September 04, 2014

Y chromosomes and mtDNA of early farmers from Hungary

A new preprint has just appeared on the bioRxiv. It's free to read so I'll just summarize some results. First:
The haplotype of the Mesolithic skeleton from the Croatian Island Korčula belongs to the mtDNA haplogroup U5b2a5 (Dataset S3). The sub-haplogroup U5b has been shown to be frequent in pre-Neolithic hunter-gatherer communities across Europe [28–30,32,33,45,46]. 
Contrary to the low mtDNA diversity reported from hunter-gatherers of Central/North Europe [28–30], we identify substantially higher variability in early farming communities of the Carpathian Basin  including the haplogroups N1a, T1, T2, J, K, H, HV, V, W, X, U2, U3, U4, and U5a (Table 1). Previous studies have shown that haplogroups N1a, T2, J, K, HV, V, W and X are most characteristic for the Central European LBK and have described these haplogroups as the mitochondrial ʻNeolithic packageʼ that had reached Central Europe in the 6th millennium BC [36,37]. Interestingly, most of these haplogroups show comparable frequencies between the STA, LBKT and LBK,
N1a is the "signature group" of the LBK based on previous publications and now it seems that it was also found in the Starcevo culture of Hungary. The mtDNA PCA plot (right) shows clearly that the Hungarian farmers are very similar to the German ones so it seems that the LBK is a direct outgrowth of the Carpathian Neolithic; some earlier models of "demic diffusion" argued that Neolithic farmers spread slowly across Europe, picking up hunter-gatherer ancestry as they went along, but now it seems that at least in the Hungary->Germany part of this journey interaction with hunter-gatherers was minimum.

mtDNA change over time in Europe is pictured in Figure 3 (left) showing a shared haplotype analysis. The Y-chromosome data genetic distance is shown on the right and shows the Balkan-Anatolian-Caucasian-Mesopotamian relationship of the early farmer Y-chromosomes. Practically, this is due to haplogroup G2a (and especially G2a2b), which has turned up in lots of ancient European farmers (including the famous Iceman):
Three STA individuals belong to the NRY haplogroup F* (M89) and two specimens can be assigned to the G2a2b (S126) haplogroup, and one each to G2a (P15) and I2a1 (P37.2) (Dataset S3, S5). The two investigated LBKT samples carry haplogroups G2a2b (S126) and I1 (M253). Furthermore, the incomplete SNP profiles of eight specimens potentially belong to the same haplogroups; STA: three G2a2b (S126), two G2a (P15), and one I (M170); LBKT: one G2a2b (S126) and one F* (M89) (Dataset S5).
I believe this is the first ancient finding of haplogroup I1 which attains a peak in modern Swedes. This might be useful to those who have tied this to Germanic migrations because of this, as it was already in Central Europe with the earliest farmers.

Surprisingly, Y chromosome haplogroups, such as E1b1b1 (M35), E1b1b1a1 (M78), E1b1b1b2a (M123), J2 (M172), J1 (M267), and R1b1a2 (M269), which were claimed to be associated with the Neolithic expansion [23–25], have not been found so far in the 6th millennium BC of the Carpathian Basin and Central Europe. Intriguingly, R1a and R1b, which represent the most frequent European Y chromosome haplogroups today, have been reported from cultures that emerged in Central Europe during the 3rd/2nd millennium BC, while a basal R type has been reported from a Palaeolithic sample in Siberia [60] in agreement with a proposed Central Asian/Siberian origin of this lineage. In contrast, G2a has not been detected yet in late Neolithic cultures [42,43]. This suggests further demographic events in later Neolithic or post-Neolithic periods.
A cautionary tale against over-reliance on modern distributions to trace ancient origins.

Considering the entire set of 32 published NRY records available for Neolithic Europe thus far, the low paternal diversity is indeed quite remarkable: G2a is the prevailing haplogroup in the Central European and Carpathian Basin Neolithic, and in French and Iberian Neolithic datasets [36,40,41]. There are only two exceptions, namely one E1b1b (V13) [41] individual from the Avellaner cave in Spain (~5,000-4,500 BC), and two I2a [40] individuals from Treilles, France (~3,000 BC).


Tracing the genetic origin of Europe's first farmers reveals insights into their social organization

Anna Szécsényi-Nagy et al.

Farming was established in Central Europe by the Linearbandkeramik culture (LBK), a well-investigated archaeological horizon, which emerged in the Carpathian Basin, in today's Hungary. However, the genetic background of the LBK genesis has not been revealed yet. Here we present 9 Y chromosomal and 84 mitochondrial DNA profiles from Mesolithic, Neolithic Starčevo and LBK sites (7th/6th millennium BC) from the Carpathian Basin and south-eastern Europe. We detect genetic continuity of both maternal and paternal elements during the initial spread of agriculture, and confirm the substantial genetic impact of early farming south-eastern European and Carpathian Basin cultures on Central European populations of the 6th-4th millennium BC. Our comprehensive Y chromosomal and mitochondrial DNA population genetic analyses demonstrate a clear affinity of the early farmers to the modern Near East and Caucasus, tracing the expansion from that region through south-eastern Europe and the Carpathian Basin into Central Europe. Our results also reveal contrasting patterns for male and female genetic diversity in the European Neolithic, suggesting patrilineal descent system and patrilocal residential rules among the early farmers.



bellbeakerblogger said...

Hell or highwater, I will have the Bell Beaker genetics page up today.

It is more or less a genetic explanation or the emergence of the Beaker phenomenon.

Lathdrinor said...

Mmm, still no R. The mid-late Neolithic are shaping up to be an event horizon for populations in West and East Eurasia. There is a vast gap in Y-DNA distributions before and after ~4000-5000 BP, although this paper along with others have shown that G2a, I1, I2, and even E1b1b were all present in Europe during the mid Neolithic, albeit not necessarily at the levels and locations they are at today. To a degree, then, 'survivalist' narratives are sustained. But to survive is not to thrive. For the ultimate winners of the Y contest, we have to wait till the late Neolithic expansions.

Daniel Szelkey said...

So it is now obvious that the Linearback had hunter-European blood from the fathers side as they had I y-dna.

Alberto said...

Yes, still no R among the Neolithic farmers. But still a bit early to conclude that R1b came to Western Europe during the Bronze Age with the Indo-European invasions. After all, how much Y-DNA do we have from Mesolithic (or earlier) Western Europe? One sample comes to mind (La Brana-1, and it's C6!). Any other sample that I'm forgetting (not counting Sweden, since it's not a country currently dominated by R1b, but by I1)? With one sample it's just too early to jump to conclusions (and besides, it has not been yet found in HG from Eastern Europe either, unlike R1a).

What does seem clear by now is that Mesolithic populations from North, Central and Eastern Europe (not so conclusively Western E) were almost exclusively mtDNA haplogroup U (U5, mostly). But nowadays in Lithuania (as a good example of Mesolithic Europeans) it only accounts for 11% (15% if we add U4), roughly the same as the north of Spain while H alone is 45% (also the same as the north of Spain).

It would be interesting to know if the H mtDNA from Lithuania (and rest of Northern and Eastern Europe) came from the Neolithic Farmers or from Iberia (where it has been found in Paleo and Mesolithic times).

Slumbery said...

Daniel Szelkey

Either that or YHg-I was present in the Mesolithic Asia Minor to begin with. Or the combination of both. Well, we do not even know accurately or for sure where the LBK people came from...

BTW, the theory that the current distribution of some YHg-I sub-branches is a result of some "Eastern Germanic tribes caught up in the Slavic expansion" - as someone said in a comment here - seemed to be a bit surreal for me anyway.

Grognard said...


A study a while back shows H has been heavy in western europe since early neolithic. Since H and r1b are pretty heavily linked coming in the bronze age is not really possible.

Blasonario Cremonese said...

I'm sorry Alberto, but Lathdrinor didn't speak about R1b... he only spoke about R. So... why did you point immediatly your finger against R1b? Excusatio non petita, accusatio manifesta...

For what we know, R1b could have come from Mars... until, obviously, we will find any evidence of its existence in Paleolithic or in Neolithic sites.

It's a great discover the I1 sample.

eurologist said...

"some earlier models of "demic diffusion" argued that Neolithic farmers spread slowly across Europe, picking up hunter-gatherer ancestry as they went along, but now it seems that at least in the Hungary->Germany part of this journey interaction with hunter-gatherers was minimum."

While I agree that the speed of spread varied (slightly), I have always emphasized that mtDNA was picked up by early farmers along the way, and clearly, most of that early on. So, the variability of mtDNA first increased tremendously throughout the Balkans, only to decrease (through a lack of novel inputs and founder effects) when entering the less variable NW Central Europe and beyond.

G2a2b is rather rare now in most of Europe (except along the Croatian and Slovenian coast, and subsequent river ways flowing through Austria and Bavaria and Italy into Tyrol that I have pointed out before): so, this incredibly important lineage of farmers did not succeed, in the long run.

As to y-DNA R1a and R1b, it certainly was present for at least 20 - 30 millennia in the steppe, Urals, and general E European regions - so they are one of the first European y-DNA groups after I and perhaps some branches of G.

apostateimpressions said...

Is it possible that "southerns" (E1b/ J1/2) and "northerns" (R1a/b) overran Europe during the late Neolithic/ metal ages? The northerns may have been autosomally similar to I hunters and the southerns similar to G1a farmers, perhaps with the addition of "west asian" admixture. That way the haplos would change but the autosomal would remain similar. We know that G1a was largely replaced in Europe and we have found no R1b before the Bronze Age.

eurologist said...

"BTW, the theory that the current distribution of some YHg-I sub-branches is a result of some "Eastern Germanic tribes caught up in the Slavic expansion" - as someone said in a comment here - seemed to be a bit surreal for me anyway."

Are you talking to me? The major I-subgroups of the Balkans are by all studies less than 2,000 years old. Figure that.

Slumbery said...


I could not remember who said that, so I cannot really say that I was talking to you.

That date maybe compatible with East Germanic origin but does not prove it. The least that should be explained why it is not connected to the known historical distribution of East Germanic people at all and how a supposedly East Germanic Y-Hg became the dominant haplotype of populations that do not have the slightest trace of East Germanic roots.
And then I just accepted the picked time estimation at face value. If that is off just with one millennium (even older estimations can be found), then the picture is immediately very different.

I admit thought: that part of my comment is not actually connected to this article and it was possibly a bit snarky too. Also possibly drags me into a debate I cannot follow trough, because I do not have the time to read articles.

apostateimpressions said...

The Eupedia article on the history of I1 has been updated in the light of the new Neolithic find. It argues that I1 may have dispersed from early LBK Hungary into Germany with LBK and into Scandinavia with Funnlebeakers.

It is fascinating that we are still at a stage of anthropology that a single find can lead to the revision of theories. Clearly the more data made available the better.

On the correspondence of I1 distribution with Germanic migrations, quote:

The Germanic migrations dispersed I1 lineages to Britain (Anglo-Saxons), Belgium (Franks, Saxons), France (Franks, Visigoths and Burgundians), South Germany (Franks, Alamanni, Suebi, Marcomanni, Thuringii and others), Switzerland (Alamanni, Suebi, Burgundians), Iberia (Visigoths, Suebi and Vandals), Italy (Goths, Vandals, Lombards), Austria and Slovenia (Ostrogoths, Lombards, Bavarians), Ukraine and Moldova (Goths), as well as around Hungary and northern Serbia (Gepids). The I1 found among the Poles (6%), Czechs (11%), Slovaks (6%) and Hungarians (8%) is also the result of centuries of influence from their German and Austrian neighbours. The relativelemy high frequency of I1 around Serbia and western Bulgaria (5% to 10%) could be owed to the Goths who settled in the Eastern Roman Empire in the 3rd and 4th centuries.

The Danish and Norwegian Vikings brought more I1 to Britain, Ireland, the Isle of Man, Normandy, Flanders, Iberia, Sicily... The Swedish Vikings (Varangians) set up colonies in Russia and Ukraine, and outposts as far as the Byzantine Empire, the Caucasus and Persia. The higher frequency of I1 in Northwest Russia (east of the Baltic) hints at had a particularly strong Varangian presence, which is concordant with the establishment of the Kievan Rus' by the Swedes.

eurologist said...


No problem.

The way I see it, at the fall of the Roman empire, East Germanic and emerging migrating Slavic tribes (mostly from the ~C/S Ukraine) met and merged. Some of them went to the NW Balkans, others to the East and NE.

I think Poland is a good example of a country in which this newly arriving dominating culture and language seems to make up only about 10% to 15% of y-DNA (based on the percentage of the very recent R1a subgroups)- because people battled and joined and mixed along the way. So, autosomally, Poland today represents N/C Europeans quite well - with a bit of a lack of Germanic contribution (from expulsion), and more of a Baltic contribution in the N.

Also, NW Polish people don't look anything like S Ukrainians (the putative original Slavic source population), because of long migrations into territory that was NC European for more than 4,000 years before: more light skinned, more blond, taller.

NE Balkans have the same features (in fact, they are some of the tallest people in Europe) - which makes it very likely that SE European Slavic speakers admixed with East Germanic peoples during this migration - as also historically attested.

Y-DNA haplogroup I is a signature of C Europe over a wide range of latitudes - separating R1b to the west and R1a to the east. However, while each goes back to more than 30,000 years in the East, I has been able to stick around and represent C Europe for the longest.

The NW Balkans are among the most extreme in Europe in terms of y-DNA, in the sense that a huge fraction of males are descending from just a clan or two ~1,500 ya.

eurologist said...

"NE Balkans"

Sorry - this is supposed to state NW Balkans.

Simon_W said...

@ eurologist

The location of the Slavic Urheimat is a debated topic. According to the Encyclopedia of Indo-European Culture it encompassed the southeasternmost part of the traditional Polish territory and a large deal of the traditional Ukrainian territory, namely almost all of its western part, plus the central part eastwards up to the Dnepr, but notably not the southernmost parts. To the east of them there were Iranian speakers, to the north they had Baltic neighbours. According to Jürgen Udolph, the present-day southeastern Poland and westernmost Ukraine are the most ancient Slavic nucleus. So it's not quite clear why the phenotypes of southern Ukrainians ought to be regarded as the Slavic „signature type“. According to Ilse Schwidetzky, the cranial types that were common to all earliest attested Slavic groups, were a long faced, robust Nordid type and a broad faced, but long headed „Osteuropid“ type. It might be thought that the Nordid type is Germanic legacy, but that would be unfounded, because that type was already present in the far flung Corded Ware. I think it's partly Germanic though, at least in the early western Slavs, where there had been a seizable, sedentary Germanic population before the Slavic expansion.

Afaik Polish people and Ukrainians are quite close autosomally, closer than either of them is to Germans. That was evident in the Dodecad experiments for instance, and they are also close with regards to cranial measurements, as demonstrated by Bunak. So this makes it clear that the Poles cannot be a mere 10 – 15% Slavic, as the y-DNA allegedly suggests. The error lies probably in the idea that only young clades can be a legacy of the Slavic invaders. But actually there is no reason why the early Slavs shouldn't have carried a certain amount of older clades too. (The same applies to the young clades of the W Balkans.)

As for Polish looks, first of all you have to bear in mind that present-day NW Poland was settled by Poles only after 1945, and these came not all from the same area. So the only original NW Poles are those in Pomerelia. And I wouldn't say that these look particularly NC European. Some certainly do, but others look strikingly eastern. There are some common misconceptions about Polish looks, I think. For instance, a long shaped face (high facial index, typical for N-NW Europe) is most typical for southeastern Poland, not for the NW. The distribution of light hair is very patchy, the only clear pattern is the lower incidence of light hair along the (pre-WW II) eastern border (save the NE).

And as I've said before, the tall statures of the W Balkans peak in parts of Bosnia, Montenegro and adjacent parts of Hercegovina and Serbia. At least judging from Coon's map. It may be different today, I don't know, the better living conditions in the NW might favour the body growth there, but this doesn't change the genetic potentials. Now the Germanic impact in the aforementioned regions is quite minimal, the Goths presumably didn't march through the difficult highlands, but more along the rivers. Moreover, the Germanic marker R1b-U106 is less than 1% in these places. And this indicates that the W Balkan high stature phenomenon has developped independently from Germanic input.

Y-DNA haplogroup I2a1 (old name) has a strong presence throughout SE Europe, including the Ukraine. It's much less common now in Germanic territories like Germany and Scandinavia. In consequence it seems hard to believe that a lot of the southeastern presence is due to Germanic influence. It ought to be investigated though, by building a dated tree of the different clades, and by studying their geographic distributions.

Unknown said...

Okay, here’s the questionable logic at the core of this paper.

We see a drastic difference between early neolithic and modern mtDNA in Central Europe.

So,...something big must have happened in Europe in the meantime.

So,... we are going to use MODERN Near Eastern mtDNA to trace “the origin of the first European farmers.”

Because it makes PERFECT sense to assume that there was no big change in the Near East, even though there was a big change in Europe?

In fact, it’s the other way around.

Everything we know says that the Near East would have undergone much more severe genetic alteration than Central Europe.

So, there is no justification for this article “tracing the origins of the first European farmers” with such an analysis.

It is simply misleading.

Here’s the explanation they give:
“Recent aDNA study from 8000 BC Near Eastern farmers raises the question whether modern Near Eastern mtDNA can be used as a proxy for the Near Eastern Neolithic variability [44]. In our opinion, these newly described seven different incomplete HVS-I haplotypes (np 16095-16369) only provide a limited basis for comparative aDNA analyses, and we thus still consider modern-day Near Eastern genetic data sufficient proxies, when tracing the origin of the first European farmers.”

Simon_W said...

Now for the paper:

Certainly stunning that y-DNA I1 was found in the LBKT. So it isn't the legacy of northern hunter-gatherers after all, but it arrived in Scandinavia with Sardinian-like farmers! If it had been necessary to prove that the origin of a haplogroup doesn't always lie in the present-day peak of the haplogroup, there you have it.

Interesting that mt-DNA T1 was found in Starcevo and LBKT. In central Europe it appeared much later, and I had associated it with West Asian admixture. But still, Baalberge, Corded Ware, etc had more of it than Starcevo and LBKT.

Also interesting to see some U2 in LBKT. It's a haplogroup that is associated with rather eastern hunter-gatherers. So their influence extended westwards to western Hungary.

From the PCA of the mt-DNA and the multidimensional scaling plot I would conclude that there were two different waves of early farmers: the Danubian wave and the Cardium wave. The Danubian one gave rise to many central European Neolithic cultures, up to Salzmünde 3400 – 3100 / 3025 BC. The Cardium wave is close to the Neolithic Basque country & Navarre, and it got strongly admixed with southwestern hunter-gatherers by the time it reached Portugal. The late Neolithic to early Bronze Age cultures of central Europe are similar to each other, with the distinction that Bernburg didn't have any U2. The latter appear only admixed with local, central European hunter-gatherers. All the others had admixture from eastern hunter-gatherers.

And this is an important point: Even the German Bell Beakers had U2 according to this study. In fact they have more U2 than the Corded Ware! According to the recent paper on Chalcolithic El Mirador, the Bell Beakers had 0% U2. Apparently they authors didn't take all the evidence into account. Actually I had noted before that there were some incidences of U2 in Bell Beaker people.

This takes me to the puzzle of R1b origins. R1b in Europe is associated with strong ANE admixture, also with some ANI/Gedrosia/MDLP Indo_Iranian admixture, but not with Caucasus/general West_Asian admixture (note the Basques!). Therefore R1b-M269 cannot have reached Europe via Western Asia. And neither via Iberia. Therefore it must be from eastern hunter-gatherers. And the fact that there is some mt-DNA U2 along with some y-DNA R1b in Bell Beaker people from Germany is well in line with this. It would be interesting to know how and when R1b went from hunter-gatherers to farmers. Dienekes' Rolloff experiment with the French (an R1b dominated population) as a mixture of Sardinians and Burusho suggests that this happened in the early 5th millennium BC. Indeed there is cranial evidence for Cromagnoid hunter-gatherer admixture in the Tisza-Polgar culture of the early 5th millennium.

I had speculated before that R1b might have reached Portugal with an early Neolithic wave that was distinct from the Cardium wave. However, now it's clear that the peculiarity of the Neolithic Portuguese mt-DNA is merely the result of heavy admixture from local hunter-gatherers, and the PCA shows that the NPO is nonetheless rooted in the Cardium wave. Autosomally the Portuguese aren't hunter-gatherer-like but predominantly Mediterranean, i.e. farmer like. This must have been introduced mostly by males then. In theory, R1b might have been picked up from local hunter-gatherers, too, but its strong eastern autosomal associations make this very unlikely. So in all likelihood it didn't expand to central Europe with Bell Beakers either. And, fittingly, the German Bell Beakers don't cluster with Neolithic Portugal in this paper.

Finally, note how the mt-DNA distance maps show quite some Starcevo/LBKT ancestry in Italy and Switzerland, but hardly in southwestern Europe west of it. That's because of the Cardium ancestry there. Only northern Portugal and Galicia show some Starcevo/LBKT mt-DNA, probably from central European invaders.

Unknown said...

TO: Simon_W:
"first of all you have to bear in mind that present-day NW Poland was settled by Poles only after 1945"
But the traditional border was the Oder River. That only changed in the meantime. More importantly Western Slavic speakers -- the Wends or Polabian groups -- were all the way to the Elbe in Charlemagne's time. And some to this day some of these folk still reside in the Sorbian areas in eastern Germany. So, that border would work both ways.

"So this makes it clear that the Poles cannot be a mere 10 – 15% Slavic, as the y-DNA allegedly suggests."

And what is this "Slavic" y-DNA and where does one find it. If you are going to connect y-DNA to a language, then I'd definitely call whatever y-DNA the Poles are "Slavic."

andrew said...

It is worth noting that G2a has a very specific distribution today. It is predominant in the NW Caucasus Mountains among speakers of the NW Caucasian (e.g. Circassian, Adygo, and Abkazian and Shapsugs) languages (G2a1) and the Iranian language Ossets (G2a3).

In contrast, NE Caucasians such as Chechens are mostly Y-DNA J2 (also predominant in Turkey suggesting connection pre-Hittite peoples of Anatolia), and Dagestanis are mostly Y-DNA J1, which has strong Near Eastern affiliations.

Thus, NW Caucasus may be the relict population of first wave Neolithic peoples rather than the source of the first wave Neolithic, however, while Chechens and Dagestanis may be relict populations associated with later metal age farmer waves.

Simon_W said...

Unknown, no, the traditional border wasn't the Oder river, in fact the present-day Oder-Neisse border is quite arbitrary. Calling this the „traditional border“ would be a projection of the current state into the past, as if it was bound to happen one day that the Oder became the border again. In reality it was a back-and-forth: In antiquity there were the Burgundians, Vandals, Goths and other eastern Germanic tribes in what is now Poland. But back then neither Germans nor Poles did yet exist. Then the eastern Germanics migrated off and Slavic tribes expanded far westwards, even behind the Oder, as you said it, into what is now eastern Germany. At that time there were no Poles either, there were just different Slavic tribes, one of which were the Polanes in what is now Wielkopolska (Greater Poland). And the Germans were as well just starting to grow together from different Germanic tribes in the eastern Frankish empire. Then the Polish state got founded, and its western border was per chance similar (but not identical) to the present-day border. But soon afterwards Silesia and Pomerania became part of the German (Holy Roman) empire and got settled with Germans from the west, in the 13th and 14th century. The Slavs in eastern Germany, Silesia and Pomerania became assimilated and then merged with the Germans. In some areas where the German settlement was less dense the opposite happened, and the Germans got Polonized. The ethnic (not the political) borders remained more or less the same from the 13th-14th century till 1945.

The Sorbs are merely the last relic of the medieval Slavic population of what is now eastern Germany, there used to be other long-lasting relics. The Slavs surely didn't disappear without a trace, they were assimilated and merged with the Germans. Their legacy is in the genes and, more obviously, in plenty of place names.

However, my point to Eurologist was merely that it's illusory to think the NW Poles are more NW-like than other Poles, first of all because they haven't been NW for that long.

By „Slavic y-DNA“ we had the legacy of the expanding Proto-Slavs in mind. There once must have been a Proto-Slavic population and the lineages they founded, that's what we were talking about. But of course, during their expansion, pre-Slavic people became Slavs and their lineages further participated in the Slavic expansion and in the end all present-day Slavs are Slavic. But in theory it would be possible to trace some of their lineages back to other ethnic groups of the past.

In my opinion Udolph was right, and this is what happened:

So, similarities between Poles, Czechs, Slovenians and Croats may easily be the result of common Slavic ancestry. Alternatively, if the similarity includes western central Europeans, it may be from the pre-Slavic and pre-Germanic population layer. The Lusatian culture in Poland was probably neither Slavic nor Germanic, it was related with the Urnfield complex of cultures. And the Tumulus and Urnfield cultures were also present in Slovenia and northern Croatia.

Unknown said...

This paper just doesn’t help things. When new aDNA shows up, researchers should really be careful about the conclusions they draw. Especially if they come in with a pre-disposed conclusion.

As mentioned before they justify using current DNA data for the Near East and pretty much reject ancient DNA from Fernandez (2014) for a bunch of bad reasons.

But after saying this, they somehow come to the conclusion that G2a Y-DNA is “Near Eastern.”

“Modern” G2a is probabilistically NOT NearEastern. Not in the sense they are using the word. Not in terms of modern diversity or frequency or likely history.

Omitting this from the research is pretty blatant.
SEE ESPECIALLY Siiri Rootsi, Distinguishing the co-ancestries of haplogroup G Y-chromosomes in the populations of Europe and the Caucasus (2012)
The most reasonable conclusion is that the so-called signature Y-DNA of the neolithic should be both in origin and in modern terms be located in the Caucasus, northeastern Anatolia and Armenia.

The current archaeological evidence is that livestock domestication in fact traveled very slowly through southwestern Anatolia from an origin in the Zagros mountains, near Lake Vann and much closer to where G2a most likely originated. SEE B. Arbuckle, Data Sharing Reveals Complexity in the Westward Spread of Domestic Animals across Neolithic Turkey (2014).

So, the authors miss this problem on the most basic level. The ancient Y-DNA evidence does NOT point to the Near East. It points to a location at the eastern end of the Black Sea, in vicinity of the Caucasus, eastern Anatolia, Armenia and western Iran.

Unknown said...

Sorry about the additional posting. There are other problems with this article that are hopefully worth mentioning.

#1 - the article says that there is no H mtDNA found in LBK. The chart Dienekes posted below from the El Mirador article definitely says something different. I’m wondering how that could be.

In fact, the chart in this article omits some important evidence included in the El Mirador chart, including Funnel Beaker, Pitted Ware, Hunter-Gatherer mtDNA from the south, central and east, plus Bronze Age Siberia and Kazakhstan.

There’s no doubt that there appears to be a demographic change in mtDNA in Europe, but it does not look like much of it came from the east.

The Siberian, Kazakhstan, HG east and Pitted Ware are pretty far off -- They would be off the chart in a multivariate pictogram.

The authors have been too selective in how they compare the changes.

Gómez-Sánchez, Mitochondrial DNA from El Mirador Cave

To the chart here

#2 - The dating of Starcevo is controversial. If it actually only dates from around 5500 BCE, then there are plenty of other “cultures” in the Balkans and near the Danube that might be as reflective of the “Neolithic” gene pool -- particularly in contrast to groups like Unetice. Sites like Durankulak -- which, as happens too often, was misdated as a “kurgan” settlement, about 2000 years too early, when it was first discovered -- and Vinca point to very different economies and perhaps genetic groups.

Use of the term “Late Neolithic” is also confusing, given the metallurgical economies evident throughout southern Europe before 3000 BC. We have gold reported at Varna around 4500 BCE and bronze reported for Vinca
about at the same time. This kind of technology is not found in “TransDanubian” LBK. Cultures as important as Tripolye and Vinca have not yielded DNA as far as I know. This is a major gap.

LBK was a very specialized technology group. It did seem to keep the Mediterranean approach and a clear contrast to other cultures in the Balkans, to TRB and south western Europe. That makes this article’s title seem inappropriate.

#3 - There has been at least one report of G Y-DNA at a Corded Ware site (along with another non-R Y-DNA individual). It was not mentioned in the article.

#4 - The changing diversity of the mtDNA throughout the early Neolithic may show it was “patrilocal.” But if the males were finding brides from the outside, where were those brides from? An exchange between adjacent LBK settlements certainly should not have yielded a different diversity, if we can go by these ratios of haplogroups the authors are using. Once again, the El Mirador chart shows quite a different result for every “culture.” So we are either talking about long distance, mail-order brides. Or the kind of extreme random drift that always happens in simulations of small populations.

Simon_W said...

But Unknown, why then do the early European farmers lack any sign of the highland West Asian autosomal component? I think this pretty much rules out an origin of the earliest European farmers in the vicinity of the Caucasus, eastern Anatolia and Armenia. Judging from the Eupedia map of haplogroup G, that haplogroup is still quite common in central Anatolia and even more common in parts of western Anatolia. Yes, it's much more common in parts of the Caucasus, but that may be because it's a refuge area. G has been similarly common in Neolithic Europe, but has been much replaced since then.

Then, does the article really say that there is no H in LBK? That would be strange, because according to table S7 (in the supplementary data), there is 14.81% H in LBK, plus 1.85% H5; the reduced samples (with asterisk) have even slightly higher percentages. I very much suppose that this has been taken into account in the analyses.

Funnel Beaker, Pitted Ware, Bronze Age Siberia and Kazakhstan are indeed not included. The sample HGSW is from Spain and Portugal. And the sample HGCN (hunter-gatherers of central and north Europe) is from Germany, Luxemburg, Poland and Lithuania, so you can't say central European hunter-gatherers are omitted, and Lithuania is even quite eastern.

You say the authors have been too selective, but I think even the charts they presented don't support the idea that the demographic changes in the later Neolithic are just a consequence of eastern immigration. Because the Bernburg culture clusters with the other late Neolithic – early Bronze Age cultures, yet it has no U2. The difference between the earlier central European Neolithic cultures and Bernburg is most of all the much higher incidence of hunter-gatherer haplogroups in Bernburg. And it would make sense if this was parallelled on the autosomal side with a higher North European component. So I would be very hesitant to ascribe this all to eastern invaders.

Yet it has to be said also that Unetice does deviate a lot in the eastern direction, also in the El Mirador chart. Unetice was an important central European culture of the early Bronze Age, and it was considerably more eastern than the Final Neolithic Corded Ware. Since you seem to hold the El Mirador study in higher esteem than this study, please notice that the El Mirador study omitted data as well: As I said, according to it, the Bell Beaker people from Germany had 0% U2. According to the present study however, there were some Bell Beaker individuals with U2, and this can also be seen here:

Where did you get that the dating of Starcevo is controversial, from Wikipedia? Starcevo can't date from 5500 BC, because at that date, Vinca already followed. Vinca had a lot of geographical overlap with Starcevo. Furthermore, Starcevo formed a cultural complex with Körös and Cris, they have to be contemporary. The source for the lower age of Starcevo indicated in Wikipedia is just a historical atlas, which perhaps used uncalibrated C14 dates.

The use of the term „Late Neolithic“ isn't confusing, if one bears in mind that (as always) this is not an absolute category but regionally different. E.g., in central Europe the late Neolithic started about 3500 BC, whereas in northern Europe it started after 2400 BC. But surely, the use of gold is by no means at odds with a classification of a culture as Neolithic. Not even the use of copper is at odds with it. Because the Neolithic lasted until the beginning of the Bronze Age. That's why even cultures like the Corded Ware and Bell Beaker are actually nothing but Neolithic. The term Chalcolithic („Copper-/Stone Age“) roughly coincides with the Upper to Final Neolithic, at least in central Europe. In terminological questions like these, you'll always find differing opinions, however.

Simon_W said...


Tripolye will probably be very similar genetically to the other early Neolithic farmers. At least that's what the craniometrical evidence suggests.

You compared the LBK to contemporary cultures of the Balkans and to the TRB. I don't quite see the problem however. The TRB started more than 1000 years later than the LBK. And although the LBK was an outgrowth of earlier Balkanic cultures, it can't be expected to be on the same technological level as contemporary Balkanic culture. It simply wasn't on the latest level, it lagged behind, the innovations needed time until they reached central Europe.

I think the graph with the mtDNA change through time isn't that instructive. It's mostly a consequence of the mtDNA mutating.

Simon_W said...

By the way, to prevent any misunderstandings: Even though the current Polish-German border cannot claim to have a lot of tradition, it's nonetheless the border now and should be respected. Nothing could justify another ethnic cleansing. And anyway, Germany and Poland are friends now and members of the EU, which is like a big family. I'm against the exploitation of history for political means. History is about finding the truth, and when it gets exploited, the truth always suffers.

Unknown said...

Reply to Simon_W
1.The article, as already quoted above, says “...Previous studies have shown that haplogroups N1a, T2, J, K, HV, V, W and X are most characteristic for the Central European LBK.”

How did they lose H? Well, mtDNA H is very inconvenient if you are trying to show that the early Neolithic disappeared from modern Europe.

We’ve seen this before. In Brandt, Haak 2013, which Dienekes posted.
Take a close look at how they graphed mtDNA H. Completely separate from all the Neolithic cultures where it occurs.

The justification seems to be that H appeared in mesolithic sites. As if that was an excuse for making it invisible. Especially if you say you are comparing LBK to modern populations, where H dominates.

If the El Mirador charts and tables are correct, then H, K and T2 all outnumber what they call the “signature” mtDNA -- N1a -- in LBK.

In fact, if the El Mirador chart is correct, N1a’s presence dwindles very soon after LBK splits (see columns for Rossen, Schoningen and Bernburg).
So there is no surprise from the central European Neolithic that N1a is still not around.

Meantime, H -- one of the major haplogroups in LBK -- only keeps getting a bigger share in the middle and late Neolithic, in Corded Ware AND even appears to be the largest mtDNA group in Bronze Age Unetice.

So, by these standards, mtDNA found throughout the European Neolithic not only survived but prospered.

On the other hand, it looks like the “signature” mtDNA of the Bronze Age Western Steppes is C4. How does that work with modern European populations?

The dominance of C4 among the kurgans probably reflects the Bronze Age invasion of Europe by Papuan warlords. :)
2.For much later dates for Starcevo, see the radiocarbon reports collected in Biagi 2005 at the URL below. Almost all the BP dates you’ll see there are from around 5000-4500 BC. More recent C-14 have yielded quite earlier dates, but the material culture described is quite different from LBK, right down to the underground houses and lack of cows.

Be aware that was a quite different Impressed Ware culture dating from about 6000 BC along the Adriatic Balkan coast that was producing finely fired pottery, quite beyond what Starcevo was doing, that may have been very influential on Vinca’s metallurgical efforts.

Read about them here:

Unknown said...

Continued Reply to Simon_W

3.Simon -- when you mention “highland West Asian autosomal component” that has not been found in the European neolithic, I’m not sure what that is. There is Lazaridis et al, “Ancient human genomes suggest three ancestral populations for present-day Europeans” that Dienekes posted and has just been published in Nature. It identifies an “Ancient North Eurasian” element not found in early European farmers or hunter-gatherers, but is now found in Europe and the Near East. There’s been much press about the article, where
David Reich and the others are quoted as calling it “major surprise.”

If I understand it right, they used MA-1 as the “Ancient North Eurasian” point. If that’s correct, then the surprise should be why the degree of “Ancient North Eurasian” ancestry isn’t higher. After all, MA-1’s Y-DNA should be in some way ancestral to the R1a and R1b, major modern European haplogroups. I suspect both groups escalated at different times out of the areas where domestication became a core process, the triangle created by the Zagros mountains, the Caucasus and western Iran. That would be the source of your “Ancient North Eurasian” ancestry. G2a originally came from roughly the same place, but probably a bit earlier.

Unknown said...

A correction about the original C-14 dating of Starcevo I gave in an earlier post. Biagi did in fact report uncalibrated BP dates for Starcevo sites no earlier than 5750 BC. However it's been pointed out to me thathe also gave Cris culture cal dates going back to around 6000 BC, with one going back to about 6500 BC, and that they are arguably interchangeable cultural designations. In fact, the article at the URL posted below gives dates starting at @6100 BC and finding type A period dates for as late as 5000 BC. It also concludes that Starcevo folk mastered "the skill of building above-ground structures, which were presumably used only for special purposes."

Simon_W said...

@ Unknown

I'm not convinced that Brandt, Haak et al. treated H separately because of a hidden agenda. To me it rather looks like they were uncertain if H originally came from the hunter-gatherers only or if the earliest farmers also had carried some H along. Because, in the map they added „H?“ to event A, the arrival of the earliest farmers in Europe. Since it was uncertain, it would be confusing to list it among the early farmer haplogroups, when the basic distinction is between hunter-gatherer haplogroups and early farmer haplogroups. H was simply both. And they didn't make it invisible in the timeline. It's plotted separately from early farmer haplogroups, but one can still see in the same plot how common it (together with U3 and others) was in the various Neolithic cultures.
I neither think there is a clear trend for an increase of H. It was already very common in Rössen, at an early date, it was again similarly common in Salzmünde, it was most common in Bell Beaker, and it was less common in Corded Ware and Unetice, in spite of still being the most common haplogroup there.

The occurence of C4 in the Mesolithic Ukraine, and again in the Chalcolithic Yamnaya culture and subsequent cultures, is of course interesting. But you have to see the relations: Much more mt-DNA data has been reported for Yamnaya (27 samples afaik), and the most common haplogroup was H (7 times), followed by U5 and T1 (3 each). C4 has been one in Yamnaya proper, another one in its Catacomb (late) phase, and one in the still later Mnogovalikovaya (KMK), or Babino culture.

As for the West Asian autosomal component: Dienekes has blogged a lot about this a couple of years ago. He discovered it in the Dodecad Ancestry Project where he used ADMIXTURE on autosomal samples provided by project participants and complemented it with samples that are publicly available. It is absent in early European farmers and at least in Europe it's correlated with Indo-European languages, which led him to hypothesize that it's a signal of the Indo-European expansion. A similar component has been found in the MDLP project. However, last Monday it became clear to me that this West Asian component is a compound of older population layers + a smaller, more specific component that variously peaks in Balochistan, southern central Asia or in the Kalash, see the comment section of the blog post on the ASHG 2014 titles and abstracts. So therefore I have to renounce my objection to you, as the higlands of eastern Anatolia and Armenia need not have been autosomally West Asian at the time of the earliest farmers.

The paper by Lazaridis et al. isn't new to regular readers of this blog, a preprint was already made available last December, you can find it here:
Unfortunately its huge importance kind of submerged in the epic battle between German Dziebel and the others on quite a peripheral topic, maybe also in the high days and holidays. Others like „Genetiker“ found the paper stupid, dumb, useless, not at all saying something new etc. While it only and rightly deserves to be called a fascinating, important peace of research.

However, be aware that the ANE component isn't identical with the above mentioned West Asian component. As the paper showed, MA-1 is related with various modern components, only one of them being the Kalash centered component. The PCAs show that there are two different peaks of ANE admixture in Western Eurasians: One is centered in Caucasus populations, particularly in Lezgins (who have a lot of R1b), the other peak is more northern, in Russians and Mordovians (known to have lots of R1a).

Slumbery said...


"So therefore I have to renounce my objection to you, as the higlands of eastern Anatolia and Armenia need not have been autosomally West Asian at the time of the earliest farmers."

Yes, this was one of conclusions here years ago. There are also other signs of significant changes in that region. Plus thinking a bit "reversely": it is hard to imagine how could be the first European farmers completely without it if it's distribution in the Levant/Middle East was even remotely similar to the modern distribution.

Simon_W said...

I'm not so sure if that was clear here years ago. My impression of Dienekes' theory had been that the West Asian component had expanded from the West Asian highlands east of central Anatolia and north of the Levant, around the place where Gamkrelidse and Ivanov suspected the PIE homeland, i.e. eastern Anatolia, Armenia, the Lake Van region. That would still allow for a different EEF population, if that population originated not exactly there.

Unknown said...

Sorry for the late reply.
Simon_W wrote:

“I'm not convinced that Brandt, Haak et al. treated H separately because of a hidden agenda. To me it rather looks like they were uncertain if H originally came from the hunter-gatherers only or if the earliest farmers also had carried some H along.”

No, that can’t be right. Unless a haplogroup arose during the Neolithic, all these haplogroups had to be “originally” hunter-gatherers. Everybody was mesolithic at some point.

Where H came from hardly matters, IF the question is what parental DNA lineage has been found consistently in the Neolithic.

The problem once again is the effect of the kurgan-steppes theory. Any strong continuity with modern populations is going to be avoided so some massive change can be attributed to the steppes. It’s warped the archaeology in the Balkans and Anatolia -- I can give glaring examples.

And of course it’s warping the obvious implications of H as a gene that is clearly representative of the Neolithic. That chart from Brandt is a steaming example.

Matching proportions of mtDNA found in a culture is a little silly. No one seriously thinks that the exact same mix of mtdna in individuals would move from the old village to a new village exactly the same every time one was founded.

What is consistent is that H appears in the reports in nearly every single report across Europe across the Neolithic and up to the Bronze Age. And it is nearly always among the most common genes.

I’m not sure how the subclades work, but the point that Brotherton, Haak made in their 2013 paper on mtdna H is that it is the signature gene of the European Neolithic and Copper Age cultures that have supplied aDNA so far.

And doesn’t look at all like it came from the steppes. It does look like it ended up in the steppes, just like food production did. It may have been the H female who converted at least the European hunter-gatherer to start domesticating. And that makes a lot of sense.

Simon_W said...

Well, what would be the alternative to the kind of plot that Brandt, Haak et al. made? The only feasible alternative would be a graph like in the El Mirador paper, with the percentage shares of the haplogroups for each included culture. Imho that wasn't a bad solution either, it was instructive, too. The downside being that with that many colours it isn't immediately very evident, it has to be carefully studied. Therefore a solution where the haplogroups are integrated into four groups has it's positive aspects too.

Though it may not seem correct to divide haplogroups in hunter-gatherer haplogroups and farmer haplogroups, there is a clear criterion to distinguish between them: In what context they first appear in central Germany. And as long as you bear in mind that H is also part of the early/middle Neolithic farming cultures, I see no problem. Actually all four plotted curves up to the late Neolithic are 100% part of the early/middle Neolithic farmers, because, apart from the single HGC sample at the start, they are plotted above Neolithic farming cultures.

Since the fourth category, „other“, mostly consists of H, the plot does show that H was consistently an important part of the Neolithic cultures. You just have to know that U3 and others don't matter much to see that. Alright, maybe it would have been fairer to make H a single fourth category and then add a minor fifth one with U3 and the rest.

As regards your critique of the Kurgan theory: As Lazaridis et al. 2014 clearly showed, and as has become even clearer with the paper by Cristina Gamba et a., there must have been a massive impact of eastern hunter-gatherers on the modern European gene pool. What exactly happened still has to be worked out, but one thing is clear: Those who had been categorically denying such an impact and dramatic changes in the final Neolithic were proven wrong. Yes, there was continuity, but also strong change. You can't have modern Europeans without both.

Unknown said...

Simon_W wrote:
“Alright, maybe it would have been fairer to make H a single fourth category and then add a minor fifth one with U3 and the rest.”

Probably not. mtdna H has not been found in mesolithic Central Europe, the Balkans, Scandinavia or as far as I know the mesolithic “Steppes.” (See Lillie, Potekhina, et al (2012) Wildea, Timpson, et al (2012))

Brandt notes a part of that absence in his latest paper. (“haplogroup H... has not been observed in Mesolithic Central Europe and Scandinavia”)

If H was mesolithic in Iberia, it started being nothing but Neolithic right after that. And there’s a pretty fair chance that H wasn’t in there in the mesolithic in the first place.

And that would mean the major mtDNA in Europe today that spans the continent came to Europe with farming.

There is tremendous pressure in many academic circles to defend the Kurgan idea come hell or high water and that has skewed the research and prompted misleading charts like Brandt’s.

And that why the weasel about H as neolithic.

BellBeakerBlogger also has some nice thoughts about the questionable archaeology that made H some how separate from “Neolithic”

Simon_W wrote:
“As regards your critique of the Kurgan theory: As Lazaridis et al. 2014 clearly showed, and as has become even clearer with the paper by Cristina Gamba et a., there must have been a massive impact of eastern hunter-gatherers on the modern European gene pool.”

Let’s get this clear. Lazardis says no more than that the so-called ANE admixture reaches about 14% and averages 10% in modern Europeans. (TABLE 14.11) The same paper also says that it reaches 13% and 11% in modern Cypriots and Druze.

That is NOT massive impact. ~10% can just as easily be attributed to wide spread migrant workers who moved to Europe from the Steppes over the millennia for the sake of simply finding work as agricultural workers.

80%+ of the modern European population is attributable to Neolithic farmers and people who were there before the farmers came.

The Lazardis paper names the Northeastern Caucasus as the main center of modern ANE genes >>> where NO IE languages are local. They are either Turkic or Northern Caucasian languages.

The most surprising thing about the Gamba is that a fundamentally Near Eastern Y-haplotype was on the Danube in the Bronze Age. Which suggests that even the ANE element in the European population may have also come along the same route earlier agri-pasturalist used. And that was not the Steppes.

Simon_W said...

Table 14.11? Do you mean Table S14.11 in the latest version of the paper? This shows the ratio of EEF ancestry. Table S14.10 shows ANE ancestry, in the third column. I didn't calculate the average of alle European populations, but it looks obvious that this must be more than 10%, since almost all European populations have more than 10% ANE according to Table S14.10. The strongest ANE admixture is found in Estonians (18.7%), Hungarians (17.4%) and the Scottish (17.1%).

But yes, according to Table S14.15 Cypriots have 13.3% and Druze 11.8% ANE.

You say that these are not massive levels of admixture. But you have to take into account that the people who carried ANE into Europe in all likelihood didn't have 100% ANE! So far we don't know anything certain. But we will see soon how much ANE the Yamnaya people had, when the paper comes out. If they had 50% ANE, which is a generous guess, we'd need an admixture of 28% to reach a level of 14%. According to a presentation of the paper at a conference, Yamnaya people can be modelled as 50% Karelian hunter-gatherers + 50% modern Armenians. So 50% ANE is going to be rather too generous.

And no, it doesn't look like a slow change that took millennia to accomplish. The population in the Carpathian Basin changed from the Sardinian- + Basque-like CO1 with no ANE to the East German-like BR1 with about 12% ANE within 800 years at most.

And no, the latest version of the Lazaridis et al. paper explicitly refutes the idea that the ANE admixture in Europe originated in Western Asia or the Caucasus, on p.126 of the supplemental info:
„The finding of high ANE ancestry in the North Caucasus might suggest that the Caucasus is a potential source of this type of ancestry in Europe. However, when we try to fit present-day Europeans as a 3-way mixture of a North Caucasian population+EEF+WHG in the structure of Fig. S14.20 this model is successful for only 5 populations (Bergamo, Tuscan, Italian_South, Bulgarian, Spanish_North, using Lezgins as a sister group to the admixing population). Admixture from the Caucasus would need to be substantial to account for observed ANE levels in Europe (e.g., for a European population with ~15% ANE ancestry, almost half of its ancestry must come from a Lezgin-like population with ~29% ANE ancestry; this would account for the ANE ancestry but would greatly dilute its WHG-related ancestry, and yet present-day Europeans have increased affinity to WHG in Extended Data Fig. 4 relative to Stuttgart).“

Unknown said...

Yes, sorry, Table 14.10. And, yes, the paper says ANE admixture reaches 17.1% mean in Scotland and the average across the continent probably comes to near 14% across the continent. A case of confirmation biased memory on my part.

While on the subject of confirmation bias, I must apologize, but the results of paper don’t support Kurgan-Indo-European idea in any way as much as you or some of the authors had hoped. The paper is obviously trying to support some variety of the Kurgan theory.

It doesn’t work.

More than 80% of modern European genetic descent is NOT from ANE. Across the board. That’s the real news.

Logically, no matter what additional mixtures these ANE types carried, they DID NOT change that overall genetic relatedness of more than 80% of modern Europeans to those proxies of “Western Hunter-Gatherers” and “Early European Farmers.”

This would appear, prima facie, to pretty much eliminate the idea of population replacement. More than 80% of modern European descent can be attributed to the time of or before the early Neolithic.

It does leave open the possibility of a small number of “Indo-European masters” dragging their horses and kurgans from the steppes and forcing the other 80% to take Indo-European language lessons. But...

As you have pointed out, the largest ANE shares are found in Estonia and Hungary.
These countries are not Indo-European language speakers. The status of the Pictish spoken in early Scotland is unclear, but it might also arguably be non-IE.

Based on this concentration in non-IE speaking communities, we should probably first conclude that ANE carriers were not IE speakers. This is further confirmed by the two much more heavily ANE populations given in the paper -- Lezgin (28%) and Georgians (27%) -- are also NOT IE speakers.

It’s ridiculous of course to use the modern population of the Caucasus to prove that ANE could not have come from the Caucasus. The paper notes it has no Near East DNA to work with. Loads of Steppes DNA, but no Near East DNA. How did that happen?

As I wrote before, if MA-1 was basal R Y-dna, then we probably can trace later R types moving to Iran, Armenia, Anatolia and the southern Caucasus, where the evidence is that R1a and R1b originated. This is probably where Indo-European also originated.

They should both have been carrying some ANE genes with them.
So we can project that R1b (among others) moved along Southern Europe, North Africa and the Mediterranean and to Iberia speaking Centum IE. Centum IE spread from Iberia across Europe with Bell Beaker. While R1a moved from Iran onto the Steppes, etc., speaking satem IE, Indo-Iranian and Slavic.

Cristina Gamba’s paper shows the Bronze Age DNA specimens nestled nicely among the Spanish and French, with Russians as far away as Sardinians. So we probably should say these individuals came from the west with Bell Beaker. The very distant Iron Age specimen lines up with modern Armenians and so we can probably look again to Anatolia.

Simon_W said...

You have misunderstood my point. I wasn't arguing in favour of the Kurgan theory of IE origins at all, I merely pointed out that steppe populations had an important, non-negligible impact. The question where the PIE originated is still controversial and I'm not quite sure either. But even if you are convinced of a West Asian origin, this shouldn't lead you to deny the facts, namely that EHG admixture was considerable. Actually these issues might be discussed without making reference to language at all.

The authors didn't say anything about Indo-Europeans either. We will see what they'll have to say on that topic in their upcoming paper. Presumably they're discussing their findings right now and are uncertain themselves. I've recently read that Patterson seems to support a PIE origin in West Asia, for the reason that Armenians don't seem to have northern HG admixture, and supposedly Indians neither. And then, as I said, Yamnaya seems to be modelled as a 50-50 mix of Karelian HG and modern Armenians in their upcoming paper. Strange you didn't pick this up. This leaves ample room for argument about which side of the mix introduced IE to Yamnaya.

Your calculations about the share of pure ANE vs. EEF + WHG in Europe are pointless. I think nobody would equal pure ANE with Proto-Indo-Europeans. Pure ANE, that's Mal'ta boy on Lake Baikal, 22'000 BC. That's very far, both in time and place, from seriously discussed models of IE origins. In the model by Lazaridis et al. what finally admixed with EEFs was a „hunter“ population inbetween WHG and ANE, and that's much more realistic, even if it's a simplification.

Nah, Georgians never have 27% ANE, they have much less. Chechens have 27%, and like the Lezgins they belong to the Northeast Caucasian language family. It's in Northeast Caucasians that we find the highest concentrations of ANE in the Caucasus. As I said, I wouldn't equate ANE with IE, that would be silly.

The ANE in the Caucasus has probably arrived both from the north and from the south, but probably much more from the south, since R1b came from the south and R1b is what's strong in Dagestan.

I'm not so sure that R1a came from the south too, there are alternative explanations for the high SNP diversity there. Actually that region in southeastern Anatolia and Kurdistan, where R1a diversity is high, also has elevated levels of I2a, suggesting that there was some Balkan influence there which probably has brought some European R1a along.

The previews I've seen on the upcoming paper stated that Karelian HGs and HGs from the Samara valley had plenty of ANE, and even the Scandinavian HGs had some ANE, so it's beyond doubt that ANE wasn't spread westwards exclusively along a southern route.

The southern stream in any case didn't have WHG admixture, so you'd have to assume a population with very high ANE levels for your model to work. I don't know if that's very plausible. In the end probably both, the southern and the northern stream brought ANE to western Europe, the question remains who brought more.

Personally I'm thinking that R1b and Bell Beaker were not associated with IE languages, rather with Iberian and Basque.

Simon_W said...

(contd.) And instead of relying on the PCA in Gamba's paper, which just shows the first two dimensions and probably suffers from projection bias, I prefer analyses with admixture calculators, like the MDLP K23b on GEDmatch. BR2 is closest to southern Germans, according to this calculator. Unfortunately BR1 isn't on GEDmatch, but David from Eurogenes, who has good calculators as well, determined that she is closest to eastern Germans. Now you could say that doesn't change a lot, so they came from Germany. But this would be faulty reasoning. First of all, there is not a shred of evidence that western or central Europeans of the Copper Age were like modern western or central Europeans – that would be absurd, when Copper Age Europeans of Hungary were like CO1, who was like a typical EEF. CO1 is most similar to Sardinians with a bit of Basque thrown in (because of slightly higher WHG admixture). In admixture calculators, also ones by Dienekes, these EEF people had very low North European components and very high Mediterranean components. No wonder, since they were so close to Sardinians! Check this clip by David showing the K15 admixture components of the Hungarian samples:

You can see, the farmers, starting with NE1 had a very strong West Med component, also a considerable Atlantic component, but almost no North Sea and Baltic, and zero Eastern Euro. Besides this, NE1 had also a strong East Med component and a bit of Red Sea. NE7 was similar, but had even more West Med at the expense of East Med and Red Sea – likely the result of some WHG admixture, making it more similar to modern western Mediterraneans. Accordingly, the North Sea component was slightly stronger, but still weak and the Baltic and Eastern Euro nonexistent. CO1 was very similar, she had again a bit more East Med and less North Sea. And now look the tremendous change between CO1 and BR1. The Med components were seriously pushed down and North Sea and Baltic shot up amazingly. There also appears some slight Eastern Euro for the first time after KO1. BR2 is similar, but has more Eastern Euro and at the same time some East Med and Red Sea again, in agreement with his J2 yDNA. IR1 has even more Eastern Euro, he already has a lot of that, and at the same time he has a lot of the West Asian component which we see here for the first time. And now explain the change from CO1 to BR1 with immigrants from southwestern Europe, it makes no sense at all.

Actually you have to ask what population CO1 must have admixed with to produce BR1 since there was hardly a complete population wipe-out. Moreover BR1 belongs to the Mako culture which started at 2800 BC, at that time there were no Bell Beakers in central Europe. And anyway the actual Bell Beaker presence in Hungary was very modest. What we see much more are syncretistic cultures of the Bronze Age picking up some Bell Beaker influence.

IR1, according to MDLP K23b analysis is like a mix of Basques + Norwegians + Chechens + Udmurts. This means, about half of his ancestry was rather local, the Bronze-Age like mix of EEFs and more north European-like people, the other half was like a mix of Caucasus populations with more northern eastern Europeans. It has been suggested that his culture his culture can be identified with the ancient Cimmerians, the Thraco-Cimmerian horizon, and these people may have originated near the Caucasus in the Kuban steppe. Anatolia doesn't work out however, his Eastern Euro component is much too high for that. Moreover he had y-haplogroup N which isn't typical for Anatolia.