New blogs

Leherensuge was replaced in October 2010 by two new blogs: For what they were... we are and For what we are... they will be. Check them out.

Saturday, July 3, 2010

Central Eurasian genetic specifity detected


Found at
Dienekes and GNXP:

Hui Li et al., Genetic Landscape of Eurasia and “Admixture” in Uyghurs. AJHG 2009. Now freely available at PubMed.

The interesting stuff is in figure 1:


So we have now a cluster centered at the Khanty (an East Uralic population) but strong in Central Asians that is distinct from both West and East Eurasians.

I think this is the coolest discovery of autosomal genetics in quite a long time.

Dienekes protests about the inconsistence of this with Y-DNA but I fail to see the connection, because Y-DNA or even mtDNA, surely scattered in the early Upper Paleolithic when Eurasians were still very much undifferentiated, while these components are surely shallower and represent regional homogeneization processes that happened surely only after the LGM. Also the impact of Y-DNA flows may be really weak: Buriats and Finns share Y-DNA lineage, as do Polish and many Indians or West Europeans and some Central Africans but it's obvious that these don't correlate too well with autosomal genetic clustering which has its own processes of regional homogeneization, while haploid genetics and specially Y-DNA is subject to high odds of fixation by mere drift and founder effects.

Razib says that the authors seem to be arguing for greater number of (strategically chosen?) populations, what makes total sense and has served, when done, to add many shades to the tricomy of the old HapMap continental hyper-simplification.

However it's very possible that a deeper cluster analysis would have revealed futher clusters, maybe an Indian one (very small sample though) or a distinction between the Khanty and Central Asians or...

Where does this Central Eurasian cluster stem from. I'd say, based on haploid genetics, that it was probably created by the admixture of a West Eurasian and East Eurasian migration converging in that peripheral area and then coalescing into a rather homogeneous cluster on its own right. That would explain why the population appears almost exactly intermediate between the two major continental groups and also why there is no haploid lineage that specifically belongs to that region (all are shared with either West Eurasia or East Asia).

East Asian variations

It is noticeable that three quite distinct populations emerge in East Asia: inland, coastal and south. These are probably essentially comparable with the blue, yellow and one (or all together) of the SE Asian components detected in the HUGO Consortium paper.

31 comments:

Anonymous said...

Your opinion makes sense except that I can't help noticing the aDNA studies in Kazakhstan * and south Siberia ** seem to show a major trend, a change in these populations from bronze age to iron age in these regions.
First the results show a population with almost exclusively west eurasian haplogroups then the apparition, more and more visible, of east Asian haplogroups.

Do you think it is just a coincidence and that more aDNA samples in both studies (around 30 samples in both, IIRC) would have shown evidences of a mix between these populations from the start, in the most ancient samples? Two different studies would show the same false trend by mere coincidence?

What makes me doubt your theory is also that several elements seems to support the arrival of a caucasoid population deep into Asia during chalcolithic-bronze age. I'll go directly to what seems the most obvious case :

In south Siberia, aDNA almost exclusively R1a1a + typical west Eurasian haplogroups with europoid skeletons (and quite different from the skeletons of their neighbours - this dichotomy rather goes against the central Asian paleolithical mix theory, I think) that were light-pigmented individuals (at least 60 % of the samples), that I assume are originally related to the Afanasevo culture (a culture of pastoralists having some resemblances with the Yamna culture) starting about 3,500 BC (it could go well with the fact that Tocharian language was centum and fits with the Kurgan theory as the date could go well with the spreading of the west IE centum languages whose Tocharians have so much similarities with, and the European archeological cultures seen as being early IE). Tested on pigmentation-related loci, they were classified as Europeans too ( http://1.bp.blogspot.com/_Ish7688voT0/Sgl-jY6gjjI/AAAAAAAABdQ/OsUUZB04OhQ/s320/triangular_siberian_pigmentation.jpg ).

* Ancient DNA provides new insights into the history of south Siberian Kurgan people (http://www.springerlink.com/content/4462755368m322k8/)

** Unravelling migrations in the steppe: mitochondrial DNA sequences from ancient central Asians. ( http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1691686 )

Maju said...

I am not totally sure of what do you mean by "a major trend, a change in these populations from bronze age to iron age in these regions". I presume that you mean the quite evident advance of East Asian lineages c. 600 BCE and onwards, along with what is generally understood as the expansion of Turkic peoples from somewhere in or near modern Mongolia.

This fact is also reflected in the Kazakh and Uyghur datasets, with about 1/3 of East Asian components at k=6, mostly the "Mongol" (or rather "Baima") gray component.

What seems to change is that rather than the West Eurasian lineages detected in aDNA being mostly of direct European origin they now seem to be of local Central Eurasian origin, as evidenced by the distinct cluster, which essentially replaces the European/South Asian one detected in older papers (and you can see how in the transition from k=5 to k=6 in this paper).

It's a clear case of "hidden component", caused by not running cluster analysis software deep enough, IMO (or may also be aided by insufficient sampling). Once the "hidden component" is found, it shrinks the other components, sometimes very dramatically.

"Do you think it is just a coincidence and that more aDNA samples in both studies (around 30 samples in both, IIRC) would have shown evidences of a mix between these populations from the start, in the most ancient samples?"

I admit I haven't dedicated enough time to the relatively abundant aDNA research from the steppe belt. But Jean Manco has all the data sorted here.

You can see that the most ancient mtDNA from Tarim Basin, c. 2000 years ago, is mostly C4, with some M*, R* (F?) and only very few clearly Western lineages (H and K). However all Y-DNA is "Western" R1a (except for one C(xC3)).

However just a few centuries later in Andronovo (Central Asia but very close to Europe) the lineages are exclusively Western (U, H, T) but already with one case of Z. This seems to be also the case of IE-related Altai and Kazakh sites in the Bronze Age.

There's not a single case in Uyghuristan at least where Western mtDNA dominates at any moment, west and north of there there is some case for greater doubt because it does fluctuate as you say.

"What makes me doubt your theory is also that several elements seems to support the arrival of a caucasoid population deep into Asia during chalcolithic-bronze age".

Yes. But was this "Caucasoid" population "European"? Maybe it was specifically Central Asian, as this paper seems to suggest.

It's hard to say because its genetics should have been best preserved not in Kazakhstan or Uyghuristan but in southern Central Asia (Uzbekistan and surroundings), where the impact of Turkic genetics is clearly lower. Another interesting population to sample would have been Altaians, who are still pretty much unusual with their Y-DNA R1a+Q pool, maybe a reminder (at least in the Q side) of a time when West-Central Siberia was "a world apart".

Anonymous said...

"I am not totally sure of what do you mean by "a major trend, (...) it shrinks the other components, sometimes very dramatically. "

Well, you did say "that it was probably created by the admixture of a West Eurasian and East Eurasian migration converging in that peripheral area and then coalescing into a rather homogeneous cluster on its own right" (and you made it clear that you thought it occured during paleolithical) so I was pointing to the fact that two major aDNA studies were showing a different situation haplogroup-wise (almost exclusively west Eurasian haplogroups during bronze age, then more and more east Asian ones, not a west-east ancient mix of population).

"You can see that the most ancient mtDNA from Tarim Basin (...)"

I don't think the Tarim Basin is the best place to observe the situation of Central Asia before, during and after chalcolithic/bronze age (beside the fact that it's apparently difficult to access from the west, the south and the north, it's also a quite desertic place that we don't have prooves of settlement prior to bronze age, if I'm not mistaken).
If, as it is supposed, the early Tarim inhabitants came from Afanasevo population (Afanasevo's generally a good candidate because it could fit with the features of Tocharian language, and some potteries of the Tarim have been considered linked to Afanasevo according to a Mallory book I've read), they had plenty of time to mix with east Asian before 2,000 BC (date of the oldest tested Xiaohe peoples from Tarim IIRC).

"However just a few centuries later in Andronovo (Central Asia but very close to Europe) the lineages are exclusively Western (U, H, T) but already with one case of Z.
This seems to be also the case of IE-related Altai and Kazakh sites in the Bronze Age. However just a few centuries later in Andronovo (Central Asia but very close to Europe) the lineages are exclusively Western (U, H, T) but already with one case of Z. This seems to be also the case of IE-related Altai and Kazakh sites in the Bronze Age. "


"There's not a single case in Uyghuristan at least where Western mtDNA dominates at any moment"

IMO the south Siberian samples are related to a population of pastoralists from the west arriving around 3,500 BCE in this region (I think it also fits with what we can reconstruct from their phenotypes) that's why the bronze age (as early as 1,800 BC in Keyser et al 2009) samples are with 90% west Eurasian mtDNA hgs while the iron age samples are only 67 % with west eurasian mtDNA hgs.
The more recent (than the Tarim ones) Andronovo samples are almost completely west Eurasian during bronze age, unlike in the Tarim basin, because they are the source (IMO the early Tarim R1a1a probably came from south Siberia). The population of R1a was largelty mixed in 2,000 BC because their ancestors probably were at the fringe of west Eurasian and east Asian populations before migrating in Xinjiang.

Anonymous said...

But was this "Caucasoid" population "European"? Maybe it was specifically Central Asian, as this paper seems to suggest.

I admit that your view makes sense, there is a logic in it, but at the moment I lean towards a gene flow from the west during chalcolithic. Among the reasons is the fact that I don't think mtDNA haplogroups such as H5a, H6, HV, T1, T3, T4, K, U2, U4, U5a1 and I, were in south Siberia (or even in central Asia) as early as the paleolithic (well, some of them could have been, of course, but not all of these IMO).
Maybe the truth is kind of in the middle though.

"Another interesting population to sample would have been Altaians, who are still pretty much unusual with their Y-DNA R1a+Q pool, maybe a reminder (at least in the Q side) of a time when West-Central Siberia was "a world apart". "

This I also doubt. It's kind of counter-intuitive given the location, but since during bronze age 90% of the mtDNA was east eurasian (67% during bronze age and much less nowadays IIRC) with R1a and that nowadays R1a is by far the most frequent haplogroup in the Alataians, I assume at the beginning the region was first populated by R1a1a + west eurasian mtDNA hgs (maybe some of the y-DNA D which i think is found in the Altai is a track than a more ancient local population though, I wouldn't know) and that the east Asian haplogroups (Q and C) arrived later, otherwise East Asian lineages would have been more visible during bronze age.
The other possibility is that there was a social bias (hierarchy : R1a1 from the west imposing on an east Asia local population) "altering" the general picture since the bones were taken from Kurgans if I'm not mistaken, but I'm not convinced anyway given their culture (related to the west and pastoralist) their supposed phenotypes (apparently, europoids different from their neighbors) which could really point to a "recent" migration and above all the overwhelming presence of west Eurasian mtDNA lineages (this wouldn't reflect any social bias, east Asian lineages should be visible from the start, like in the Tarim Xiaohe peoples).

Maju said...

Hi again, Wagg.

Regarding haploid lineages, we should not forget that the Khanty are an Ob-Ugrian population high in Y-DNA N, of undoubted East Asian origin ultimately (N's diversity is highest in some parts of China, rather to the south). Along with this Y-DNA N, mtDNA CZ and D also arrived to NW Eurasia, spreading somewhat into Europe too. Additionally, mtDNA G is often claimed to have colonized Central Asia very early on (a bit unsure myself but the claim is there and the lineage is old enough for that). Let's not forget either the relatively high frequency/diversity of Y-DNA D in Central Asia either, another arrival from the East probably long before Turkic migrations.

So my hypothesis is that there may well have been a confluence of migratory waves in that area of Central Asia/Siberia which is well reflected in the distribution of Y-DNA lineages P (specially Q but also R1) and N, both of which would seem to signal (together with C3 but this one only in the East) the colonization of the Eurasian North from the tropics, one from the West (P) and the other from the East (N). MtDNA to some extent at least is coincident with this double pattern.

But I can't be totally sure in regards to Central Asia anyhow because I'd like to see a deeper and better sampled analysis(better West Asian sample, larger South Asian one, Uzbek, Tibetan, Altaian samples), in order to have a clearer vision of the matter.

"I don't think the Tarim Basin is the best place to observe the situation of Central Asia before, during and after chalcolithic/bronze age (beside the fact that it's apparently difficult to access from the west, the south and the north, it's also a quite desertic place that we don't have prooves of settlement prior to bronze age, if I'm not mistaken)".

Actually (Chunxiang Li 2010) it is the oldest archaeological site in Central Asia to have been genetically tested, and it's from c. 2000 BCE (described as "Bronze Age" in the paper). The authors acknowledge that all the Y-DNA is Western (R1a1a) but that the mtDNA is mixed (in fact mostly Eastern, though there's doubt about what the R* may mean). The pigmentation markers strongly suggest deep brown to black hair color (unlike what happens in other Central Asian sites, see Jean Manco's list for details).

(continues)

Maju said...

...

"If, as it is supposed, the early Tarim inhabitants came from Afanasevo population"...

There's nothing to reject that. Just that the mtDNA is largely and probably mostly Eastern, so we'd be talking already of West-originated IE males mixing with local women at the edge of China proper.

My question is: which was the autosomal affinity of these IE guys with R1a?

After all now we know that R1a probably spread from South Asia and is not only a IE marker but something surely quite older, at least Neolithic and I'd say Paleolithic (some subclades may still be associated with IE expansions like R1a1a7 in Europe). So, if R1a has been "floating" in Central Asia and East Europe since so old (i.e. maybe since the scatter of Q northwards and eastwards or just slightly later) and that is also the case of N and the various mtDNA lineages, we can well think that their main autosomal affinity was with that Central Asian specific component now detected.

"Andronovo samples are almost completely west Eurasian during bronze age"...

Yes, I already acknowledged that. But Andronovo is very far West, technically in European Russia, even if slightly East of the Volga. It's not more Asian than the early Kurgan culture (Samara and such) and interacted most intensely with other East European IE cultures like Srubna.

It's timeline is also quite late, mostly of the 2nd millenium BCE, i.e. more recent than Afanasevo and the Tarim Basin sites.

Andronovo expanded eastwards, yes, but we may be seeing only the "invaders' fortresses" and not the bulk of the native peoples. I don't really know but this kind of potential bias cannot be ignored.

"IMO the early Tarim R1a1a probably came from south Siberia".

Altai? It's a possibility but we can't ignore the possibility that it arrived from Central Asia proper.

Whatever the case, it did not arrive straight away from Europe and the lineage carriers may not even have been at Europe ever if R1a was "floating" in the South-Central Asia-East Europe corridor since Paleolithic times.

Anonymous said...

Me again. I have to react on a few points :

"Regarding haploid lineages, we should not forget that the Khanty are an Ob-Ugrian population high in Y-DNA N, of undoubted East Asian origin ultimately (N's diversity is highest in some parts of China, rather to the south). Along with this Y-DNA N, mtDNA CZ and D also arrived to NW Eurasia, spreading somewhat into Europe too."

Don't you mean the Mansi * (the name you put in your article)?
Their mtDNA seem to be 63% west Eurasian and 37% east Asian and a lot of their hgs can indeed point to a paleolithical west eurasian movement east, but I have trouble associating hg such as T1 or K to some paleolithical movement from the west (some of their hgs could be a of more recent contact). Overall, these west Eurasian hgs are not really similar to the south Siberian ones (except a few, that could also be local since the paleolithic like U4, U5a1 and U2 for instance (but not necessarily either as they are also present in the east of Europe)) and the presence of U7, U8, H1, H2, H3 for instance really don't fit well the types of hgs of ancient south Siberia IMO.

* http://www.ncbi.nlm.nih.gov/pmc/articles/PMC379094/

"My question is: which was the autosomal affinity of these IE guys with R1a?"

Isn't this kind of revealing? :

http://1.bp.blogspot.com/_Ish7688voT0/Sgl-jY6gjjI/AAAAAAAABdQ/OsUUZB04OhQ/s320/triangular_siberian_pigmentation.jpg

"After all now we know that R1a probably spread from South Asia and is not only a IE marker but something surely quite older, at least Neolithic and I'd say Paleolithic "

Indeed, I realize that.

Anonymous said...

"(some subclades may still be associated with IE expansions like R1a1a7 in Europe). So, if R1a has been "floating" in Central Asia and East Europe since so old (i.e. maybe since the scatter of Q northwards and eastwards or just slightly later) and that is also the case of N and the various mtDNA lineages, we can well think that their main autosomal affinity was with that Central Asian specific component now detected."

The actual spread of r1a1a in central Asia is not that uniform and it's actually basically absent in "revealing" places. Not sure it works so well with this theory of a r1a1 continuum throughout central Asia since paleolithic, even though it makes sense on the paper (I used to think that way).

http://upload.wikimedia.org/wikipedia/en/thumb/a/a4/GlobalR1a1a.png/300px-GlobalR1a1a.png

In the mentionned studies, the "west eurasian" components makes 90-100% of the mtDNA lineages during bronze age (70-80 % overall if we add up bronze and iron age). Nowadays, if I'm not mistaken, the west Eurasian mtDNA lineages represent apparently like 28% of the gene pool of all central Asia.
Don't you think this central Asian component would be more representative of the population that flowed after bronze age and apparently changed a big part of the population of central Asia?
Unless you're thinking of a social bias (as I think the remains were mostly from Kurgan) that alters the big picture in the mentionned aDNA studies in favor of west Eurasian hgs ? ->

"Andronovo expanded eastwards, yes, but we may be seeing only the "invaders' fortresses" and not the bulk of the native peoples. I don't really know but this kind of potential bias cannot be ignored."

Yes I adressed the "social hierachy" bias thing that could be linked to Kurgan sepultures' remains but as I said earlier, the social bias would probably be not visible in the mtDNA (it would concern only the y-DNA hg (think of the Xiaohe mtDNA or even of the fact that there are more and more east Asian mtDNA hgsin south Siberia through the time but it still shows only R1a1 (except for one y-DNA C)).

Anonymous said...

"Actually (Chunxiang Li 2010) it is the oldest archaeological site in Central Asia to have been genetically tested, and it's from c. 2000 BCE (described as "Bronze Age" in the paper). The authors acknowledge that all the Y-DNA is Western (R1a1a) but that the mtDNA is mixed (in fact mostly Eastern, though there's doubt about what the R* may mean)."

Yes, I also adressed it earlier, my point is that the south Siberian samples, despite being from bronze age, are actually more likely representative of the substrate living there since chalcolithic (Afanasevo) than of some Andronovo-related wave (BTW look again at the R1a1 map and you'll see that the Altai R1a is pretty concentrated in that zone as if it really was the result of a particular specific migration).
IMO, typical of Afanasevo is a pretty safe assumption (It makes sense to conjecture that Europoid skeletons compared to the ones of Yamna and a culture with a link with Yamna appearing in 3,500 BCE in south Siberia + mtDNA so close of the European ones can easily be a strong signal of a population movement from the west).

"The pigmentation markers strongly suggest deep brown to black hair color (unlike what happens in other Central Asian sites"

Pretty logical as they were mixed peoples.

me : "Andronovo samples are almost completely west Eurasian during bronze age"...
you : "Yes, I already acknowledged that. But Andronovo is very far West"

My bad I wasn't clear enough, I was not only referring to Kazakhstan but also of the south Siberian samples of the bronze age timeframe, which obviously are mainly from the Afanasevo population substrate IMO (they're the ones that were 90% west eurasian during bronze age (the Kazakhstan ones were 100% west Eurasian between 1,300 BC to 700 BC IIRC, in the few samples tested, that is)).

BTW, why would you favor the fact that all these hgs had to travel from the west during the paleolithic while we have tracks of cultural (and linguitic) spreading from the west during chalcolithic/bronze age.
If you cringe on the idea of such population movement (probably not massive anyway, central Asia was probably never much populated compared to much of the other eurasian regions and this R1a bulge in the Altai certainly built up through time) what makes it more acceptable if it's during paleolithic? Is that because it can imply a very slow and gradual movement while the one from north of the black sea to south Siberia would implies at best a few centuries?

Anonymous said...

"Actually (Chunxiang Li 2010) it is the oldest archaeological site in Central Asia to have been genetically tested, and it's from c. 2000 BCE (described as "Bronze Age" in the paper). The authors acknowledge that all the Y-DNA is Western (R1a1a) but that the mtDNA is mixed (in fact mostly Eastern, though there's doubt about what the R* may mean)."

Yes, I also adressed it earlier, my point is that the south Siberian samples, despite being from bronze age, are actually more likely representative of the substrate living there since chalcolithic (Afanasevo) than of some Andronovo-related wave (BTW look again at the R1a1 map and you'll see that the Altai R1a is pretty concentrated in that zone as if it really was the result of a particular specific migration).
IMO, typical of Afanasevo is a pretty safe assumption (It makes sense to conjecture that Europoid skeletons compared to the ones of Yamna and a culture with a link with Yamna appearing in 3,500 BCE in south Siberia + mtDNA so close of the European ones can easily be a strong signal of a population movement from the west).

"The pigmentation markers strongly suggest deep brown to black hair color (unlike what happens in other Central Asian sites"

Pretty logical as they were mixed peoples.

me : "Andronovo samples are almost completely west Eurasian during bronze age"...
you : "Yes, I already acknowledged that. But Andronovo is very far West"

My bad I wasn't clear enough, I was not only referring to Kazakhstan but also of the south Siberian samples of the bronze age timeframe, which obviously are mainly from the Afanasevo population substrate IMO (they're the ones that were 90% west eurasian during bronze age (the Kazakhstan ones were 100% west Eurasian between 1,300 BC to 700 BC IIRC, in the few samples tested, that is)).

BTW, why would you favor the fact that all these hgs had to travel from the west during the paleolithic while we have tracks of cultural (and linguitic) spreading from the west during chalcolithic/bronze age.
If you cringe on the idea of such population movement (probably not massive anyway, central Asia was probably never much populated compared to much of the other eurasian regions and this R1a bulge in the Altai certainly built up through time) what makes it more acceptable if it's during paleolithic? Is that because it can imply a very slow and gradual movement while the one from north of the black sea to south Siberia would implies at best a few centuries?

Maju said...

"Don't you mean the Mansi * (the name you put in your article)?"

My error: it's actually the Khanty, as you can see in the graph. It does not seem to matter much because both are Ob-Ugric peoples from the same area, close cousins, but I had an error and, thanks to you, I have now corrected it. Thanks.

"Their mtDNA seem to be 63% west Eurasian and 37% east Asian and a lot of their hgs can indeed point to a paleolithical west eurasian movement east"...

Or north via Central Asia, as distinct from NW via West Asia, which would be the main flow. Hard to tell. Some lineages like U7 and even U4 may be talking of the time when early West Eurasians, coming from India/Pakistan, branched around the Caspian, with some branches (the most successful ones) heading West and others North. While the Ob Ugrians are not known for their Y-DNA Q, nearby Samoyedic(also Uralic) peoples (Nenets, Selkups) are instead.

"... some of their hgs could be a of more recent contact"...

I can agree with that. T1, K and maybe others are surely Neolithic era erratics. H3 also seems to imply a flow from further west in Europe, because it's a lineage not known to exist in West Asia and mostly concentrated in SW Europe and North Africa. But maybe is a distinct sub-branch from the time of the first colonizations, as happens with H2 IMO, which displays different Western and Eastern European sublineages.

"and the presence of U7, U8, H1, H2, H3 for instance really don't fit well the types of hgs of ancient south Siberia IMO".

U7 is a South/Central/West Asian lineage, very rare in Europe, so it does fit very well with my theory. There's no U8 nor H1 among the Mansi (I have already mentioned H2 and H3 in the previous paragraph).

My impression is from your link is:

· Clearly "Neolithic" lineages (K, T1): 6.2%.
· NE European specific lineage (U4): 16.3%
· Other clearly old European lineages (H2, H3, V): 10.2%
· Unclear WEA lineages (U5, T(x(T1)): 6.1%
· West Asian lineages (J): 12.2%
· South/Central Asian lineages (U7): 5.1 %
· East Asian lineages (A, C, D, F, G, M*?): 36.7%

Plus their Y-DNA pool is dominated by N, of East Asian origin, which should weight something too. My rule of thumb is: mtDNA tells of some 50% of the real overall ancestry, Y-DNA of some 25% and the rest is uncertain (alternatively 2/3 and 1/3, removing the uncertainty). It's just a very rough approximation but it generally works quite well.

"Isn't this kind of revealing?"

No because they are only measuring some sort of haploid DNA (mtDNA?), not autosomal DNA. It says nothing as such on the autosomal, i.e. overall population affinity, level.

Remember that haploid lineages, some 8 generations back (a couple of centuries only), only tell of less than 1% (0.78%) of the overall ancestry and that, because of patri/matri-locality these tend to concentrate in few lineages (this is particularly true of Y-DNA, of course), specially when population size is small (drift).

A guy can perfectly be R1a1 and his overall genetics (autosomal DNA) be 100% (or >99%) Chinese or Nigerian. Of course, this distortion tends to be smaller in whole populations but it can still happen, specially if we consider Paleolithic population history and relatively isolated and low density areas such as Central and North Eurasia. So I'd be wary to rely too much on haploid genetics alone when considering overall ancestry, specially if, as in this case, the autosomal results are contradictory.

Cheers.

Maju said...

I think that somehow I missed your last three posts, Wagg. Let's address them here.

"I was not only referring to Kazakhstan but also of the south Siberian samples of the bronze age timeframe, which obviously are mainly from the Afanasevo population substrate IMO".

You must mean the Altai-Khakassia individuals. Following again Jean Manco's synthesis, we see the following pattern:

1. Karasuk culture: 2 individuals: one U4 (NE European/Central Asian) and the other U5a1 (generic West Eurasian AFAIK).

2. Tagar culture: 10 individuals. All males R1a1 (East Europe/Central Asia/South Asia). Largest single mtDNA lineage is T3 (what is this?! Not found in Phylo Tree), followed by CRS (H or U or even HV*) and then all are singletons: I4 (a most rare sublineage of I), H5a (undoubtedly European), C, G2a and F1b (all three Oriental).

[Note on "T3": HVS-1 markers correspond with JT (126), T (294), T1 (189) and T2b3a (292). Correct me if I am wrong. Notice that the Mansy discussed above were also high in T(xT1), which may be this T3 lineage and have an specific Central Eurasian distribution].

So after detailed analysis only H5a seems unmistakably West Eurasian to the exclusion of Central Asia (one case: 10%), with the rest being only generically West Eurasian (in the wide sense) with some very strinking peculiarities of their own or NE Asian. Most of the mtDNA and Y-DNA pool can be thought as Paleolithic (or arguably Neolithic if you favor a shorter time frame) immigrants from the South or the East, rather than Bronze Age immigrants from Europe.

3. Tachtyk Culture (historical period): 5 individuals of which two mtDNA lineages are Western (CRS, HV and T1) and two are Eastern (N9a and C). The single Y-DNA is still R1a1 in spite of being in a time frame after the Turkic migrations, what suggests that the area remained genetically stable overall also for male lineages (in fact Altaians still have plenty of R1a).

(continues)

Maju said...

"BTW, why would you favor the fact that all these hgs had to travel from the west during the paleolithic while we have tracks of cultural (and linguitic) spreading from the west during chalcolithic/bronze age".

Not clear enough. We only know of Andronovo culture (loose sense) spreading, as well as some Kurgan affinities in other places (Afanasevo), which may well indicate ethno-linguistic affinity. But neither Andronovo nor any likely offshoot does reach so far NE and we must not forget that early Indoeuropeans arose mysteriously at the very edge of Europe in an area that the ancients considered to be Asia (the Volga was the real border, not the much smaller Ural river). They were distinct from other East European archaeological cultures such as those from the Dniepr-Don (probably rooted in the local Paleolithic of Gravettian tradition) and if they are effectively linked in Europe to the spread of R1a, that may well mean a more specific Central Asian (and ultimately South Asian) affinity than properly European. At least there's certain ambiguity in all this and we can't forget that the horse now appears to have been domesticated in Kazakhstan (Botai culture) rather than Russia or Ukraine, as well as the intense links with Uralic language populations of PIE (Indo-Uralic).

I've never thought of early IEs as specifically European but rather ambiguously Eurasian or Central Asian. I changed my impression with the genetic apparent findings of the decade but these new findings (South Asian origin of R1a, Botai culture, Indo-Uralic and the Central Eurasian specific autosomal cluster in partial association with WEA lineages) are making me retake the original suspicion I had about early IEs being only ambiguously European and having a Central Asian affinity which they spread in Europe to a limited extent (stuff like blood group B or partial epicanthic fold).

"central Asia was probably never much populated compared to much of the other eurasian regions"

Well, Altai has been populated all the time since the MP-UP transition, displaying an outstanding archaeological record of its own. The rest of the Steppe-Siberia area was surely less densely populated but it almost certainly had their own groups of nomads, from two (or even more) different origins criss-crossing their paths and surely coalescing into specific regional clusters such as the one we find here.

Further south, the Uzbek area has been populated since at least the Neolithic but I suspect a much older time line as because it must have been in the route from South/West Asia to the North. The archaeological record available is limited but the Altaian "Aurignacoid" culture seems to have affinities in Afghanistan and maybe Iran too for what I have read here and there, as well as some late expansion into Mongolia and NW China.

So, coupling this with DNA evidence such as Y-DNA Q and R1a (and maybe even the ubiquitous J2), as well as mtDNA U (U2, U7 specially but maybe also U4 and U5), HV/H, X and now T too it seems, I do think that there was a very early, though still somewhat unclear in the details, migration from the South-West Eurasian corridor in that direction.

...

Maju said...

...

"... and this R1a bulge in the Altai (certainly built up through time) what makes it more acceptable if it's during paleolithic?"

I haven't seen any specific data yet on Altaian or otherwise Central Asian/Siberian R1a and how it fits in the wider picture. The whole debate has been very shallow, focused on the European/South Asian duality and often simply assuming happily that R1a merely meant an IE expansion signal.

I am also guilty of this, admittedly, but now I realize that it must be older because we know of no archaeologically justified migration from South Asia in the NW direction after the original wave of colonization of West Eurasia c. 50-40 Ka ago. Furthermore, South Asian lithic industry becomes microlithic soon after those dates, and there's no sign of microlithism elsewhere until the Epipaleolithic. Of course one hypothesize that Western microlithism arrived from South Asia in a demic migration but there's no evidence of that I know of, rather instead evidence of local parallel evolution from non-microlithic industries such as Magdalenian.

"It makes sense to conjecture that Europoid skeletons compared to the ones of Yamna and a culture with a link with Yamna appearing in 3,500 BCE in south Siberia + mtDNA so close of the European ones can easily be a strong signal of a population movement from the west".

Yamna is the original Kurgan more or less: they are "Asians" in the East-of-the-Volga sense. Yamna is a second wave of IE expansion in Europe (limited to parts of Eastern Europe) and AFAIK there's no indication whatsoever of Western IEs, already established in Central Europe (main Western IE branch) and the Balkans (other diverse branches maybe leading to Greek, Albanian, etc.), gave them any sort of meaningful feedback (rather I suspect they fought against them, as the Corded Ware culture and it's precursors of Lubon and Globular Amphorae expanded eastwards dramatically from their Polish homeland, taking the place of the somewhat "Danubianized" Dniepr-Don "last stand" around modern Kiev).

We have to understand that Western IEs experienced high admixture and specially cultural "contamination" from the Danubian cultural area, which was still powerful for some time after their initial arrival (Baalberge culture). In that period, after the recession and fragmentation of Baalberge at the apogee of Baden culture (Danubian), they became at least partly European, properly speaking, by means of absorption of Danubian cultural elements such as burial in fetal position, etc. They remained primarily Kurgan (IE) culturally but quite less markedly than their cousins of Yamna and such, which in my opinion represent the most genuine IE group, leading to the Scythians and related Iranian groups of the steppe eventually.

Anonymous said...

"So after detailed analysis only H5a seems unmistakably West Eurasian to the exclusion of Central Asia (one case: 10%), with the rest being only generically West Eurasian (in the wide sense) with some very strinking peculiarities of their own or NE Asian. Most of the mtDNA and Y-DNA pool can be thought as Paleolithic (or arguably Neolithic if you favor a shorter time frame) immigrants from the South or the East, rather than Bronze Age immigrants from Europe."

But since all these hgs (IIRC) are found in Europe, it's not really resolving anything for sure (the fact that some particular hg have a specific central Asian "physionomy" don't necessarily tell much, it also depends of the time of the mutation and we can't be sure of the original location of its first bearer either - but yes I admit that some of them are probably "autochtonous" to central Asia for a long time - I never said the gene pool had to be 100% from the chalcolithic Pontic steppes).

I mean, even H6 is found in Europe :

"In contrast, H2 and H6 are both common in eastern Europe and the Caucasus, although there are hints that they may have dispersed from western Europe. In particular, the basal type of H6 is exclusively European, and there is a single derived type that is common in eastern Europe and the Caucasus." (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC540273/)

It's interesting that the Caucasus is mentionned because in the aDNA kazakhstan study (Lalueza-fox et al 2004) several mtDNA hgs were apparently particularly common in the Caucasus (I believe it's also the case of U3 that was found several time in 2,000 yrs old (IIRC) Xinjiang remains).

Anonymous said...

"Not clear enough. We only know of Andronovo culture (loose sense) spreading, as well as some Kurgan affinities in other places (Afanasevo), which may well indicate ethno-linguistic affinity."

That's a point of view... With Afanasevo we're talking of (apparently) Europoid pastoralist individuals (r1a1a + almost completely west Eurasian mtDNA hgs) with similar sepultures' style than in Yamna (even the ritual artefacts resemble the one from the pontic steppes), with potteries that are also matching the Yamna style IIRC, probably the use of wheel even if not directly attested, the use of domestic horses, also knowing the copper metallurgy (also gold and silver) - and yet at a huge distance from the pontic steppes. It's quite tempting to see a direct population movement. (source : a Mallory book)

What is your explanation for the origin of Tocharian language?
Don't you think the date of 3,500 BC for Afanasevo could go well with the date of the surmised IE cultures appearing in the north half of Europe and could explain the fact that Tocharian resembles more the western IE languages (roughly the same stage of developpement of the IE language).

For after all, the eastern IE language has many similarities with the western IE languages (Germanic, celtic, slavic, etc... + Hittite IIRC - it also has a few finno-ugric borrowings apparently). These languages' ancestors had to be derived from the same source within a few centuries span, I assume.

Thanks for your thorough replies.
I'll probably drop a comment on the rest later. Good night.

Maju said...

"I mean, even H6 is found in Europe"...

Sure, I think it scattered at the colonization of Europe, together with other H, some of which arrived to Central Asia, specially H8 and some less known ones.

H6 is not listed in the paper you linked as part of the Mansy genetic pool. There's no mention of H6 anywhere but H*, H3 and H2. Probably all arrived from Europe but when? IMO early in the UP. There's also no mention to H6 in the list of Kayser 2009 haplogroups I have seen.

"With Afanasevo we're talking of (apparently) Europoid pastoralist individuals (r1a1a + almost completely west Eurasian mtDNA hgs)"...

I don't think the case is clear at all. R1a1a is not "Europoid" but has Central Eurasian distribution between India and Scandinavia, and the mtDNA is, as I discussed above, only occasionally linked to Europe clearly.

"What is your explanation for the origin of Tocharian language?"

I presume it's related to Afanasevo, yes. But that may mean little regarding the genetic pool: languages are not genes. A population can totally switch language in very few generations (my great-grandparents were Basque speakers with difficulty when using Spanish, my father is Spanish Speaker who can barely speak Basque) and they will invariably switch (if they do) towards the language of the elites of the moment (albeit typically creolizing it into a new dialect).

"For after all, the eastern IE language has many similarities with the western IE languages"...

I don't think so. It is a total isolate (or even maybe two) inside Indoeuropean. In all classifications I have seen it stands apart being the first one of all IE languages to split off. That means that:

A. If it divergence must be measured only in time, it broke apart before Albanian and Hittite, and certainly before the main Western IE branch (from which Celtic, Latin, Germanic and Balto-Slavic sprung).

... or ...

B. If divergence is also caused by deformation with creolization (which I think it is the case), it may have diverged at a similar time but be much more affected by this substrate influence.

In case A, as Western IE broke apart c. 3500-3000 BCE, Tocharian would have diverged with the arrival of Neolithic or something like that. In case B, the impact of the Altai/Sinkiang substrate would have bee stronger than the impact of the Danubian one on Western IE, something hard to explain.

So I do suspect that Tocharian is a pretty old offshoot of IE - but the info we have is not really conclusive.

"+ Hittite IIRC"...

Hittite is usually considered a totally different branch. One of the three or four "Central IE" branches, whose classification and origins are most unclear, the Anatolian one to be specific. This particular branch possibly arrived via the Caucasus after the Maikop culture period.

"Thanks for your thorough replies".

Thanks to you, I'm enjoying the discussion. It's a most interesting debate.

Anonymous said...

Me again.

So, how do you interpret the fact that during bronze age 100-90% of the samples from Kazakhstan and south Siberia (75-80% if we add up bronze age + iron age) while nowadays west eurasian mtDNA lineages is apparently closer to 30 %.
Shouldn't this mean that this detected central Asian component (that includes parts of north Siberia) is more related to the Asian lineages that apparently flowed after bronze age than to these west eurasian hgs (I also mentionned why I doubt the social bias thing on the mTDNA side, earlier)?

"early Indoeuropeans arose mysteriously at the very edge of Europe in an area that the ancients considered to be Asia (the Volga was the real border, not the much smaller Ural river). They were distinct from other East European archaeological cultures such as those from the Dniepr-Don (probably rooted in the local Paleolithic of Gravettian tradition)"

"Yamna is the original Kurgan more or less: they are "Asians" "

Well, Asia or Europe is irrelevant here as I was pointing to their (alledgely) Europoid skeletal remains similar to those of Yamna and Sredny stog (according to Mallory, from Russian archeologists' studies). A fronteer is a human convention. It doesn't block human movement (unless the fronteer uses a natural obstacle but it's never totally blocked).
The nowadays Russian are autosomally close to the west European and they live in these yamna/samara area. The fact that their culture was quite different from the population living west of them, don't imply much to me concerning their genetic classification.

"At least there's certain ambiguity in all this and we can't forget that the horse now appears to have been domesticated in Kazakhstan (Botai culture) rather than Russia or Ukraine"

Yes, the Botai domesticated the horse by 3,500 BC. But the general picture is still unclear AFAIK, the Dereivka site of Ukraine is supposed to have earlier tracks of horse domestication (though it's not totally clear IIRC).
And what about the scepters in the shape of horse head foun first in the pontic steppes (the oldest are around 4,000 BC IIRC) then in the south eastern Europe after the arrival of people from the steppes?

Anonymous said...

What do you think of the R1a repartition map, especially concerning the south siberian region * ?
The mtDNA and their phenotypes go against the idea that its a remnant of the early R1a1a migration north, separated from the west R1a1a.

* http://upload.wikimedia.org/wikipedia/en/thumb/a/a4/GlobalR1a1a.png/300px-GlobalR1a1a.png

I could believe an hg such as U7 came with the early R1a1a into Europe/central Asia (after all 2 Danish U7 in the aDNA of the Viking age were found and in the north west of Europe it seems there are some ancient R1a1a* there IIRC and all this could be lingering tracks of the arrival of R1a1a), but obviously the European and central Asian R1a are associated with west mtDNA eurasian hgs in central Asia. It doesn't fit very well with this idea of these central Asian R1a1a being the remains of the initial R1a1a spread from south Asia IMO (I think that's an idea you proposed in an earlier post, right?).

And as I mentionned earlier, the fact that several mtDNA hgs seem particularly linked to the Caucasus (and as such also the adjacent regions) could also be a hint of a post-paleolithical movement east, especially since mtDNA hg such as K is also found up to Bronze age Xinjiang and south Siberia.

Anonymous said...

me : "I mean, even H6 is found in Europe"...
you : "H6 is not listed in the paper you linked as part of the Mansy genetic pool. There's no mention of H6 anywhere but H*, H3 and H2. Probably all arrived from Europe but when? IMO early in the UP. There's also no mention to H6 in the list of Kayser 2009 haplogroups I have seen."


I was refering to the south Siberian samples and your comment that only H5a could be clearly seen as "european". H6 is really present in the south Siberian samples of Keyser 2009, it's sample S13 in the article (I have got the PDF).

me : "With Afanasevo we're talking of (apparently) Europoid pastoralist individuals (r1a1a + almost completely west Eurasian mtDNA hgs)"...

you : "I don't think the case is clear at all. R1a1a is not "Europoid" but has Central Eurasian distribution between India and Scandinavia, and the mtDNA is, as I discussed above, only occasionally linked to Europe clearly."


I badly expressed myself. The general idea was that their mtDNA haplogroups gives more strength to the idea that they came from the west during chalcolithic, I was more thinking of the cultural signals from far away to th west and their phenotypes that have been found similar to those of Yamna and sredny stog. I also think the R1a1a map I gave the link (from the 2009 study that determined R1a1a7 and the others) support that view more than yours.

"What is your explanation for the origin of Tocharian language?"
I presume it's related to Afanasevo, yes. But that may mean little regarding the genetic pool: languages are not genes. A population can totally switch language in very few generations.


I realize that, still, the archeology + the distribution of R1a + mtDNA makes a good candidate for some migration porcess IMO.

Anonymous said...

me : "For after all, the eastern IE language has many similarities with the western IE languages"...

you : "I don't think so. It is a total isolate (or even maybe two) inside Indoeuropean. In all classifications I have seen it stands apart being the first one of all IE languages to split off. That means that:"


I'm surprised you say that because I can't remember seeing it as the first language to separate even though I did see it as the second to separate after Hittite (but I can't say I'm a specialist of this matter). Generally, Hittite is considered the most archaic.
Still, Tocharian do have MANY similarities with the western centum languages in its vocabulary (känt = 100 ; mälk : to milk ; ko : cow (and german kuh) ; warto : garden (p.gmc *gardon) ; tápärk : now (-> Russian teper'); wrauña : crow (-> Russian : voron); luks : to illuminate (-> latin lux) etc... the list is long) As for its specificities, let's not forget the first tracks of Tocharian language are about 500 AD.
If it really derived from the language of Afanasevo, that's 4,000 yrs later...

" Hittite is usually considered a totally different branch."

It's generally suspected to date from before proto-indo-european. The first Kurgan wave ravaging Bulgaria around 4,400 BC (destruction of Karanovo VI) is thought by some to be at its origin.

Maju said...

"So, how do you interpret the fact that during bronze age 100-90% of the samples from Kazakhstan and south Siberia (75-80% if we add up bronze age + iron age) while nowadays west eurasian mtDNA lineages is apparently closer to 30 %".

I don't know that. From memory (lost my bookmarks in a computer crash, sorry) Central Asian mtDNA is at least 50% Western, with mtDNA H being clearly dominant. Any particular reference?

However Kazakhstan, which is quite unequally sampled (all samples are from the East AFAIK) seems to be the region where Turkic demic replacement was more complete. So I'd look at other populations like Altaians, Southern Central Asians and West Siberians for an improved reference.

"A fronteer is a human convention. It doesn't block human movement (unless the fronteer uses a natural obstacle but it's never totally blocked)".

True but the Volga was very wide in the Paleolithic and, while not impracticable, it may have been an effective border. I'm not sure but I think it's a relevant issue to be explored, specially considering the apparent Asian affinity of both Uralic and Indoeuropean early groups.

"The nowadays Russian are autosomally close to the west European and they live in these yamna/samara area".

Yes but we don't have specific samples from Samara area, do we? Anyhow it's perfectly possible that Russians made an ethnic cleansing there, as they did in other regions, when they conquered it from the Tatars (Samara was Tatar/Mongol country not so long ago - Russians expanded into that area but are original from farther West as any map of Kievan Russia, prior to Mongol expansion, will tell you).

Russians from the North are clearly distinct and look Uralic peoples in fact. It's possible that SE European Russians also have that kind of admixture but I can't tell for sure. I'd be interested in knowing, of course.

"The fact that their culture was quite different from the population living west of them, don't imply much to me concerning their genetic classification".

It potentially does. Because we are trying to discern the origins and we know nothing of possible Paleolithic inhabitants of the Samara area. So they (IEs) appear out of nowhere in the Neolithic and they are not related to their Western Neighbors and carry a lineage with clear South Asian origins (R1a) and likely Central Asian scatter. Ok, not conclusive but intriguing to say the least.

"But the general picture is still unclear AFAIK, the Dereivka site of Ukraine is supposed to have earlier tracks of horse domestication (though it's not totally clear IIRC)".

It's possible. I'm not totally persuaded. In fact I do suspect a Magdalenian domestication of the horse too but it may have happened several times in different places and even be lost thereafter.

But Botai show clear indisputable signs of horse domestication and riding (nose rings, bridles), which make up for a clear early case not just of domestication but of a full economy based on the horse, as it's attributed to early IEs soon after.

"And what about the scepters in the shape of horse head foun first in the pontic steppes (the oldest are around 4,000 BC IIRC) then in the south eastern Europe after the arrival of people from the steppes?"

Honestly, the ones I have seen (you mean the East Balkans, right?) look like crocodiles and hippopotamuses, what is even more striking (an Egyptian import or I'm just imagining things and they were merely bad artists?). IDK.

In any case it's within the time frame of IE expansion, which we know was based on horse riding. It does not question Botai horse domestication at all.

...

Maju said...

...

"What do you think of the R1a repartition map, especially concerning the south siberian region * ?"

That Kazakhstan may have suffered widespread replacement in the Turkic expansion. Being a steppe-desert piece of land is even reasonable. However Kazakhstan's genetics are generally extrapolated from single samples from the East, what is not as good as we would like.

The map is not totally accurate because one of the areas with highest R1a density is Poland and in the map it appears as not top. Russia is not as dense surely and Sweden is also higher probably. I'm unaware of R1a among the Sakha (Yakuts), who are essentially Y-DNA Q or N (can't recall, I think it's Q with very few founding fathers).

The map anyhow would illustrate well enough the possible corridor of arrival into Europe from South Asia via Central Asia, and a branch heading to Altai. As I said before, Kazakhstan is quite irrelevant as it's essentially empty space.

"I could believe an hg such as U7 came with the early R1a1a into Europe/central Asia"

Maybe. U7 is a rare lineage in Europe anyhow. Its highest frequencies are in NE India and Iran.

"but obviously the European and central Asian R1a are associated with west mtDNA eurasian hgs in central Asia".

Only to a very limited extent, as I have discussed previously.

"It doesn't fit very well with this idea of these central Asian R1a1a being the remains of the initial R1a1a spread from south Asia IMO (I think that's an idea you proposed in an earlier post, right?)".

Essentially yes. There may be specific sublineages that have arrived from Europe or whatever but that's what I'm thinking now... while I wait for further data on Central Asian and South Asian R1a, further resolution of the R1a tree that should clarify things up to a point.

"And as I mentionned earlier, the fact that several mtDNA hgs seem particularly linked to the Caucasus (and as such also the adjacent regions) could also be a hint of a post-paleolithical movement east"...

Again: it's the anecdote rather than the bulk of the data. Was there some minor flow from Europe? Possibly. Was there an overwhelming flow from Europe? Probably not. Was this flow from Europe a purely Chalcolithic event? Probably not either.

"... especially since mtDNA hg such as K is also found up to Bronze age Xinjiang and south Siberia".

But it's not even clear if K is European by origin or West Asian. I am in doubt but mainstream genetics tend to see it as a Neolithic arrival to Europe AFAIK. After all it's a direct relative of U8b, which is only found in the Eastern Mediterranean (Italy, Jordan).

And, in any case, it's one individual out of many: an erratic. Can you build a theory on that? I cannot.

...

Maju said...

...

"H6 is really present in the south Siberian samples of Keyser 2009, it's sample S13 in the article (I have got the PDF)".

Ok, I was ignoring the group labeled as "Andronovo culture" because Andronovo site is in Europe. But actually these are from as far East as the Krasnoyarsk Krai, so guess that fair enough. However even in this group, only two individuals display clearly West Eurasian lineages (H6 and T1), while there are many others that look more or less specific to the region: U4, U2, T4 and Z1.

U4 is the most arguable. It is not found in West Europe and seems a peculiarly steppary lineage. Can't be specifically IE because it's been found in subneolithic peoples of the Baltic but it has that kind of scatter through the far North that suggests it had a Central Asian pathway, like U7 and U2 (and unlike U5, U3, U6 and U8 possibly).

"... I was more thinking of the cultural signals from far away to th west and their phenotypes that have been found similar to those of Yamna and sredny stog".

What do we know of the phenotypes of Paleolithic Altaians? Also the deductions on genetic traits of phenotype may be misleading (only one known SNP seems to be quite accurate for blue eyes, for instance, the rest is vague "probability" and often do not even have a clear Western affinity). I just mentioned this at Dienekes a couple of days ago.

"I realize that, still, the archeology + the distribution of R1a + mtDNA makes a good candidate for some migration porcess IMO".

Some yes. But from "some" to "a lot" there is great difference.

"I'm surprised you say that because I can't remember seeing it as the first language to separate even though I did see it as the second to separate after Hittite".

You seem to be correct on that. My memory failed me. Still it is clearly a branch that broke apart before the main West-East division and/or was heavily creolized/transformed after split.

"Still, Tocharian do have MANY similarities with the western centum languages in its vocabulary"...

For what I know, the Tocharian (and other dead languages) adscription to Centum group has been challenged. And anyhow the mainstream theory seems to be that Centum is old PIE and Satem a secondary fashion. Hard to tell.

If I had to follow this map, then I'd argue that Centum is actually a Danubian borrowing and that Satem is the true IE thing. But who knows? It's not like it is central in language phylogeny in any case.

I even recall certain English linguist who argued that Basque "ehun" (100) was related to English "a hundred" (a hund(-dred)) somehow [that and the well known Basque borrowing Jingo (< Jainko, God, surely adopted in the Hundred Years' War) were the only ones he could find, though 'bizarre' is too for sure, via Sp. 'bizarro' and I have a small list of others such as 'ash' and 'ill/kill']. It's not totally impossible so things are kinda unclear.

"[Hittite] It's generally suspected to date from before proto-indo-european. The first Kurgan wave ravaging Bulgaria around 4,400 BC (destruction of Karanovo VI) is thought by some to be at its origin".

It cannot be. That wave would have given birth to Balcanic languages such as Thracian, Albanian and maybe Greek. (My dates are from almost one millennium after yours btw but same event - check here for related Central European chronology, it was just before Baden culture, which is dated not earlier than 3600 BCE).

The Anatolian branch should have coalesced in the Caucasus, maybe via creolization (accelerated drift in a mixed cultural group). It may also be an older offshoot of some sort.

Anonymous said...

I'm not fond of very long internet discussions, particularly through blog comments, especially since we know the result, so I'll just reply a few points and leave it that way. Thanks for your input, I guess we'll have to wait a little more in order to have certitudes about it all.

"I'm unaware of R1a among the Sakha (Yakuts), who are essentially Y-DNA Q or N (can't recall, I think it's Q with very few founding fathers)."

Yes Yakuts do have R1a (but also R1b and I, mtDNA H too and a U5b1b was also found).

http://s155239215.onlinehome.us/turkic/60_Genetics/TurkicGeneticsGraphs.htm

"I was ignoring the group labeled as "Andronovo culture" because Andronovo site is in Europe."

Andronovo is more an archeological horizon, a group of related culture that spread up to south Siberia IIRC.

"What do we know of the phenotypes of Paleolithic Altaians?"

The Mallory allegations were based on the works of Russian specialists whose name were not mentionned in the book (but I remember Polak (the user at Dienekes') mentionning their names and studies so it must be findable. Too lazy to search now).

"If I had to follow this map (centum/satem repartition), then I'd argue that Centum is actually a Danubian borrowing and that Satem is the true IE thing."

A popular theory among kurganists is that the satem languages are an innovation (starting maybe at most at 2,500 BC) spread by subsequent conquests/population movements and that the languages on the fringe (west IE languages) and Tocharians are more conservative of the oldest IE form.

For what I know, the Tocharian (and other dead languages) adscription to Centum group has been challenged

I'm no specialist on the question, but spontaneously I'd say "Kante" and "känt" (= hundred), in the 2 tocharian languages, seem typically centum.

Also, I'm not sure why the 4,400 BC wave in Bulgaria from the pontic steppes + alleged apparition of south-east europe related cultural elements in Anatolia around 3,000 BC + destructive population movement in west and south central Anatolia around 2,700 BC couldn't have anything to do with populations of Anatolian branch speakers (elements mentionned in a Mallory book I've read), even though it's totally hypothetical. I didn't get your explanation, maybe the lack of depth in my knowledge in archeology is to be blamed.

Maju said...

"Yes Yakuts do have R1a (but also R1b and I, mtDNA H too and a U5b1b was also found)".

At very low levels all. That charter says they are like 90% N3!

The link anyhow is a nice reference, thanks.

"A popular theory among kurganists is that the satem languages are an innovation (starting maybe at most at 2,500 BC) spread by subsequent conquests/population movements and that the languages on the fringe (west IE languages) and Tocharians are more conservative of the oldest IE form".

I already said that with other words. However since neither Tocharian nor Anatolian are clearly centum, nor is Albanian (another possibly old branch - Thracian?), we have to consider it may well be a European-specific creole evolution.

Alternatively it's worth to consider that, after all, languages do not exist in a frozen uniform academic form but evolve along geography and time. It's also possible that the original PIE had more than one dialect, be this a difference in the Samara area or between both sides of the Volga after the Sredny-Stog expansion.

It's almost a must because, after all, Balto-Slavic is Western IE, so the very seed of Western IE was surely both satem and centum.

"I'm no specialist on the question, but spontaneously I'd say "Kante" and "känt" (= hundred)".

Maybe but why would a bunch of seminomadic barbarians need a word for 100, to count sheep? It's not about this word but a whole lot of phonetic traits which make the two groups. Kant in fact makes me think of cantidad (quantity) in Spanish and kam (how much) in Turk... it might well not mean 100 but "amount, lot".

Anyhow, if you read on the Centum-Satem divide, at Wikipedia for instance, it's said that Tocharian is in fact different:

"Tocharian combined all rows into a single velar row by merging the palatovelars into the velars as in Centum and the labiovelars into the velars as in Satem".

It seems that the creole proto-Tocharian group simply did not like velar diversity of either dialect.

In turn Anatolian did not even experience the PIE-Centum change so it is clearly pre-Centum:

"The Proto-Anatolian language apparently did not undergo either the Satem or the Centum sound change, as the velar rows remain separate in Luwian. The closely related Hittite is a Centum language, which it may have become secondarily, but the exact sequence is unclear".

"Also, I'm not sure why the 4,400 BC wave in Bulgaria from the pontic steppes + alleged apparition of south-east europe related cultural elements in Anatolia around 3,000 BC + destructive population movement in west and south central Anatolia around 2,700 BC couldn't have anything to do with populations of Anatolian branch speakers"...

I am unaware of any Western influx in the Hittite genesis. Anatolian seems unrelated to Greek, Thracian, etc., which would be the derivatives of the Cernavoda expansion in the East Balkans.

"I didn't get your explanation, maybe the lack of depth in my knowledge in archeology is to be blamed".

I may be the one lacking info here (IDK), I am just totally unaware of any flow from the Balkans into Anatolia before Tracho-Phrygians and Greeks. In general the mainstream theory on the origin of Anatolians links them to Maikop but the issue does not seem to be sufficiently clear.

An additional indication on this matter is the linguistic distance: while Greek and possibly "Daco-Thracian" (one or two subfamilies?) may be related to Western IE (Ger-Cel-Ita-Bal-Sla) (Mycenaean Greek sounds like Latin often), this is clearly not the case with Anatolian.

I would think that if the Hittites arrived from the Balkans, we'd see more clear direct correlations with Balcanic IE and even Western IE, when in fact Anatolian stands totally on its own.

I guess it's an open question anyhow.

Anonymous said...

Hi.
Reading one of your comment at Dienekes' I understand that you don't have the "keyser et al 2009" PDF (I should have guessed from your comments here).
I can send it to you via e-mail, like you sent me the R1b-V88 one. Are you interested?

Maju said...

Hi, Wagg. Thanks a lot but I already have it. It happened that with a PC crash I lost my bookmarks and stored materials a couple of weeks ago but I have recovered some already.

Thanks anyhow.

Mason said...

Here's a new paper:

Toward a more uniform sampling of human genetic diversity: A survey of worldwide populations by high-density genotyping (Xing et al. 2010)

http://viewer.zoho.com/docs/mrcBW

http://eurogenes.blogspot.com/2010/07/thirteen-new-sample-sets-make-their.html

Maju said...

Very interesting, Natsuya, thanks a lot. :)

Maybe most important is that the NJ tree establishes a division in Eurasia as follows:

1. Between South/West Eurasia and East Asia/America

This is very consistent with haploid genetics, both Y-DNA and specially mtDNA. At least the way I see it.

2a. In the eastern branch, the first to diverge (or the ones isolated for the longest time, if you wish) are Native Americans. Nepalis hang from a higher branch but they are a mixed population, what really dismisses their comment about Amerindians having a Central Asian affinity. I don't see why it would be that way at all.

Further down in the Eastern branch, it seems a division happened between NE Asians and SE Asians with the "border" running through modern China.

2b. In the western branch, the first division is between Southern Indians (excluding Brahmins who are also mixed) and the rest. Even with the caveats of the Bengal area not being sampled and that there may have been some back-migrations into South Asia, it seems reasonable to assume this represents a separation before the colonization of West Eurasia, i.e. still in South Asia.

This tree seems to satisfy my expectations, as does the one at supp. fig. 1 (very similar). However the supp. fig. 2 is totally different. Maybe it is distorted by the North African/West Asian admixture in Europeans and South Asians? It is quite intriguing and difficult to answer.

Things like Basque hanging from the French make perfect sense but that Italians hang from Basques is more difficult to understand, as is that the two Tuscan samples occupy very different positions in that tree.

I'll probably write something tomorrow. Thanks again.

Mason said...

You're welcome, Maju.

Looking forward to your introduction of the new paper.

Best,
Natsuya