semi-desert facies known as Harifian, which is often understood as being as the root of Semitic (would still be proto-Semitic), via the so-called Circum-Arabian Pastoralist Complex (CAPC).

"From what I've read, Egypt was barely populated at the beginning of Holocene".

That's probably not true. However archaeological research has been difficult in the country because much of what is now desert was inhabited precisely in the early Holocene. Also the main focus of Egyptian archaeology has been ancient Egypt, with only limited interest on the Neolithic. Hence the data seems limited to a few Delta sites like Fayyum, the occasional oasis like Kharga (apparently related to the Capsian genesis in NW Africa) and inference from neighbors, including Sudan and a scatter of Saharan sites. I'm sadly not sufficiently knowledgeable to explain everything in the detail I'd like to anyhow but I'm very sure that Egypt was not uninhabited, just that we don't know enough.

Karl_K said...

@Maju

"I would like to see clear data on this matter but you do have a point, at least potentially so. My contention is that it is not very clear."

Indeed. It is not very clear. You are quite astute.

Europeans are 'known' to have less Neanderthal ancestry than East Asians, but these amounts are based on setting various African populations to zero. If these populations actually have European (and/or 'basal eurasian') ancestry, then this will give the impression that Europeans have less. On the other hand, if Basal Eurasians did not have Neanderthal ancestry, a similar situation would arise.

It is hard to distinguish without an ancient 'basal eurasian' genome.

Gaspar said...

This Thessalian founder area may best be described by the recent supplementary paper on the Haak paper.
see bottom map
http://rspb.royalsocietypublishing.org/content/royprsb/suppl/2015/03/20/rspb.2015.0339.DC1/rspb20150339supp1.pdf


In regards to hunters and Gathers, if they did not become farmers, they would have not existed. The Australian aboriginals where pure 100% hunters and gathers ( never farmed) , existed for 70000 years and only barely passed 2 million people when the first white man arrived.

Alberto said...

@Maju

Me: "- Modern Near Easterners (lowlands) are a mix of both."

You: "Mostly not in the case of Palestinians, peninsular Arabs. Syrians and Iraqis yes but clinally so. My impression is that in spite of the time passed since then, much of the original duality persists. Hence Palestinians are somewhat admixed but not hugely admixed, they still look quite like the pre-EEF West Asian population."

Since we're taking in colours now, I guess it's me who now has to ask you about being colour blind.

EEF are mostly orange, with a part of blue (which accounts for that UHG that you mentioned). So Ancient Palestinians would be basically 100% orange.

Did you look at modern Palestinians? About 50% Orange, 40% Teal, 10% Somali-like African. What makes you think that they are the best proxy for Ancient Palestinians, when we already have EEF, some of which are like 90% Orange? Do you know something that we don't? Please share your source.

Krefter said...

@Gasper
"The Australian aboriginals where pure 100% hunters and gathers ( never farmed) , existed for 70000 year"

70,000 years is an old-school assumption like people assumed R1b had been in west Europe for 30,000 years. Who knows what happened in Australia, over the last so many 1,000s of years.

Unknown said...

Returning to the subject of Sardinians, EEF, Basal Eurasian, and Dinka. I noticed that Francalacci et al, (2013) found A-M13 in a large sample of Sardinians. A-M13 reaches peak world frequency (61.5%) in Dinka Sudanese.
They also found E-M33 - which reaches peak world frequency in Cameroon Fulani (53%); plus E-M78 [subclade(s) unknown]; E-M81; and E-M165. And finally, R-V88, which reaches peak world frequency in Cameroon Uldeme (95.5%). Together, all of the above account for almost 15% of Y DNA haplogroups among the 1200 Sardinians sampled.
Also, with IJ-M429 and I-M438 being found in Biaka pygmies, IJ-M429 in Yoruba, and I-S154 and I-P37.2 in different populations in Tanzania [including Sandawe], this suggests to me a more ancient presence in Africa and I have to wonder if the significant levels of 'I' in Sardinians might have resulted from a founder effect due to a very early migration from Africa. And lastly, R-M124 [R2a] was found in the Sardinians. This haplogroup also turns up in Sudan and in Biaka pygmies.

Chad said...

Balaji,

I don't like using Ust_Ishim, as he occupies a strange place in the tree. He is not Basal Eurasian, like farmers, but something else. Some like to call this a Basal Crown Eurasian. I think that something between Basal Eurasian and right before West and East Eurasians split, is more likely. Look at how he is equidistant from LBK, Bell Beaker,Corded Ware, etc. That should not happen if he was Basal or fully-derived Eurasian. I am still looking for a better proxy for basal, but unfortunately, too many have recent African ancestry, which skews them in any Dstat.

Unknown said...

@ Maju

Me "From what I've read, Egypt was barely populated at the beginning of Holocene".

* You: "That's probably not true. However archaeological research has been difficult in the country because much of what is now desert was inhabited precisely in the early Holocene"

NOT according to most recent summaries of radiocarbon dates. The lower Nile valley was deserted until 8500 BC, so well after the commencement of Neolithic in middle East.

("Climate-Controlled Holocene Occupation in the Sahara: Motor of Africa’s Evolution"; Kupfer)

Unknown said...

@ Chirs Dvies

" have to wonder if the significant levels of 'I' in Sardinians might have resulted from a founder effect due to a very early migration from Africa. "

Definitely not.

Chad said...

The orange EEF component dominating modern West Asians is not a joke. There was no mass mixing in Europe, by farmers. They came as nearly full-on early farmers. BedouinB are closer to European farmers than their neighbors.

Ju_hoan_North BedouinB Stuttgart Loschbour -0.0275 -7.121
Ju_hoan_North BedouinB Stuttgart LBK_EN_SW 0.0004 0.124
Ju_hoan_North BedouinB Stuttgart Spain_EN -0.0023 -0.767
Ju_hoan_North BedouinB Stuttgart Armenian -0.0099 -3.593
Ju_hoan_North BedouinB LBK_EN_SW Armenian -0.0104 -4.887
Loschbour BedouinB Mbuti Chimp -0.0089 -2.553
LBK_EN_SW BedouinB Mbuti Chimp -0.0004 -0.136
Stuttgart BedouinB Mbuti Chimp -0.0060 -1.956
Spain_EN BedouinB Mbuti Chimp -0.0020 -0.778
Armenian BedouinB Mbuti Chimp -0.0009 -0.622
Loschbour BedouinB Yoruba Chimp -0.0086 -2.570
LBK_EN_SW BedouinB Yoruba Chimp -0.0012 -0.466
Stuttgart BedouinB Yoruba Chimp -0.0044 -1.519
Spain_EN BedouinB Yoruba Chimp -0.0020 -0.848
Armenian BedouinB Yoruba Chimp -0.0017 -1.264
Chimp BedouinB Stuttgart Egyptian -0.0548 -17.819
Chimp BedouinB LBK_EN_SW Egyptian -0.0576 -21.433
Chimp BedouinB Stuttgart Iranian -0.0240 -7.463
Chimp BedouinB LBK_EN_SW Iranian -0.0268 -9.397
Chimp BedouinB Stuttgart Armenian -0.0083 -2.648
Chimp BedouinB LBK_EN_SW Armenian -0.0111 -3.989
Chimp BedouinB Stuttgart Palestinian -0.0295 -9.688
Chimp BedouinB LBK_EN_SW Palestinian -0.0324 -12.402
Chimp BedouinB Stuttgart Iraqi_Jew -0.0108 -3.278
Chimp BedouinB LBK_EN_SW Iraqi_Jew -0.0138 -4.979
Chimp BedouinB Stuttgart Yemenite_Jew -0.0162 -4.814
Chimp BedouinB LBK_EN_SW Yemenite_Jew -0.0190 -6.753
Chimp BedouinB Stuttgart Lebanese -0.0236 -7.395
Chimp BedouinB LBK_EN_SW Lebanese -0.0264 -9.454

Grey said...

@alberto

"I don't see how those small details are relevant."

If they exist they can be used to track population movements like footprints in the dna snow.

.

@davidski

"..populations carrying ANE that moved into the Near East after the ancestors of early European farmers migrated from the Near East to Europe."

That's how it seems to me - the Levant farmers expanded mostly seawards and a second set of farmers came in from the landward side at a later date.

.

@Mark Thomas

"From what I've read, Egypt was barely populated at the beginning of Holocene".

Thing is what was the Nile like before farmers built drainage ditches? If it was a big swamp full of swarms of HGs then the evidence was 1) likely all wooden all wooden and 2) may all be under water.

.

@Davidski

"The Atlantic component represents genetic drift along the Atlantic Facade"

"I'd say Spain_EN is what the Atlantic component is mostly based on"

If the frequency of LP found in ancient dna stays around 1% then given the modern frequency of LP along the Atlantic coast the Atlantic component may be the result of a massive fluke i.e. the dna of the handful of people who happened to have LP when it was critical.

Maju said...

@Gaspar said: "The Australian aboriginals where pure 100% hunters and gathers ( never farmed) , existed for 70000 years and only barely passed 2 million people when the first white man arrived".

That's approx. 10% of the modern industrialized population of Australia (23 mil.), quite a bit.

For comparison, at the advent of Modern Age (mid 15th century, according to N.G. Pounds), the population of "highly advanced" agricultural and trading European countries was similarly low or lower:
→ Belgium-Luxemburg: 1.2-1.5 mil. (~12% of present)
→ France: 18 mil. (27% of present)
→ Germany-Holland-Bohemia-Austria: 10-12 mil. (<10% of present)
→ Switzerland: 600,000 (7%)
→ Poland: 1.5-3 mil. (4-8%)
→ Italy: 7-9 mil. (13%)
→ Iberia: 5-7 mil. (11%)

So IF Aboriginal Australians were at 10% of the record industrial population of the tinderbox continent (your figure!), they were clearly doing extremely well, at levels comparable to those of Renaissance Europe.

Maju said...

@Alberto: "Did you look at modern Palestinians? About 50% Orange, 40% Teal, 10% Somali-like African".

As orange is EEF and not its West Asian precursor you have still to substract the percentage attributable to UHG (~WHG), which can be 30% and weights against EEF affinity with Palestinians (who have no WHG for all I know). That can make Palestinians ~65-70% {EEF minus UHG}.

In PCA analysis, EEF always appear intermediate between WHG and Palestinians or other Lowland West Asians. For example in fig. S19-2 of Lazaridis the straight line drawn (by me) from Lochsbour to Stuttgart cuts the West Asian scatter right on Saudis and only slightly below Palestinians (Yemeni Jews are also in that cluster).

But you are right at least in the sense that modern Palestinians do seem to have too much Highlander West Asian ("teal") component, so probably using peninsular Arabians (even possibly individuals pre-selected among those for their lower Highlander admixture) is better for reference. See: http://forwhattheywereweare.blogspot.com/2013/03/west-asian-autosomal-genetics-two.html

Gaspar said...

@Maju

I do not know where you got the 15th century pop, numbers from , but better to read your compatriots figures ( Fernand Braudel) there is 13M for Italy, 14M for FRA, 14M for GER, 6M for SPA, 1.5M for POR, 8M for POL ...etc etc........and that was year 1420.

Point is , Hunters and Gathers do not multiply as fast as Farmers in society , be it from 18th century or to the bronze-age.

Gaspar said...

@Maju

Wh are these Palestians, where are they from?

I recently watched a program from 2013, with Israeli archaeologists investigating Philistines.....they stated these people are nearly the exact same in pots syles, jewellery styles and genetics and they match their other investigations that match East-Minoan society......so what did minoans have in % of EEF?

Maju said...

@Gaspar: My figures are from N.J.G Pounds "An Economic History of Medieval Europe" (1974), translated to Spanish (my edition 1981, Ed. Crítica, basic book in Economics faculty 1st. grade - when I was younger, of course).

The figures you present are not substantially different except for Poland. All them say that Renaissance economy in Europe could support around a 10%, maybe a bit more, of the present day record industrial population. So, assuming your figure for Australia is correct (no idea), then Aborigines seem to demonstrate that a pre-agricultural society can thrive almost as well as an agricultural one.

Notice I'm not saying that myself because I actually suspect that your figure for pre-colonial Australian population is probably too large, it's YOU who is implying it. And, if confirmed, that would be quite a revelation, even for me.

...

As for what you say about Israeli documentary on Palestinians and Minoans, they were probably dealing with Philistines (a Sea People, probably Greeks from Crete). Palestine does take the name from Philistines but Philistines inhabited only the area around Gaza and had vanished long before the time the Romans renamed the region Palestine, name that stuck till (at least) the 20th century.

Gaspar said...

@Maju

I went for the high end of the aboriginal pop, which was 2.1M..the low end is 750000.
We can guess what we want, it will never be found out because half the pop. died from "white-Mans diseases" before the year 1850.

Gill said...

I'm beginning to think "Teal" (i.e, Gedrosian) might have been a real sort of population once upon a time.

If you take South Asians' admixture results in any calculator with Gedrosian and South Asian components, then add up everything else and use that as admixture for a missing population, then run that through an Oracle, and you'll get some kind of Central Asian cross between Yamnaya, Karelia EHG, Native Americans, and North Europeans, with some North Caucasus and modern Central Asian populations (Tadjik, Tatar, Udmurd) scattered in the results. Top matches are repeatedly Tatar, Udmurd, Yamnaya.

I posted a whole bunch of such runs in the ancient South Asian DNA thread on Anthrogenica.

This "third" component peaks at around 30-40% in Haryana, North India, but the composition is remarkably similar throughout all of North India, no matter how little of it is left (i.e, 5 to 15% in Bengalis).

As you move west towards Afghanistan or to the southwest within India, a lot of West Asian-leaning admixture is represented (more ENF, less WHG, less Uralic/Siberian/etc) instead. I saw 50% Karelia_EHG, 25% Otzi, 25% Ste7. Sometimes Gok4. Basically the same as India but with 25% Middle Eastern which could indicate a separate layer of admixture altogether (4 populations instead of 3).

Krefter said...

New mtDNA data ranging the Neolithic-Middle ages from central Poland, and calls in rs4988235(lactose SNP).

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0122384

Krefter said...

The mtDNA predictions are way off as usual.

LCT-13910*T is absent with the Neolithic samples. Present in the Early Iron age Hallstatt but not at modern frequencies, and at modern frequencies starting just after 0AD and into the middle ages.

Alberto said...

@Krefter

Yes, it confirms what I said the other day. 5000 YBP LCT was found at 31% in samples from a site in the Basque Country and at 0% in samples from Yamnaya (and 0% in CWC and just in one BB from 4400 YPB). So the theories about LP coming from R1b people from the steppe should be put to sleep.

Maju said...

@Krefter: Very interesting, thanks.

Table 1 strongly suggests that the LP allele frequency was strongly increased in the "Roman" (Germanic) period (in Rogowo, not so much in Linowo) and then collapsed with the Slavic migrations (Middle Ages), reaching low frequencies still retained by modern Poles.

Davidski said...

Maju,

It's not that simple. Gruczno (LP 82%) and SBK-4 (LP 86%) are both medieval sites from what was already Poland back then.

Grey said...

@Krefter

very cool, ty

@Alberto

"So the theories about LP coming from R1b people from the steppe should be put to sleep."

Given the (so far) very low initial frequencies it seems wherever it came from it was brought by only a handful of individuals and exploded in place.

Grey said...

@Alberto

"5000 YBP LCT was found at 31% in samples from a site in the Basque Country"

Interesting to place that thought alongside Maciamo's map of the atlantic component from Eurogenes K15.

http://www.eupedia.com/forum/threads/30941-New-map-of-the-Atlantic-admixture-(Eurogenes-K15)?p=451268&viewfull=1#post451268

Karl_K said...

@Grey

"Given the (so far) very low initial frequencies it seems wherever it came from it was brought by only a handful of individuals and exploded in place. "

Not necessarily. It could have been a "mass migration" of otherwise highly similar people. Actually explaining the spread of LCT alleles in Europe will be very interesting.

Because it was actually so recent, they will be able to get excellent resolution on the dynamics (eventually).

Simon_W said...

The MDLP K23b spreadsheet is interesting to study, because of the unsurpassed number of population samples. I've just compared the different Greek samples, in particular I checked the ratio of Caucasus to Near Eastern ancestry. This list shows in sequence: Caucasus / Near Eastern / North African + the ratio Caucasus : Near Eastern. It's sorted from lowest to highest ratio.

Cretan 43.92 / 13.13 / 4.47 = 3.345
Phokaia 41.64 / 11.90 / 2.70 = 3.499
Smyrna 43.07 / 12.17 / 2.74 = 3.539
Isles 44.66 / 12.13 / 3.67 = 3.682
Macedonia 42.33 / 9.76 / 3.17 = 4.337
Peloponnese 40.61 / 8.16 / 2.12 = 4.977
Athens 42.64 / 8.49 / 3.42 = 5.022
Thessal(y?) 38.52 / 6.98 / 2.19 = 5.519
Thessal(oniki?) 38.90 / 6.82 / 2.98 = 5.704
Azov 43.96 / 7.68 / 2.31 = 5.724

Obviously the ratio Caucasus : Near Eastern is lowest on Crete and on the West Anatolian coast (near Lydia!). It's relatively high on the Peloponnese and in Athens, also towards the north (Thessaly, Thessaloniki). I think this means that the incoming Greeks had a higher ratio than the pre-Greek substrate which was preserved better on Crete and on the Anatolian coast. (Crete also has the strongest North African component.)

The fact that the core Greeks have a higher ratio than those from western Anatolia make me doubt that the Greeks were originally from Anatolia. Maybe a northern origin with Yamnaya is still the better variant.

To me these questions are also interesting because my ancestors from the Romagna seem to have had also a relatively low ratio. Which might be attributed to Etruscan ancestry, and which in turn would go well with a Lydian origin of the Etruscans. Actually Dionysius of Halicarnassus' scenario of a Peloponnesian origin of the Pelasgians could be reconciled with a Lydian origin of the Etruscans: If there was a pre-Greek substrate which got dispersed into various directions, by the incoming Greeks, also to northwestern Anatolia, and that's in fact what Dionysius wrote.

For comparison:

Cypriot 45.67 / 16.81 / 5.78 = 2.717
Armenian 51.94 / 12.97 / 1.87 = 4.005

Gaspar said...

@Simon

You must know that the Myceneans of Greece had also West-Anatolian mix, while their replacements in Greece by the Dorians had no Anatolian.

Maju said...

@Davidski: I'm focusing on what could be described as the fixation index: both C/T and T/T produce LP phenotypes but they do not mean the same in terms genetic. And T/T is very low in Medieval and Poland (as well as "Hallstatt" - didn't know there was Hallstatt over there, so I guess it's a bit like the "Roman" label, indicating influence rather than presence).

Davidski said...

Maju,

T/T has a higher frequency at Gruczno, one of the four medieval Polish sites, than at Rogowo.

Also, keep in mind that the Cedynia site, which has the lowest LP frequency in the dataset, was at the very edge of Poland at the time, and in fact mostly outside of it, and shows a rather strange mtDNA composition in this study.

Maju said...

Yes, you're right David, I missed that Gruczno site or I mixed it with the Roman period ones, with the result I confused it all. My bad. It's a Northern thing (in the context of that region) rather than a migrational one: the Roman period sites also show some significant difference along the N-S axis.

On another note it is interesting that we see no obvious signal of spread of the allele in such a small area along time, suggesting that people could perfectly live without it, what seems to contradict selective pressure theories, right?

Krefter said...

I've gone through around 2,500 HV1 profiles from Southwest Asia(Jordan, Lebanon, Palestine, Saudi Arabia, Syria, Yeman) the same way I did with ancient mtDNA.

I've put just about all the HV1 haplotypes from Lebanon(~1,000 HV1 profiles) in a spreadsheet. I'll compare all the SouthWest Asians to each other like I did with ancient samples. Also, I'll compare them to Neolithic Central Europeans and Ancient Steppe people, focusing on typically non-WHG/EHG mtDNA considered to be Near Eastern/ENF.

https://docs.google.com/spreadsheets/d/110Lksf33feYHiLe-Gch7m1m2s2X91x0KR9ltk9Ydpz8/edit?usp=sharing

In every haplogroup which can be broken down into subclades easily with HV1, SouthWest Asians are very differnt from Neolithic Central Europeans.

Mesolithic Balkan ancestry or coming from a differnt set of west Asians could be why.

Davidski said...

Maju,

It looks to me as if the frequency of the LP allele varied wildly in Poland since the Iron Age, even between communities that very likely spoke early west Slavic languages and weren't recent migrants from the west or east. Modern Poles, it seems, a mixture of these groups.

But I have no idea what it says about the LP allele. I'm now more confused on the issue than ever before.

Krefter,

Stick the ancient Polish samples into your analysis if you can.

Grey said...

@Maju

"suggesting that people could perfectly live without it, what seems to contradict selective pressure theories, right?"

If a trait went from close to 0% to up to 98% in some places then it was clearly selected for but selection only implies the trait was critical at the time it was selected for.

If the neolithic crop package initially had low yields in the northern and Atlantic climate zones but it was found cattle raising for milk could provide the necessary calories in those same regions but only for people who were LP then you have a clear selective pressure.

It would be no different to Bedouin in Arabia having LP. It's not that LP turns them into ubermensch (which I assume is the reason people have such a problem with LP as a trait) but that it allowed people to survive in a particular climate zone - like Bedouin in the desert.

So as the neolithic crop package got adapted to the northern and Atlantic climates and/or technology advances e.g. the heavy plow in Europe switching northern Europe from cattle to wheat, then the selection pressure declines.

That seems the simplest explanation.

.

Separately, even if milk was particularly beneficial in itself and not just at a particular time and place then any selection pressure from that might be proportional to the amount in the diet e.g. 40% of calorie intake as milk != 5%, so if the proportion of milk in the diet declines the selection pressure might decline with it.

And why would people switch from a 40% diet to a 5% diet if it was less beneficial?

Numbers.

If you could get x number of big dudes from cattle and milk on a piece of land but 2.5 times that number of slightly smaller dudes on the same piece of land from wheat then wheat wins.

However now that intensive dairying is possible and large quantities of milk can be produced on smaller amounts of land it would be interesting to find out if there are distinct benefits from milk consumption as i imagine it would be possible to manufacture milk for non LP people if enough brains are applied to the problem.


Grey said...

@Krefter

"In every haplogroup which can be broken down into subclades easily with HV1, SouthWest Asians are very differnt from Neolithic Central Europeans.

Mesolithic Balkan ancestry or coming from a differnt set of west Asians could be why."

Very cool.

.

@Davidski

My guess is communities from the same time period and region in Poland* with widely varying LP might be either

1) on a north/south cline with higher LP towards the north i.e. cattle raising increasing as a substitute for lower crop yields further north

or

2) local terrain suitability for crops vs cattle

(*along the Atlantic coast both a north/south cline for latitude and a west/east cline for Atlantic rainfall hence the epicenter in the NW)

Unknown said...

Krefter, are you talking about modern SW Asians?

Grey said...

@Davidski

Alternatively, if rainfall was a significant factor in the relative suitability of land for cattle raising then

http://printfree.cn/sites/printfree.cn/files/Image/20090314151759166.jpg

in Poland's case the LP distribution might be more in the north and south (near Czech/Slovakia) and less in the middle.

Unknown said...

Krefter

Where can I find the mtDNA spreadsheet of ancient mtDNA ?

Krefter said...

@Chad,

I'm talking about modern SW Asians. There is Early Neolithic mtDNA from Syria. It's very strange!!

It's made up of K, R0, H, U*, N*, L3(xM, N), and HV. Most had K, but none had 16093(which a large portion of EEF and modern SW Asians have).

I think the R0 is R0* but I'll have to check if HV1 was tested.

Krefter said...

Mike,

here

https://docs.google.com/document/d/1qTizIyzlsbGl3hVxFO0BMfl58Kd2qLQXhglk4jeDCBg/edit

Unknown said...

It's about location, time, and number of samples. Don't let it bug you too much.

Unknown said...

Thanks Krefter

And don't worry - wait for the BEAN project
There'll be 100s of samples; and we'll discern the pattern easily ..

Balaji said...

The D(Chimp, Ust_Ishim;LBK_EN, Pop) values calculated by Davidski look reasonable. For example, for Palestinians, this value is -0.0296 (Z=13.2) and for Lithuanians it is 0.0053 (z=1.92). However the D values for Southern Europeans such as Maltese, Italian_South, Sicilians, Cypriots, Tuscans, Spanish, Albanians, Greeks, Bulgarians and French are all negative. This is even though middle neolithic Europeans such as Iceman have positive D values showing that they have less BEA than LBK_EN. This is evidence that Southern Europe experienced another episode of migration in the late neolithic from the Near East that increased BEA. This migration also must have brought in some ENA that is evident at K=6 in the Haak ADMIXTURE figure. I have listed the D values in ascending order in the following file.

https://drive.google.com/file/d/0B-aM8HL5oc9WeHljRGZPNzNSLUE/view?usp=sharing

Balaji said...

Davidski, I have a request for some more D values. These are to test the preferences of the various populations for LBK_EN as compared to Germany_MN etc. I have created the files below.

https://drive.google.com/file/d/0B-aM8HL5oc9WRElUa1JGTzQxRXM/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WUmZRalh0eDFlanc/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WUW1NQ0dwTVpCRGs/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WYjRjV0ktcGdISUE/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WYnpDWmRFeEVqM1U/view?usp=sharing

Karl_K said...

@Balaji

"This is evidence that Southern Europe experienced another episode of migration in the late neolithic from the Near East that increased BEA."

Is there actually any evidence that this migration took place in the late neolothic? Why not later?

Ashkinazi Jews also have this component, and presumably came in through Southern Europe (with the most similarity to southern Italians). If the religion migrated along with the genetics, it might have occurred much later.

Maju said...

@Grey: "If a trait went from close to 0% to up to 98% in some places then it was clearly selected for"

It can also mean general demographic replacement by, incidentally, T-allele dominated peoples (from the West or North, as the paper suggests).

The European conditions are different from semi-desert ones. You can't extrapolate whatever happened in Arabia to Western Europe, where rain is common and vegetation luxuriant.

In any case I also have good theories in favor of selection (goat milk for the poor!) but what I don't have yet is clear evidence of such selection actually happening. Demic replacement is a serious alternative possibility, considering what we see in other genetic data.

Davidski said...

Balaji,

https://drive.google.com/file/d/0B9o3EYTdM8lQVGdXU3ExMkpVR1E/view?usp=sharing

Tobus said...

Balaji,

You need to be careful when mixing sample ages in a Dstat. For example, in D(Chimp, UI; LBK_EN, modern), LBK_EN is 7,000 years closer to UI than the modern. Even if the modern is a direct LBK_EN descendent with no admixture the result will be negative to some degree. I wouldn't assume insignificant or borderline significant negative results like French -0.0014 -0.632, Greek -0.0031 -1.402 or Bulgarian -0.0027 -1.125 are necessarily caused by extra post-LBK BEA.

Simon_W said...

As regards the question in the title of this blog post: I'd say the fact that the teal component is bipolar, with peaks in Kalash, Balochi, Brahui and in Georgians and Abkhazians, and with these peaks "by chance" being concurrent with the strongest ANE affinity in West Asia suggests that it's really a mixed component of ANE + Ane-free West Asians.

Actually the ANE affinity is somewhat higher in the northeastern Caucasus than in Georgians and Abkhazians, although the northeastern Caucasus is less teal. But instead the blue HG component is stronger there. Apparently they had more steppe admixture than could be swallowed by the teal component.

That the Yamnaya had some admixture from the Caucasus area is quite obvious in the Eurogenes K15 results. The West Asian component in K15 is really a Caucasus component, one which was absent in all EEF (except for a slight trace in Starcevo_EN), but quite present in Yamnaya. Yet the Yamnaya don't show any EEF/orange in Haak et al., even though that's present in the Caucasus. It must have been swallowed up by the teal component.

Grey said...

@Maju

"It can also mean general demographic replacement"

True (and I expect that's partly what happened) but unless that potential initial group were 90%+ LP to start with it was strongly selected for in them.

.

"The European conditions are different from semi-desert ones. You can't extrapolate whatever happened in Arabia to Western Europe"

Sure but the Arab example is an example of, for want of a better phrase, environmental selection i.e. selection for a specific bio-region.

.

Anyway, the weird thing about that study is the naming; maybe it's a translation thing but "northern route" is not a very good description of what looks more like the Atlantic extension of the med. route: the Atlantic-Med. route would be better imo.

A northern route would be from the steppe then north of the Carpathians and along the Baltic and North Sea coasts to Denmark or Holland whereas the arrows on their map are all from Anatolia again.

Grey said...

After actually looking at their nice map (doh) it's obviously not a northern or north and south distribution so that leaves

2) local suitability for crops vs cattle

or possibly

3) Vistula Germans?

http://en.wikipedia.org/wiki/Che%C5%82mno_Land

Apparently there was a lot of German settlers in medieval times - Chelmno Land in particular being given to the Teutonic Knights in exchange for killing Prussians - and they dominated the towns while the Poles dominated the countryside

so

is there a German vs Polish ancestry element in the four medieval sites?

Grey said...

I'm not sure if it's the same place but one of the high LP sites:

Stary Brześć Kujawski

seems to be this town

http://en.wikipedia.org/wiki/Brze%C5%9B%C4%87_Kujawski

"The earliest written mention of the town dates back to 23 April 1228, when a ceremony took place in Brześć granting the land to the Teutonic Order by Konrad I of Masovia."

Gruczno the other high LP place has a Mennonite mill (Mennonites in Poland were apparently mostly from Friesland which is very high LP IIRC) which doesn't prove anything on its own but is interesting although Mennonites apparently arrived in the 1500s so may be later than the data.

The Mennonites were imported to drain swamps and such as they had a lot of knowledge of that kind of thing from Holland.

Gaspar said...

@Grey

correcing your Prussians, they are neither Germans nor Poles, but baltic tribes , closer to lithuanians and latvians.....a link from your link below
http://en.wikipedia.org/wiki/Old_Prussians

Alberto said...

Looking at those D-stats that Balaji requested I keep seeing this pattern of EHG only having strong affinity with Yamnaya and CW:

EHG LBK_EN Chimp Yamnaya -0.0521 -12.006 344315
WHG LBK_EN Chimp Yamnaya -0.0178 -5.225 352505

EHG LBK_EN Chimp Corded_Ware_LN -0.0231 -5.25 344338
WHG LBK_EN Chimp Corded_Ware_LN -0.0032 -0.879 352607

In modern European populations it seems to dilute a lot:

EHG LBK_EN Chimp Belarusian -0.0065 -1.675 344727
WHG LBK_EN Chimp Belarusian -0.0165 -5.223 353565
EHG LBK_EN Chimp Lithuanian -0.0113 -2.846 344727
WHG LBK_EN Chimp Lithuanian -0.0238 -7.369 353565

Looking at Lithuanian numbers in K6:

Yamnaya: 52%
Pre-Yamnaya: 42%
WHG_extra: 6%

If Yamnaya was some 52% EHG, and Pre-Yamanya (Ötzi) some 45% WHG (being generous) that would still give higher EHG than WHG. But it doesn't, clearly.

Somehow, the EHG is quite below what is expected while the "teal" is as expected. It seems we're missing something.

Maju said...

@Grey: Well, the data from the Chalcolithic Southern Basque Country does suggest two such populations: one 100% T/T and another 100% C/C just beginning to make effective contact. The latter I understand was the same as EEFs, so the former must be "Atlantic" farmers (what else?)

"to start with it was strongly selected for in them".

Or was established by founder effect.

"Sure but the Arab example is an example of, for want of a better phrase, environmental selection i.e. selection for a specific bio-region".

But it's an example that we do not clearly see universally implemented in similar areas, even where there are other LP alleles. The actual LP phenotype of herder specialist populations is wildly varied in fact.

"... "northern route" is not a very good description of what looks more like the Atlantic extension of the med. route: the Atlantic-Med. route would be better imo".

I agree re. Atlantic but the Mediterranean samples are all C/C, just like other EEFs from Central Europe. So it's not something that seems to have originated among EEFs but somehow incorporated only in the Atlantic region, later expanding eastwards. The exact process of genesis is harder to discern, mostly because we have almost no ancient Atlantic data, particularly for this allele (even the Basque samples are from the Mediterranean basin!)

Unknown said...

@ Alberto

"It seems we're missing something."

Maybe not "missing "- but rather not considering other events.

Ie the yamnaya and corded ware samples we have a pure and early. Other admixture events then occurred. So really; the "yamnaya - corded ware genetic horizon" was but one dimension; and there was more survival of WHGs and farmers than assumed by current models.

The "dilution" is already seen in Haaks samples from the later on in the Bronze Age

Davidski said...

Karelia_HG definitely has some East Asian-related Siberian/Arctic admixture, and this might be affecting the D-stats. Samara_HG doesn't have this, as far as I can tell.

Krefter said...

@Davidski
"Karelia_HG definitely has some East Asian-related Siberian/Arctic admixture, and this might be affecting the D-stats. Samara_HG doesn't have this, as far as I can tell."

What about Samara_HG? Would this means Yamna, and most modern Euros have a tinkle of East Asian/Siberian? Considering the 4,500YBP N1c from western Russia, this gives a pretty old range for when "East Asian" N came into East Europe.

Davidski said...

Like I just said, Samara_HG doesn't seem to have it.

Unknown said...

I think they both had the Amerindian pink. Karelia had more pink, but Samara more teal.

Unknown said...

@ David

"Like I just said, Samara_HG doesn't seem to have it."

Well in ur K15 both have the same amount of Amerindian. Or is that different

@ Krefter

Thanks again for your lovely data. Have you included the new Iberian palaeolithic data in it ?

Davidski said...

Unlike Samara_HG, Karelia_HG shows the Siberian crimson in the K9.

Krefter said...

@Mike,

Yes, I added the Palaeolithic Iberian Hs but not thoroughly. I'll update it eventually. There's also Palaeolithic mtDNA from Morocco, but from a 2001 study. A few were R0a1a and R0a2c which are common in the Middle East today, but not in anywhere else. Most were R-CRS(maybe H or HV) to. If legit they're a big deal.

Unknown said...

Ok cool cheers krefter

Dave; do you have K9 spreadsheet with modern and ancient samples ?

Chad said...

I wonder if it's just about quality here. They look pretty much the same.

result: Chimp Karitiana Samara_HG Karelia_HG 0.0061 0.824 9039 8928 203451
result: Chimp Atayal Samara_HG Karelia_HG 0.0018 0.253 8835 8803 203451
result: Chimp Nganasan Samara_HG Karelia_HG 0.0073 1.023 8968 8837 203451

Maju said...

@Krefter, Davidski: maybe Samara HGs didn't have it but the presence of mtDNA C in not just Epipaleolithic Karelia but also the Ukrainian Neolithic suggests that the East Asian "trickle" was already very common in Eastern European Neolithic (not just in the Far North) and that allows for easy hitchhiking into the IE/Kurgan expansion, as well as in previous waves from Eastern Europe like Pitted Ware (apparently derived from Dniepr-Don).

@Mike: "Amerindian" may be different because it may (and often does) indicate Ma1 (ANE) rather than East Asian affinity proper.

Grey said...

@Gaspar

"correcing your Prussians, they are neither Germans nor Poles, but baltic tribes , closer to lithuanians and latvians"

Yes my mistake - apologies to Old Prussians.

I think the main point still stands though. What was the LP frequency in the two towns along the Vistula in 1939 as the expulsion of the German descended population after WWII might be a factor?

Davidski said...

Grey,

All of the medieval sites in the study predate German immigration into Poland.

Grey said...

@Maju

"But it's [Arab LP] an example that we do not clearly see universally implemented in similar areas"

Yes but I'm not arguing that particular environments will always produce particular adaptations.

I'm using Arab LP as an example of selection for a particular region as opposed to selection for general uberness.

Grey said...

@Davidski

"All of the medieval sites in the study predate German immigration into Poland."

Ah ok.

Grey said...

@Maju

Just to be clear - when I say LP was strongly selected for along the Atlantic coast I'm not saying I think it necessarily originated there - it could have come from Mars for all I care - I'm saying wherever it originated once it arrived it was strongly selected for by the environment at the time (whether it was a handful of individuals within each existing group or an external group who all had it replacing the pre-existing population over time it comes to the same thing - pressure form the regional environment).

Grey said...

@Davidski

"All of the medieval sites in the study predate German immigration into Poland."

I thought I must have read it wrong but...


1) "Populations from two medieval sites in Central Poland, Stary Brześć Kujawski-4 (SBK-4) and Gruczno, represented high level of lactase persistence (LP) as followed by the LCT-13910*T allele’s presence (0.86 and 0.82, respectively). It was twice as high as in contemporaneous Cedynia (0.4) and Åšródka (0.43)"


"Data from Cedynia and Åšródka are used as a reference for medieval sites. Both these sites are of quite short history (Cedynia 1.2–1.1 Ka BP, Åšródka 1.0–0.9 BP) and are located outside Kuyavia/the CheÅ‚mno land"


"Burials from the Middle Ages (1.0–0.6 Ka BP) and the Roman period (1.8–1.7 Ka BP) were dated according to the graves’ equipment"

so

1) contemporaneous

2)
Cedynia 1.2–1.1 Ka BP
Åšródka 1.0–0.9 BP

3) Middle Ages (1.0–0.6 Ka BP)

It was the 0.6 Ka BP that caught my eye. Contem...same time period (middle ages) relative to Roman or Neolithic etc but a possible time gap if the two lower LP sites are between 800 AD-1100 AD and the two higher LP are 1000 AD to 1400 AD as that's just when the German settlers started to arrive.

(obviously this would only matter if the German settlers came from high LP regions)

The actual dates of the graves may be given in the back data in which case my bad. It was just the 1.0-0.6 Ka BP that made me wonder. Maybe that was a typo.

Davidski said...

Grey,

SBK-4 is the youngest site in the study. It's a fairly well known Polish site dating to the 12-14th century. I haven't seen it described anywhere as a Teutonic Order site.

Teutonic Order influence in this part of Poland started during the 13th century, but it took a while to develop, maybe because of fighting with the Prussians.

Possible Scandinavian burials have been identified in this region, and they date to earlier than the 13th century, so it's possible that Scandinavian migrants upped the levels of LCT in some Polish town along the Vistula. See here...

http://polishgenes.blogspot.com.au/2013/05/polish-goths-enjoyed-their-millet-while.html

Grey said...

Davidski

"during the 13th century, but it took a while to develop, maybe because of fighting with the Prussians."

Yes, from what I've been reading if the graves were from the 1500s maybe but two heavily LP German sites along the Vistula at that period isn't very likely.

On the other hand your Scandi link looks like it might be a contender for a specifically Vistula / big river link to old viking settlements.

Grey said...

sorry to spam but curiosity has got to me

viking cemetery at Bodzia about ten km from SBK-4 according to google maps

https://www.google.co.uk/maps/place/Bodzia,+Poland/@52.6081683,18.8778514,9z/data=!4m2!3m1!1s0x471ca30c07313747:0x41219a5d3b3ce2e5

http://antiquity.ac.uk/projgall/buko330/


I did find a Polish study on LP which said they hadn't found any regional differences but would a string of high LP clusters around old Scandi sites along the big rivers be easily picked up?

Anyway, who knows, interesting anyway - to me at least :)

(If the Polish LP frequency 1000 years ago was 0.4 ish and the Scandi 0.8 ish and the modern Polish is 0.51 then it might have gone up a bit in the interim.)

Alberto said...

@Davidski, Mike

Karelia_HG might have Siberian admixture, and the Yamnaya samples are early and from the Samara region. But the Corded Ware ones are from Germany already, and they do show strong affinity to these EHG. So their presence all the way down there was real and measurable.

It's later events that seem to have made it go away, but not in the expected way. Mixing with middle neolithic is taken into account.

The only possibility that could possibly explain modern Europeans is a very significant "resurgence" of Motala-like HG, which are much closer to WHG than to EHG but still carry some 20% ANE. This population mixed with the "teal" people would seem to have supplanted the Yamnaya ancestry across Europe.

That scenario is purely theoretical by looking at the numbers, of course. But still, can we explain for example where did BR1 (and to a lesser degree BR2) came from? BR1 looks like CO1+KO1, but with a good amount of ANE from an unknown source. They could be a key to modern Euros, more than Yamnaya itself (again, by the numbers purely, but at least these samples are real).

Davidski said...

Come on, let's be serious, the vast majority of ANE and R1 in Europe comes from Bronze Age Eastern Europe.

Simon_W said...

The Eurogenes K15 West Asian component doesn't seem to be simply a mix of ANE + Near Easterners. BR1 and BR2 for instance have both EHG-related and EEF ancestry, but no trace of this West Asian.

Indeed it does look like a pincer movement, as Mike Thomas called it, for the West Asian component in Greeks is hardly from Yamnaya only.

Simon_W said...

Like I said before, the Corded people are unlike any modern European population, this can be seen in the Fst and in the distances in the Eurogenes K15 4A oracle. On the other hand, the same sources show that the Bell Beaker people in central Germany were already very similar to modern north-central-western Europeans. And the K15 4A oracle shows that at least these Bell Beaker individuals from central Germany were predominantly Corded descended. (Also according to Haak et al. about 50% Yamnaya-like.)

Considering the big difference of the pre-Corded Ware farmers to modern north-central Europeans, the Corded Ware was the missing link, the ingredient that after merging in resulted in modern north-central-northwestern Europeans. But unlike many people used to believe, the Corded people were not identical to modern Baltic and Slavic populations. There were also later Bronze Age changes in the more eastern parts of Europe.

Simon_W said...

The teal component is basically the same thing as the Gedrosia component from Dodecad K12b. For years many people, including myself, were puzzling about the origin of that Gedrosia component. It looked like associated with R1b to me. Now we know that it arrived with the Corded Ware which had plenty of Gedrosia, much more than any modern Northwest European population. And so had Yamnaya.

Davidski said...

The Corded Ware genomes from Germany show some substructure. Two or even three look mixed with local middle Neolithic groups, but the most eastern of the lot shows a high affinity to Poles.

Corded Ware genomes from Poland will probably be very similar, and the vast majority will be R1a-M417 and derived, just like modern Balto-Slavs.

Alberto said...

@David

"Come on, let's be serious, the vast majority of ANE and R1 in Europe comes from Bronze Age Eastern Europe."

That's the most plausible hypothesis with what we know today, at least for NE Europe. But it's precisely those populations that are more likely to be from Yamnaya descent that I'm looking at (Lithuanians, Belarusians) and even there something strange seems to be happening.

What is your own possible explanation for this apparent strong dilution of EHG ancestry in those areas?

The EHG genomes work fine with Yamnaya, CW and modern Siberians. Why wouldn't they work with modern Europeans?

Krefter said...

@Davidski,

Have you compared the Hinxton Celts to LN/BA at all? Do any of the LN/BA show special genealogical relation to modern pops?

Maju said...

"... the vast majority of ANE and R1 in Europe comes from Bronze Age Eastern Europe".

You mean Copper Age, right? Bronze Age is just too late: there were no meaningful migrations from East Europe in that period to elsewhere in the subcontinent.

Anyhow, those claims are in fact far from solved and IMO they don't fit the facts, particularly not R1b and the excessive ANE in NW Europe. For all we know West/North European R1b has absolutely nothing to do with Eastern Europe and actually looks like expanding from Atlantic Europe with two centers. The excessive ANE in NW Europe is a mystery that only further research in that are can solve. In addition to that R1a is ultimately from West Asia, so its presence in South Asia is probably pre-IE at least largely so (again only ancient DNA can give a definite answer anyhow), however, in Europe, R1a (and only R1a) does seem associated to flows from Eastern Europe that were in essence Indoeuropean.

Davidski said...

Alberto,

Yamnaya and CWC are much closer in age to the EHG genomes than modern Europeans, and actually have more direct EHG ancestry.

Modern Europeans acquired their EHG ancestry via Yamnaya and Corded Ware groups, but probably not always of exactly the type that we saw in Haak et al.

Also, I think when comparing ancient genomes to modern populations it's useful to keep in mind that modern populations are often very limited and highly drifted subsets of ancient populations.

Krefter,

No, I haven't yet had the chance to compare Hinxton4 to any of the Haak et al. genomes in any great detail. But obviously it's very similar to Bell_Beaker_LN I0112.

Maju,

Yeah, I meant Copper Age/early Bronze Age.

Unknown said...

@ Alberto ، David, Simon

we have no samples from Eastern Europe, apart from the Carpathians. David thinks that all pre- corded ware groups in EE are simply going to be EEF like. I don't

The other issue is the factor of population replacement in the steppe itself - from warm adapted Kostenki-like people to the R1 people of the mesolithic

Once we clarify these details, we'll have a clearer picture

But it's clear to me that modern Balts are very much like SHG; which means that the idea that SHG became extinct is simply false

Krefter said...

Maju,

" The excessive ANE in NW Europe is a mystery that only further research in that are can solve."

Haak 2015 and Gamba 2014 documented the arrival of ANE in Germany-Hungary. German Bell beaker, Unetice, Urnfield, and the Hinxton Celts all fit in NW Europe. There's no need for another admixture event.

"or all we know West/North European R1b has absolutely nothing to do with Eastern Europe and actually looks like expanding from Atlantic Europe with two centers."

There's room for R1b-L11 to be of the same geographic origin as R1a-Z283. We can see early forms of R1b and R1a coexisted in EHG, and that R1b-L23 and R1a-M417 coexisted in Russia during the Bronze age.

The Samara Yamna samples having mostly Z2105, doesn't mean their brother L151 in Europe can't be from East Europe. Samara Yamna are probably more like cousins than ancestors.

R1b-M269 would have to have originated in West Europe for R1b-L151 to be anything but a recent introduction from east. Obviously M269 originated outside of west Europe.

R1b-L11 looks like a Steppe marker just like R1a-Z283 to me. 3/3 Beakers have R1b, and Steppe ancestry.

I'm open to other possibilities, but ancient Y DNA suggests this one. Reich's opinion is that R1a and R1b in Europe are from Yamna.

Grey said...

from the east or from the west or both?

i wonder if there was a boomerang event, so from the east initially but then the handful with LP had a founder effect event in the west and bounced back eastwards again with whatever dna that handful happened to have by chance - not as a big migration but a gradual drift over time.

Helgenes50 said...

@ Davidski

Do you know the ratio of the Teal component in the Sweden Skoglund NHGs, according to the admixture of Haak, they have a little of it

Davidski said...

I don't know it, but they should be the same as Motala12 in this respect.

Helgenes50 said...

From what I read sometimes , the Pitted Ware culture could be related to a Don Dnieper culture , if this is true, a slight amount of the teal component could be explainable ?

Davidski said...

Yes, I just had a look. They do have a tiny bit of it, which does make sense if they were mixed with recent migrants from just across the Baltic.

I can't run them in my K9 though. They're missing too many markers.

Balaji said...

@Karl_K

You are right that Ashkenazi Jews migrated to Italy much later than the Late Neolithic. However they represent only a small number of people For most Southern Europeans, the Near Easterners who brought in the purple component at K-6 in the Haak ADMIXTURE figure, must have come much earlier in the Late Neolithic with advanced technology such as metallurgy to have had significant demographic impact.

@Tobus

I think you are right that modern people would have drifted more away from Ust Ishim than LBK_EN and this would have an effect on the D statistics. However there are several pieces of evidence that Southern Europeans absorbed a flux of migrants from the Near East some time after the Middle Neolithic.
(1) In the Haak paper, inclusion of Bedouins as one of the source populations improved the modeling of several Southern European populations.
(2) The ADMIXTURE figure in Haak shows a purple K=6 “Papuan” camponent in Southern European and Near Eastern populations. This component was not present in Middle Neolithic Europeans. It must have arrived in Europe with Near Eastern migrants.
(3) f3(European_pop; LBK_EN, Papuan) statistics calculated by Davidski are significantly negative for many Southern Europeans.

https://docs.google.com/spreadsheets/d/1TMLSovx76hseSwX57yeqyQA6nwm1zwrQbrV1MZgaq9k/edit?usp=sharing

Simon_W said...

Eurogenes K15 4A oracle:

Corded Ware I0049:
Least-squares method:
1 East German @ 14.521541
Gaussian method:
1 Yamnaya
2 Yamnaya
3 North German @ 8.629547

Corded Ware I0103:
Least-squares method:
1 Yamnaya
2 Southwest Finnish @ 16.460154
Gaussian method:
1 Yamnaya
2 Yamnaya
3 Yamnaya
4 Ukrainian Belgorod @ 8.867492

Corded Ware I0104:
Least-squares method:
1 North German @ 14.011082
Gaussian method:
1 North German @ 7.708659

Corded Ware I0106:
Least-squares method:
1 Motala
2 Finnish @ 13.903163
Gaussian method:
1 Yamnaya
2 Yamnaya
3 Yamnaya
4 Yamnaya
5 Kargopol Russian @ 10.207269

So I would say Corded Ware I0103 is indeed not very far from modern Slavs (Ukrainians), but only if compared via Gaussian approximation, and distance isn't very low, it's just not excessive in this case.

Simon_W said...

According to Figure S9.23 in Haak et al. Lithuanians are 74.2% LNBA + 25.8% SHG, or with three populations, which results in a somewhat lower resnorm, 67.8% LNBA + 7.9% EHG + 24.4% WHG.

A few comments:
- The Lithuanians are predominantly LNBA, even though there was no EEF substrate in the Baltic, which suggests that later migrations took the LNBA ancestry there. Indeed, the cranial evidence shows that the relatively gracile standard Nordid type appeared in the Baltic in the Bronze Age proper.
- The Baltic peoples have a lot of WHG ancestry. This contradicts the commonly held belief that they should be rather pure Corded people or EHG. According to the above Figure they have more WHG than the Scottish, Irish and Basques.

Balaji said...

Alberto made an interesting observation that modern Northern Europeans seem closer to WHG than to EHG. Davidski, can we calculate the following D statistics? They may throw some light on this issue.

https://drive.google.com/file/d/0B-aM8HL5oc9WNFFibHFmNmJFRlk/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WaEhsVjJjSXUtVFk/view?usp=sharing
https://drive.google.com/file/d/0B-aM8HL5oc9WXy1UcUd0UF9lNVk/view?usp=sharing

Krefter said...

Bell Beaker had high WHG and ANE, so whatever the reason why modern Euros do(Balts are the most extreme) it was occurring over 4,000 years ago. We don't know what is causing this yet. One way or another its excess SHG-WHG ancestry, which didn't exist in CWC, Yamna, and MNE. This is true if the models used are accurate.







Krefter said...

It only took me about 1 month(I'll be done tonight) to get the most of out 2,500 SouthWest Asian HV1 profiles. That's not very long.

There's plenty of mtDNA data online with better coverage and that will take even less time. If people start taking advantage of this, it won't be very long till a detailed knowledge of mtDNA variation in west Eurasia(then the rest of the world) is common knowledge for bloggers. People won't only know something simple like 40% H, and H1 and H3 are kind of popula

Simon_W said...

Regarding the "excessive ANE" in northwestern Europe Maju alluded to, in Haak et al. there are no calculations of ANE ancestry, the authors have abandoned doing this. In Lazaridis et al. ANE was especially high in the Scottish. Now according to Figure S9.23 in Haak et al., the Scottish are best modeled as 89.1% LNBA + 9.2% WHG + 1.6% EHG. This extra-EHG may perhaps explain it. And this is probably from Scandinavian influence, though it's hard to say from what time. But on the whole the Scottish are predominantly LNBA and thus descended from the LNBA central Europeans in Haak et al.

Simon_W said...

Excess WHG-SHG in German Bell Beakers is probably from Danish TRB and from BR1-like people in the Carpathian Basin, at least that's what the K15 4A oracle suggests. (There is no Danish TRB sample, but a mix of Gok2 and Ajv58 works well.)

Davidski said...

Simon,

You're taking the ADMIXTURE analyses and oracle stuff too seriously.

Some populations have developed more distinct genetic profiles in very recent times, mostly as a result of demographic factors, so they won't be all that close to any ancient group.

There's really no perfect solution when looking at genetic affinities across thousands of years and at such a fine scale, but I think that for individual/pairwise comparisons it's hard to beat Identical-by-State (IBS) similarity.

See here...

https://drive.google.com/file/d/0B9o3EYTdM8lQNE9NbHdkazdsTDQ/view?usp=sharing

Simon_W said...

OK, the IBS comparison is of course hard to beat.

Simon_W said...

Maju, as for the origin of R1a: It's a very old haplogroup, so the question is somewhat idle. The vast majoritiy of modern R1a is R1a1 derived. And we know that R1a1 was in a Karelian HG in far northern Europe who showed no evidence of West Asian autosomal admixture. And R1a1 apparently wasn't rare in this part of the world, as evidenced by the additional find from the Upper Dvina, c. 2500 BC. So if the R1a1a1 in the Corded Ware was from West Asia, it must have been a back-migration to the north, but Ockham would prefer the more simple variant without back-migration. The R1a in South Asia isn't old R1a, the vast majority is in R1a1a1b2-Z93 which split from the European centered R1a1a1b1-Z238 not in Mesolithic times, but rather around 3500 BC, coinciding with the beginning of the early Yamnaya horizon and close to the date inferred by Chang et al. for the split between Indo-Iranian and younger European IE families. That's hardly a coincidence and shows that the common ancestor, R1a1a1b lived there, in the steppe culture. Moreover their brother clade R1a1a1a is centered in northwestern Europe, in Germanic people, not in West Asia. West Asia apparently has a high R1a diversity, which suggested to some that R1a originated there. But an alternative explanation is that IE peoples carrying different derived variants of R1a had migrated to West Asia from the west and from the east, thus increasing the diversity.

Davidski said...

Yes, for individual comparisons like this. It's probably better to use Fst and formal stats for populations.

But like I say there's no single perfect solution. You need multiple lines of evidence, and it's likely you'll get somewhat different results with different methods.

The main pattern I'm seeing is the surprisingly high affinity to northern populations among the Yamnaya and Corded Ware individuals, and some sort of minor but real north Spanish/Basque signal for the early Bell Beakers.

Nirjhar007 said...

Simon, Don't be a fool by tagging Language branches with SNP Mutations and dates based on baseless linguistic methods about R1a of South Asia and others let aDNA answer...

Unknown said...

You guys are being rather presumptuous. Pops at 2100-2300 BCE should be rather modern looking.

Davidski said...

Chad,

The problem is with many modern populations, not the ancient ones dating back ~4,000 years or so.

If a population has gone through a rapid expansion or bottleneck, and creates a very modern and specific cluster in ADMIXTURE tests, then it won't match closely any ancient group that pre-dates that expansion/bottleneck.

That's why it's useful to look at the autosomal data from as many angles as possible, and also the uniparental markers at high resolution.

Chad said...

David,

I'm in agreement with you, as far as moderns drifting into their own components at higher K values. My disagreement is with the assumptions by some about Beaker and R1b. I do think the Spanish MN link to Beaker is real, but not because of R1b. The "Spanish" shift was started earlier as well, in Esperstedt. This was shown in the paper by Haak et al. (2015). I've maintained that was about women, for quite some time. In time, we will have R1b groups that are indistinguishable from Corded Ware. Also, those that think the EEF orange is anything but an original Near Eastern farmer.

Chad said...

Sorry, I should be more specific in my comments.

Chad said...

The question that I'm working on now, is to figure out how real that Spanish signal is, and whether it has to do with other factors.

Simon_W said...

@ Nirjhar007

I do not at all agree with people like Genetiker who excessively tag language branches with yDNA SNPs. But that R1a-Z93 is essentially Indo-Iranian can hardly be denied. Of course over time it entered other ethnic groups as well, like Arabs, Jews and Turks. The Turks BTW also probably carried some of it from central Asia to Anatolia. And R1a-Z283 is essentially European, at least in its modern distribution. It's present in West Asia, but rare. It may have migrated there as a minority along the predominantly R1b western IE invaders of this area.

Simon_W said...

@ David

What evidence is there in support of a north Spanish/ Basque signal in early Bell Beakers?

According to Haak et al. Figure S9.8 Bell Beaker can be modeled as 61.9% Spain_MN + 24.2% Karelia_HG + 13.9% Esperstedt_MN. Looks like serious Iberian influence in support of an Iberian origin of R1b and a considerable fraction of Bell Beaker individuals.

But then again according to Figure S9.7 Corded Ware can be modeled as 50.3% Yamnaya + 35.6% Spain_MN + 14.1% Karelia_HG. So did Corded Ware also have Iberian admixture?

Imho this just shows that the Spain_MN influence in Bell Beaker shouldn't be taken too literally. Spain_MN was simply an MN group that was very similar to central European MN groups. There is also the problem that in these calculations they only considered populations that were temporally preceding the populations under analysis, thus excluding important places like the Carpathian Basin after EHG admixture.

What Chad alluded to is the significant f4 stat (LBK_EN, Esperstedt_MN; Spain_EN, Chimp) = -0.00173 (Z=-3.6), showing gene flow from Spain_EN related pops into MN Germany. Rather than R1b, this may have taken I2a2 to Germany. The latter haplogroup is first evidenced in Spain_MN, but is important in modern Germans and also occured in the Unetice samples.

Simon_W said...

Perhaps it arrived via Michelsberg? That's often considered an intermediate culture between western Cardium-derived cultures like Chasséen-Cortaillod and the more eastern TRB tradition.

Krefter said...

@Simon,
"Rather than R1b, this may have taken I2a2 to Germany. The latter haplogroup is first evidenced in Spain_MN, but is important in modern Germans and also occured in the Unetice samples."

HG I clades today all look western European, and the most popular clades(I1, I2a2, maybe not most I2a1-P37 clades) were probably first absorbed by IEs then experienced new expansions and overtook their brother clades. I2a2 in central Europe today maybe be from Iberia, and could be from central Europe.

One of the PWC hunter gatherers was positive for a I2a2 clade popular today, but I can't remeber which one.

It's strange that northeast Euros don't have their own hg I clade. But they do have small amounts of several pre-IE lineages, R1b-L23(IE?), lineages with other origins, and then non-IE N1c.

I'm not informed and confused about the origins of R1b-L11. It looks like a steppe-marker to me, but it could really be from anywhere. We can get a good idea with enough modern data and archaeological knowledge, not just ancient DNA.

Maju said...

@Simon: No, R1a is not "a very old haplogroup" at least per Underhill 2014 (who used full sequences): its expansion from West Asia and soon after secondary centers in Europe and South Asia is from the early Holocene, i.e. Epipaleolithic to Neolithic (or Chalcolithic at the latest). But maybe that's "very old" for you?

Chad said...

Nothing is coming out of West Asia anywhere near the Holocene without Basal Eurasian and or the orange early farmer component. That's a no go.

Chad said...

Into the Balkans, specifically.

Unknown said...

Maju
What would you link your proposed expansion of R1a ex west asia during the Holocene.? Must be agriculture related

And how would you account for its Mesolithic presence all the way in Karelia ?

Davidski said...

I've added two more ancient genomes to the K9 spreadsheet...

Alberstedt_LN I0118

Halberstadt_LBA I0099

https://docs.google.com/spreadsheets/d/1M2drQQD7jvwwaTkpFTuBLQXllkUTVNOIfa9FReH7V5M/edit?usp=sharing

Davidski said...

Maju is stuck in the past.

R1a is an EHG marker native to Eastern Europe. South Asian R1a-Z93 is a young subset of Eastern European R1a-M417.

Krefter said...

I expected Alberstedt_LN to be like CWC, based on his ANE K7 score.

Can you post the ANE K8 scores for a Halberstadt_LBA, Alberstedt_LN, and a Unetice genome?

Also, can you post the Yamna_K6 score for one of the Beaker genomes and one of the Hinxton Celts?

Nirjhar007 said...

David,
''South Asian R1a-Z93 is a young subset of Eastern European R1a-M417.''
I guess Maju don't have Magical powers to foresee all the aDNA results from Asia.
Simon,
Buddy R1a z-93 is a Indo-Iranian marker not The! and i have said above its not very wise to conclude pre-historical language spreads with the Lack of aDNA data from Asia and Near East, the Steppe hypothesis is losing ground with time anyway....

Maju said...

@Chad: "Nothing is coming out of West Asia anywhere near the Holocene without Basal Eurasian and or the orange early farmer component. That's a no go".

I have to sharply disagree: the "teal" component in Yamna is also original from West Asia, at least partly so. That's northerner West Asian (and without BEA).

The orange component instead represents a mix at Thessaly of (1) Southern West Asian (incl. African-like BEA) and (2) Aboriginal Balcanic UHG - so orange is not just "West Asian".

Maju said...

@Mike:

"Maju
What would you link your proposed expansion of R1a ex asia during the Holocene.? Must be agriculture related".

I would think so. The simplest explanation for R1a expansion into both Eastern Europe (and some distinct Nordic erratic) and South Asia would be Neolithic flows. However there are issues relative to dating ranges (uncertainty) and other matters (archaeology), so I do not discard Epipaleolithic flows either. But right now my opinion leans in favor of R1a primary expansion belonging to Neolithic flows from the wider Zagros area and therefore associated to the "teal" component (without much ANE at first probably).

Maju said...

@Mike:

"And how would you account for its Mesolithic presence all the way in Karelia ?"

I'm ignorant of that claim. Has R1a been found in Epipaleolithic Karelia? If so, then revise my previous comment in favor of the alternative Epipaleolithic option.

Nirjhar007 said...

Maju.
''But right now my opinion leans in favor of R1a primary expansion belonging to Neolithic flows from the wider Zagros area and therefore associated to the "teal" component (without much ANE at first probably).''
Quite Practical! and my apologies for earlier tantrums i was angry with someone else and you were the victim:P.

Unknown said...

Maju

Mesolithic Karelia, yes- the Haak paper

(Btw Epipaleolithic is a term reserved for Southern Europe, where the transition was more gradual)

Unknown said...

Maju

"The orange component instead represents a mix at Thessaly of (1) Southern West Asian (incl. African-like BEA) and (2) Aboriginal Balcanic UHG - so orange is not just "West Asian"."

I agree that there is a possibility that much of the Orange in EEFs is Balkan and even Anatolian HGs.

Maju said...

Davidski said: "South Asian R1a-Z93 is a young subset of Eastern European R1a-M417."

Nirjhar replied: "I guess Maju don't have Magical powers to foresee all the aDNA results from Asia".

What I do not have is the feeling of authority to contradict Underhill 2015 ("the past"? The very recent past). M417 is not "European", much less "Eastern European", it's West Asian. I quote:

... five of the six observed R1a1-SRY10831.2*(xM417/Page7) chromosomes were also from Iran, with the sixth occurring in a Kabardin individual from the Caucasus.

I'm guessing Davidski may cling to the burning nail of the Kabardin guy, but we all know that North Caucasians cluster with Highland West Asians and not with mainline Eastern Europeans of any sort.

Of course there may be some future finding I cannot indeed foresee that topples this data but so far I have seen nothing, just stubbornness and wishful thinking.

Nirjhar007 said...

Mike, I think Jeitun can be a candidate for the genetic Profile of people who Entered in Southern Russia around ~4000 bc?, that culture is in the right place and has some flows from Near Eastern Area also recorded by archaeology....

Nirjhar007 said...

Yes Maju i agree on R1a-M417 with you but we need aDNA to seal the deal..

Maju said...

To clarify: M417* was reported by Underhill in Turkey and Norway, none of which is "Eastern Europe". One could argue, I guess, for a Norwegian (but not "Eastern European") origin of M417 in these circumstances but I cannot because:

1. The upstream levels are all clearly West Asian (Iran-centered to be specific)
2. The Norwegian and Turkish branch weight equally in theory, so there's no particular reason to favor the Norwegian origin over the Turkish one.
3. The downstream levels also partly show an undeniable West Asian centrality (Z93)
4. Archaeologically speaking Norway is not believed to be the origin of anything except maybe in the Viking era, but rather a destination of various waves.

If one wants to be absolutely equanimous about this, Turkey is the most parsimonious origin of M417 and the closest one to the geographic centroid of the four known branches of this lineage (which would fall right on the Black Sea). It's proximity to Iran (and Kurdistan, Kabardinia) also supports the notion.

Finally if one wants to imagine R1a migrating from Iran to Europe and then back to Iran before expanding into Central and South Asia... fine. But how? What's the archaeological trail that may be coherent with this? Certainly not Kurgans (nor anything else I can imagine).

Nirjhar007 said...

Maju, How we can ignore BTW that the Karelian HG who had R1a1 had no West Asian type Ancestry But EHG?

Krefter said...

@Maju,
"Of course there may be some future finding I cannot indeed foresee that topples this data but so far I have seen nothing, just stubbornness and wishful thinking."

The future is here in my opinion. The evidence for a East European source of R1a-Z283 and R1a-Z93, or all M417 is stacking up.

There's R1a-M198(xM417), R1a-M417, and R1b1(xM269, M73) from Mesolithic Russia.

EHG is closer to MA-1, who had R*, than anyone around today. MA-1's next closest relatives are Amerindians who are uni-formally Q, the brother of R. The low coverage genome; AG2, who came from the same region as MA-1 and is closely related but several thousand years younger is positive for a P and R1 SNP. He could be R1.

There's a clear pattern.

MA-1 affinity was very rich in Samara Yamna who were all R1b. MA-1 affinity first appears in central Europe with CWC and BBC, who are 6/6 R1b and R1a.

Like I've said before Yamna-type(based on mtDNA) people were in Asia(Andronovo, etc.) and were almost 100% R1a-M417, some xZ93 and some Z93+.

Prove is coming soon they were similar to Yamna autosomoally, not a complex mix of EHG and other stuff.

A month or so ago Davidski posted a link to a map created by people who sampled Afanasevo DNA, showing migrations of Yamna based on ancient genonmes. It marked Yamna migration into Asia(Afanasevo) and Europe(Corded ware).

I think someone is also sampling Andronovo genomes. What we'll get is R1a-Z93 and Yamna-like people autosomally.

Nirjhar007 said...

Krefter,
''Like I've said before Yamna-type(based on mtDNA) people were in Asia(Andronovo, etc.) and were almost 100% R1a-M417, some xZ93 and some Z93+.''
Well that's purely because of Geographical reason bud.

''Prove is coming soon they were similar to Yamna autosomoally, not a complex mix of EHG and other stuff.''
But weren't Yamnaya were already a mix of Asians-West Europeans-EHGs?


''A month or so ago Davidski posted a link to a map created by people who sampled Afanasevo DNA, showing migrations of Yamna based on ancient genonmes''
Its not conclusive.
//I think someone is also sampling Andronovo genomes. What we'll get is R1a-Z93 and Yamna-like people autosomally. //
Yes But I think they will have more Central Asian type Ancestry with some Substantial SC Asian Ancestry and Z-93 is very much expected...

Maju said...

@Mike: "Mesolithic Karelia, yes- the Haak paper"...

Wow, I didn't even notice! But you are absolutely right: "Two hunter-gatherers from Russia included in our study belonged to R1a (Karelia) and R1b (Samara)"...

In table S4.2 it is only described as R1a1, what implies, I understand that it is R1a1* (M459, two phylogenetic levels above of M417). On this phylogenetic level Underhill says: Of the 24 R1a-M420*(xSRY10831.2) chromosomes in our data set, 18 were sampled in Iran and 3 were from eastern Turkey [= Kurdistan].

In any case it does seem supportive of my early suggestion about an Epipaleolithic chronology for at least part of the R1a expansion (scroll down to "time frame?" and "update" here: "http://forwhattheywereweare.blogspot.com/2014/03/y-dna-r1a-spread-from-iran.html), which required to use the oldest range of possible chronologies: the Méndez (Anzick-calibrated) rate would give an age of 6.9-9.0 Ka to R1a, while my recalibration would point to 11 Ka BP.

"Btw Epipaleolithic is a term reserved for Southern Europe, where the transition was more gradual".

Not al all: Epipaleolithic refers precisely to those areas where the Paleolithic-Neolithic transition is sharp, which is all Europe (among other regions). Mesolithic is sometimes used loosely (including Epipaleolithic) but more strictly it means only those areas where there was a gradual transition from hunter-gathering to agriculture, notably the Fertile Crescent. So for example Natufian is Mesolithic (they gathered cereals, lived sedentarily) but Azilian is Epipaleolithic (just a late regional phase of Magdalenian with nothing pointing to Neolithic). This should be obvious from the etymology of the words: Mesolithic reads "middle stone age" (i.e. between Paleo- and Neo-) and Epipaleolithic something like "periphery of the old stone age" (i.e. same as Paleo- but terminally so).

Maju said...

Sorry I quoted Underhill for the previous phylogenetic level, the correct quote is: Similarly, five of the six observed R1a1-SRY10831.2*(xM417/Page7) chromosomes were also from Iran, with the sixth occurring in a Kabardin individual from the Caucasus.

So I'd guess that the Karelia R1a1 is plausibly related somehow to this first overflow of the lineage outside West Asia across the Caucasus, because now we don't just have there a modern Kabardin but also an ancient Karelian. But it seems very hard to follow up in the line of Davidski from here, really. It is clearly not direct support for M417 being "European", because it's not M417 yet.

Nirjhar007 said...

Maju, what about as i'm asking again that how we can ignore the Karelian HG who had R1a1 had no West Asian type Ancestry But EHG?

Maju said...

@Krefter: "There's R1a-M198(xM417), R1a-M417, and R1b1(xM269, M73) from Mesolithic Russia".

Explain me that: what I see in Haak is R1a1, which is not yet M198 (R1a1a). Can you point me to the source of that claim?

Re. M417, I have no idea of what you're talking about. That's also new to me.

Anyhow, how many data points you have from Meso- or Neolithic Iran? I mean: let's be serious!

Maju said...

@Nirjhar: "how we can ignore the Karelian HG who had R1a1 had no West Asian type Ancestry But EHG?"

In a few generations admixture, you lose any significant autosomal signal but you can keep the lineage.

→ 1 generation: 50%
→ 2 generations: 25%
→ 3 generations: 12.5%
→ 4 generations: 6%
→ 5 generations: 3%
→ 6 generations: 1.5%
→ 7 generations: <1%

At 30 years per generation, that's barely above two centuries. That's why autosomal and haploid data don't need to bear any relationship whatsoever.

But I do find funny that you're being more easily persuaded than me. XD

Davidski said...

Maju,

Karelia HG belongs to a sister clade of M417, and there's a M417* Corded Ware sample in the Haak et al. dataset as well.

It should be obvious to anyone by now that R, R1a and R1b are originally paternal markers of Eastern European and Siberian hunter-gatherers.

If there's any R1a in Mesolithic Iran, then it'll be there along with ANE/EHG admixture from the steppe. It's not important what Underhill managed to dig up in the modern Near East.

Krefter said...

Maju, It's not smart IMO to base your whole opinion on the history of R1a on one study. It isn't random almost every ANE-rich ancient Y-DNA result has come out R.

Nirjhar007 said...

Thanks Maju! now lets see what the other suggest to your suggestion....

Unknown said...

@ Maju

I'll leave the discussion on R1a on the corner for now..

But your incorrect about "Epipalaeolithic". Its not a term used for regions where the Palaeolithic - Neolithic was sharp, but rather where the Palaeolithic - Mesolithic transition was gradual.

In northern Europe it was 'sharp' because of the large changes in fauna after the LGM, with mass extinction an all. In the south, it was gradual, in fact, almost unnoticable. Lithic technology changes very gradually from 20kya to 10kya, rather than the sharp transition we see between the Gravettian and Epi-Gravettian in northern Europe.

I'd have thought a seasoned prehistorian like yourself- esp from southern Europe - is aware of these simple terminologies ?

Nirjhar007 said...

R is Not E European its Central Asia-SC Asian because of the R1 and R2 split and R1a ( Forget R1b as its history is much complex apparently)) can also be a inflow from West Asia-Central Asia since and we can't predict Autosomal structure nor we should give it a very high value as its a subject of total dilution IMO now...

Davidski said...

Kefter,

K6

Pop Hinxton4
ID ERS389798
Yamnaya_related 0.503688
WHG_extra 0.025562
ENA 1E-005
Middle_Eastern 1E-005
Pre-Yamnaya 0.47072
Sub-Saharan 1E-005


Pop Bell_Beaker_LN
ID I0112
Yamnaya_related 0.444792
WHG_extra 0.053223
ENA 1E-005
Middle_Eastern 0.024961
Pre-Yamnaya 0.471587
Sub-Saharan 0.005426

K8

Pop Alberstedt_LN
ID I0118
ANE 0.1641
South_Eurasian 0.0118
Near_Eastern 0.3972
East_Eurasian 0.0004
WHG 0.4233
Oceanian 0.0002
Pygmy 0.0026
Sub-Saharan 0.0005


Pop Halberstadt_LBA
ID I0099
ANE 0.1676
South_Eurasian 0.013
Near_Eastern 0.351
East_Eurasian 0
WHG 0.464
Oceanian 0
Pygmy 0.0042
Sub-Saharan 0.0001


Pop Unetice
ID I0047
ANE 0.1482
South_Eurasian 0
Near_Eastern 0.3559
East_Eurasian 0
WHG 0.4958
Oceanian 0
Pygmy 0.0001
Sub-Saharan 0


Pop Unetice
ID I0116
ANE 0.1642
South_Eurasian 0.0071
Near_Eastern 0.3334
East_Eurasian 0
WHG 0.4952
Oceanian 0
Pygmy 0
Sub-Saharan 0

Helgenes50 said...

@ Davidski

When comparing the different results, the teal ( in Europe) is linked to the IE,
but it's possible to have a high level of it and a low level of Yamnaya-related or the opposite. For example, the ratio of the teal component compared to the blue European is higher for the French than for the Norwegians and yet these ones are more IE.

Krefter said...

Interesting how Late Neolithic-Iron age West Euros don't score any Middle_Eastern, but modern French and western Germans(not from K6 spreadsheet) score 5-10%.

Maju said...

@David: I quote from Haak's supp. materials:

the hunter-gatherer from Karelia could only be assigned to haplogroup
R1a1 (M459:6906074A→G, (age65.2:2657176C→T) and the upstream haplogroup R1a (L145:14138745C→A, L62:17891241A→G, L63:18162834T→C, L146:23473201T→A). It was
ancestral for the downstream clade R1a1a (M515:14054623T→A, M198:15030752C→T,
M512:16315153C→T, M514:19375294C→T, L449:22966756C→T). Thus, it can be designated as belonging to haplogroup R1a1*(xR1a1a) and it occupied a basal position to the vast majority of modern Eurasian R1a-related Y-chromosomes4, although more basal (R1a-M420*) Y-chromosomes have been detected in Iran and eastern Turkey4
.

So it's not a "sister clade" to M417: it is either a "grandfather" or an "uncle" clade, depending on whether the Karelian individual was strictly ancestral to M417 or (most probably) wasn't.

The M417 in Corded Ware is not Epipaleolithic, as Mike claimed, nor Eastern European either (it's from Germany). So please do not drive me crazier mixing different bits of data, which may be or not related. We are talking of two different instances that are separated in (1) the phylogeny, (2) in time (several millennia apart) and (3) in the geography.

It's interesting in any case that the CW individual was "underived" also (as far as modern phylogenetic knowledge allows researchers to determine), so one can imagine him related to the ancestors of Eastern European R1a-Z282. However how do you imagine that same lineage arriving to Iran (and even India) before expanding into Central Asia and then Westwards to Samara as Underhill's phylogeny obliges? I have a hard time believing that, really, even if Underhill himself wants to believe in that kind of short chronology by resorting to ill-calibrated scholastic data. Rather I'd think that it's a remnant of the migrant process.

But whatever the Iranian ancient DNA supports when it comes around.

@Krefter: "Maju, It's not smart IMO to base your whole opinion on the history of R1a on one study. It isn't random almost every ANE-rich ancient Y-DNA result has come out R".

Well, at least I'm basing my opinion on something. Yours seems a patchwork of rather capricious ideas. We still don't even know what, if anything, does ANE mean but we know that haploid lineages (patri- and matrilineages) and autosomal DNA are two very different animals. Autosomal DNA reflects the overall of your ancestry while haploid DNA only one very specific line. If anything, in my experience, it is mtDNA the one most closely related to autosomal DNA, just look at Finns but there are many more examples. I still have to see one single case in which autosomal DNA is more closely related to Y-DNA than to mtDNA.

Some of you seem blinded by the cultural patriarchal association to ancestry. present for example in surname transmission, which is unreal when you go outside the patrilineage. Your father is only 50% of your ancestry, your father's father only 25% and your patrilineal ancestor from two centuries ago is 1%!

Maju said...

@Mike: I've been praised by professional archaeologists and prehistorians on my correct usage and even explanation of the term "Epipaleolithic". I know very well what I'm talking about.

It's one of those thinks continental European which the Anglosaxons don't always get right, like the metric system, you know.

Unknown said...

@ Maju

"The M417 in Corded Ware is not Epipaleolithic, as Mike claimed, nor Eastern European either"

No I didn't. I clearly stated Karelia and Mesolithic. (Who would I make as ludicrous claim as arguing that Corded Ware is Epipalaeolithic ? !)

And i didn;t make any statements as to where I believed it originated from.





«Oldest ‹Older   201 – 400 of 640   Newer› Newest»