March 16, 2018

Ancient DNA from the Balkans

This study has been for several months around but I have not discussed until now and is well worth a mention.

Ian Mathieson et al., The Genomic History Of Southeastern Europe. BioRXiv (pre-pub) 2017. doi:10.1101/135616

There is a lot of ancient autosomal DNA from the region but it basically says one thing: everything was almost exactly as expected from archaeology. The Karanovo-Gumelnita people, famed for inventing the Bronze Age a whole millenium earlier than anyone else, and nearby related cultures, were within the mainstream (Vasconic) Neolithic genetic grouping. This changed however with the kurgan invasion expressed primarily in the Ezero culture, which I've been told should be considered direct precursors of Thracians. But the change is not something radical: more genetic affinity with the steppe is visible than before and more generalized through all samples. 

Fig. 1-D - Supervised ADMIXTURE plot, modeling each ancient individual (one per row), as a mixture of
populations represented by clusters containing Anatolian Neolithic (grey), Yamnaya from Samara (yellow), EHG (pink) and WHG (green).
  (click to expand)

Well, it is not exactly everything what is as expected, if we consider Polish Globular Amphorae culture, which I would have expected to be at least somewhat steppary, Indoeuropean, already but were not. Thus it seems I have to concede on this culture and its precursors being still part of the Vasconic Neolithic. This makes Indoeuropean penetration into Central-North Europe a much more sudden episode and one directly tied to Corded Ware culture and nothing else. It must have been perceived by its victims like a massive catastrophe, because it was a huge area which they conquered and to a large extent colonized in a very short span of time.

Lots of R1b in Epipaleolithic Balkans

Most interesting anyhow is the huge hoard of ancient Y-DNA R1b in the Iron Gates region (Lepenski Vir) before the Neolithic. This not only demonstrates, again, that this haplogroup is Paleoeuropean, at least in part, but, quite intriguingly makes earlier findings on modern data suggesting a possible origin or R1b-M269 in or near modern Serbia (Morelli 2010 and Myres 2010) at least somewhat plausible. However none of the Iron Gates R1b is described as R1b-M269 and in some cases it is excluded that it could be this sublineage. 

Thus the issue of the ultimate origins of this key lineage remains open, but let me underline that these Iron Gate individuals belonged to the WHG grouping, as did Villabruna (so far the oldest R1b carrier kown) and that they breach this way the assumed haplogroup homogeneity I2 conceived on merely Central and Northern European samples. Just as happened with mtDNA U haplogroup homogeneity when mtDNA H was detected by several independent studies of Iberian ancient DNA. It is normal to expect more diversity towards the south for several reasons but maybe the most critical of them is just average temperature, which makes the southern lands naturally more fertile (notably so for crops domesticated in the Middle East) and easier to inhabit. 

This trend was only broken in the Middle Ages when the heavy plough allowed the improved exploitation of deep Oceanic soils, being useless in the Mediterranean region of shallow soils however. It was only then when the center of European development moved from south to north, to Belgium specifically, where it remains till present day. So let's take Southern Europe a bit seriously, please.


  1. IIRC Hovhannisyan et al. (2014) reported that the epicentre of variance under M269 is found in South-Eastern Anatolia, presumably extending into Northern Syria. I'd think that this cannot be explained by backmigration and the subsequent convergence of Z2103 and L51 groups alone, since those exist side by side in Eastern Europe as well. Do you think M269 could have diversified in these regions initially? Afaik, ancient samples from those parts are missing thus far.

    1. There have been several studies which have "reported" that but they don't seem to dwell into the structure but only on raw diversity, which can also be produced by multiple waves of immigration. This is clearly the the case of Armenian and to some extent Turkish R1b-M269, which includes "native" (older) local variance and terminal variance from Europe, surely arrived with Phrygians (precursors of Armenians) and other Indoeuropean invaders (Hittites maybe too, why not?)

      However there is also a peak of R1b-M269* in Iran, what makes it possible, in theory at least, that it may be original from there and that might even extend to North Kurdistan, known in Turkish nationalist parlance as "SE Anatolia" (even if it is clearly outside of Anatolia Peninsula).

      However I think I recall that study and it actually talked about Armenians and SE Anatolia understood as the real geographic area, known in the Middle Ages as Lesser Armenia. If it's that study I have in mind, then it talked only of RAW diversity and not STRUCTURED or HIERARCHICALLY sorted diversity, of BASAL diversity, and when you looked at the actual haplotypes, you realized it was pretty much made up of terminal lineages coming from elsewhere, so false alert.

    2. You are absolutely right, I hadn't considered this. The joining network in the supplementary data makes it obvious that both the Lebanese and the Armenia haplotype structure is rather shallow.

      Regarding Iran, I think there are older basal types in various populations in and around Iran as well, no? It'll be interesting to see what comes to light once we have more samples from the wider Middle East & Central Asia.

      BTW, there's a new paper on Ibermaurusian DNA:

      It really sheds light on the (now obvious) African ancestry in the Natufians. Would be cool to read your thoughts on it.

    3. Can you send me a copy to lialdamiz[at]gmail[dot]com? I could I guess register with AAS but I'm pretty sure I was registered and somehow my ID was deleted, so a bit pissed off at Science magazine. Also I have my hands full with R1b-DF27 in pre-Indoeuropean Iberia with very clear pre-Indoeuropean autosomal genome, with another huge paper on Iberian autosomal DNA and right now my brother informed me about some new paper on Denisovan ancestry which finds two different populations of this archaic species influenced us (always in tiny amounts). So getting a bit crazy again with genetics and stuff.

    4. OK, got it. Taforalt is like (aprox.) modern Afar people and ancient (c. 3000 BP) Tanzanian Luxmanda. But to attempt to understand that we have first to visit this other paper I missed, which is a treasure trove of ancient African genetic data:

      Will take me some time...

    5. Was E-L19 where also R-V88 was?
      Hi YFull, can you explain to me how is it possible that these SNPs from the last paper “Pleistocene North African genomes link Near Eastern and sub-Saharan African human populations” from Science (2018) and said belonging to aDNA from 15000 years ago and of E1b1b1a haplogroup are positive only for this sample YF02909 E-A6295 with a MRCA about 1350 years ago?
      Ya Kem, of course, because I have access only to your Group E-L19 that is upstream. But no sample of your haplogroup there was in Morocco 15000 years ago... thus that it were in Europe is possible!
      Thus 15000 years ago there weren't in Morocco the ancestors of E-M81...
      Ya Kem, I have explained above that I was wrong, because I checked the E-L19 group whereas these aDNA from Morocco belong to the E-M78 haplogroup, but, as I said that this demonstrates that anyway 15000 years ago the ancestors of L19 and M81 weren't in Morocoo, even though they could be elsewhere in Northern Africa, but your map is interesting, because it seems that the expansion of PF2431 seems the same of R-V88, thus it will be important that the E-L19 aDNA found in Morocco are checked for instance from Genetiker, for seeing if they are direct ancestors or a parallel line like those found in Gambia etc.


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (it may take some time before your comment is published, thank that to Trumptards).