Substack cometh, and lo it is good. (Pricing)

Turks: Greek or Armenian?

A new paper on Turkey, The genetic structure of the Turkish population reveals high levels of variation and admixture:

We delineated the fine-scale genetic structure of the Turkish population by using sequencing data of 3,362 unrelated Turkish individuals from different geographical origins and demonstrated the position of Turkey in terms of human migration and genetic drift. The results show that the genetic structure of present-day Anatolia was shaped by historical and modern-day migrations, high levels of admixture, and inbreeding. We observed that modern-day Turkey has close genetic relationships with the neighboring Balkan and Caucasus populations. We generated a Turkish Variome which defines the extent of variation observed in Turkey, listed homozygous loss-of-function variants and clinically relevant variants in the cohort, and generated an imputation panel for future genome-wide association studies.

First, I’m surprised how inbred the Turks are in this paper. They need to get more secular quickly and stop marrying their cousins. Second, there’s the classic issue of assuming East Asian ancestry = Turkic ancestry. The reality is that by the time the Turkic tribes arrived in Anatolia they’d already mixed with Iranian peoples in Central Asia, so they may have been 50% non-East Asian by that time. Here’s the relevant section: “Paternal gene flow based on Y chromosome haplogroups C-RPS4Y and O3-M122, which were previously implicated as Central Asian specific, ranged from 8.5 to 15.6%. Maternal gene flow based on mtDNA haplogroups D4c and G2a, which were previously suggested as Central Asian specific, was 8.13%.” For what it’s worth, 4% of the Y’s are R1a of the Slavic variant and 8% are R1a’s of the Indo-Iranian variant. The main issue with the latter is that some of this might be Turkified Kurds.

But what I’m really interested in is which populations modern Turks are genetically close to. In the argument of whether Turks are Greek or Armenian, these pooled Turks seem more Armenian in the heatmap. I pulled the Turk subgroups and created a table of Fst values. Nothing super surprising.


20 thoughts on “Turks: Greek or Armenian?

  1. Curious about the close relationship between Kumyks and Armenians/Anatolian Turks. If I’m reading the chart right, Kumyks are about as close to those groups as they are to Lezgins, and much closer than they are to any other N. Caucasus group. Lezgins, for their part, don’t have any particular closeness to Anatolian groups.

    What’s going on here? Kumyks are linguistically just another Kypchak Turkic group, and I’m not aware of any historical connection between them and Anatolia.

  2. According to the Turkish papers there are almost 9 million of Turks which have origins in Serbia (3 million in Istanbul). Some of them are several hundreds of years old emigration which still remember their Christian origins and some speak Serbian. Many Serbs lived in Asia Minor during the East Roman Empire when they gave Emperors (e.g. Justin and Justinian). During the Ottoman Empire Serbs gave (between 1543-1612) 13 Great Viziers (prime ministers), 7 GV deputies, 23 viziers, 8 admirals, numerous province governors, etc. Before Greeks came to Asia Minor, ancient Serbs who came from Balkan lived there (e.g. Brigians i.e. Phrygians, Dardanians (e.g. Troy), Lydians, etc), who were Hellenised after Alexander.

    So, there is not even mentioning that Serbs ever existed in Turkey in this paper.

  3. These genetics based research reveal much about the Seljuk-Ottoman dynasties. As an invading force from the east starting 1500 years ago, most of the Balkans, Asia Minor and Middle East succumbed to their military and political domination.

    The findings of genetic research and related histories substantiate evolution of the invading Turkic population and their concentration in what today is Turkey.

  4. You have to be careful talking about “Serbs” in Justin’s time. He was an Illyrian whose mother tongue was Latin. Justinian also was fluent in Latin. The Serbs came later (as “Slavs”) with the Avars. Illyrian DNA is, I think, R1b like the modern Albanians’.
    Centuries later, the Macedonian (in original location) dynasty which includes the famous Constantine VII and Basil II (excepting the occasional in-law like John Tzimiskis) might be Slavic at base. I should not say this of the Justins.

  5. If I am reading this correctly, Syrians are genetically closest to Armenians? And furthest away from Romanians?

    I would havve assumed Assyrians and Lebanese would be the groups closest to Syrians.

  6. @Razib, off topic, but I see from your twitter feed you’re publishing an article on steppe interactions with Europe and elsewhere.

    New paper just dropped which could have some significant findings for our understanding of how Steppe ancestry arrived in Europe –

    “Dynamic changes in genomic and social structures in third millennium BCE central Europe”

    “Europe’s prehistory oversaw dynamic and complex interactions of diverse societies, hitherto unexplored at detailed regional scales. Studying 271 human genomes dated ~4900 to 1600 BCE from the European heartland, Bohemia, we reveal unprecedented genetic changes and social processes. Major migrations preceded the arrival of “steppe” ancestry, and at ~2800 BCE, three genetically and culturally differentiated groups coexisted. Corded Ware appeared by 2900 BCE, were initially genetically diverse, did not derive all steppe ancestry from known Yamnaya, and assimilated females of diverse backgrounds. Both Corded Ware and Bell Beaker groups underwent dynamic changes, involving sharp reductions and complete replacements of Y-chromosomal diversity at ~2600 and ~2400 BCE, respectively, the latter accompanied by increased Neolithic-like ancestry. The Bronze Age saw new social organization emerge amid a ≥40% population turnover.”

    (Apologies for breaking the line of Turkish genetics discussion!)

  7. @Matt

    Fascinating stuff.
    Your post made me run over to Eurogenes to read the comments there. Got a big chuckle out of your one regarding a certain theory on where all Indo-European languages are originally derived.

  8. @David Ross

    It requires a longer response but just briefly. All what you wrote is a falsification, the most is against the common sense. Latin is a mother tongue in a remote mountain village far from anything in south-eastern Serbia? It sounds as a joke. I can imagine Justinian’s Latin speaking mother how she sings Latin songs while shepherding her sheep flock. Obviously, you know nothing about Albanians and their genetics. Illyrians lived in former Yugoslavia and in the whole region. Serbian ‘embassy’ in Vienna until the 19th c.AC was called for centuries ‘Illyrian office’. And what’s happened with Illyrians when sc. ‘Slavs’ allegedly came in the 7th c.AC? Where they disappeared? Illyrians were the iron fist and the elite legions of the Roman Empire. 40% of Serbs have I2 genetics which is present in Vinca in continuity since the Ice Age and before.

    So, you assert that Decius, Hostilianus, Claudius II “Gothicus”, Quintillus, Aurelian, Probus, Diocletian, Maximianus “Herculius”, Constantine I, Constantius II, Jovian, Valentinianus I, Valens, Gratian, Valentinianus II, Marcianus, Anastasius I, Justin I, Justinian I, etc, were Albanian Emperors?

    (Future) Albanians first time stepped on Balkan soil in 1043 AC. What’s happened with a mother tongue of all previous Emperors? Who were Brigians (Phrygians)? Thracians? Who was Alexander the Great?

    If you find only ONE primary account from anyone within 200 years after this alleged Serbian migration to Balkan in the 7.cAC (confess, you found this in wiki), I will accept all above.

  9. Since Turks have that good chunk of Central Asian related ancestry, I assume they’ll be brought closer/have lower Fst to more eastern populations, even if the main local input in them is more “western”? And the Greek sample included here doesn’t look (at least fully) Anatolian, since it has lowest Fst with Tuscans and Balkan populations so it’s probably not the best point of comparison. Looks weighted towards mainland Greeks.

    We have two Ottoman-era Central Asian admixed Anatolian samples so far and they differ quite a bit in the amounts of their Central Asian related admixture. Seeing a while back what comes up with the those and with the addition modern populations from the Eurogenes dataset in Vahaduo, the most important post Central Asian input in Anatolian Turks seemed to come from Anatolian Greek like populations overall though predictably with east to west differences (more eastern groups = more Armenian-like local input), even when considering Caucasian, Balkan, Kurdish and Aegean groups i.e. the other major potential Ottoman-era inputs. But it’s probably hard to differentiate overall with that kind of method if a few different populations are involved, especially since it’s sensitive to your setup and we still have only two very different Ottoman samples. That kind of model might change a lot with more non-contemporary samples but my two cents, anyway.


    Thanks for the tip. Seems really interesting with regard to the discussions people have been having on the issue. I’ve had a very cursory look at it so far. I notice there’s even an early Q1b2a, like we see in some of the fully Steppe_EBA-like Afanasievo samples. A point I caught with the quick look is that they identify the extra ancestry early CW has in comparison to Yamnaya as “intermediate HG-like” which would be in line with what the current early CW samples in G25 look like too in comparison to Yamnaya, though I guess there might still be some questions about potential steppe clines and those already “horizontally” different Sredni Stog and Yamnaya samples.

  10. In relation to one of previous comments and still related to the topic… In the recent podcast, David Anthony avoided to answer the question about the approximate number of Yamnaya people who allegedly brought their, sc. “Indo-European” language, which was enforced on the whole Europe and influenced all European and many Asian languages. I could not find that anyone speculated about these numbers. There is probably a reason for this.

    Let me, based on my previous readings and my common sense, speculate these numbers which could, after several iterations, make some sense and validate sc. ‘Indo-European’ hypothesis. I propose that in Europe lived not less than 1 million of people and that not more than 25.000 Yamnaya people came to Europe. So, how in this context sounds sc. IE hypothesis? I expect that readers with better knowledge correct my starting figures.

  11. 23andMe has me as 1.2% “Iranian, Caucasian, and Mesopotamian”. I do have about 10 DNA matches in that area, including two in Turkey. Also, I was stationed in Sinope, in 1970-71, and I speak Turkish. I play the baglama and the l’oud. As Razib knows, I am R1a1a1h. No known Turkish ancestry.

  12. @philip, ah. Was trying not to be too harsh but that poster I’m responding to is a fairly nice guy but… also… (paraphrasing) “You won’t believe it! The true original Indo-Europeans came more or less from the same part of Poland as I do *and* they spoke more or less exactly the same Slavic languages as I do (not fake reconstructed pIE) *and* they followed the same religion as is my personal neopagan philosophy (which is also the original source of all high philosophical and religious thought) *and* further their religion was built on ceremonies of getting drunk on beer and then singing songs and discussing important events, just like me and my mates down the bar. German ‘scientists’ have hidden this all from us! Here, let me show you via etymology…”.. I mean… C’mon.

    On the paper: “First, I’m surprised how inbred the Turks are in this paper. They need to get more secular quickly and stop marrying their cousins.”

    I was surprised by that too. Doesn’t seem much lower than Egypt, Syria, Saudi are coming out in genome wide studies.

    I had a stylized impression that Turkey was more “modern” and therefore would have lower cousin marriage (because “the high cousin marriage stops the modernization!” + “Modern life means you meet more people and cousin marriage ends”). Doesn’t seem to be so or pretty weakly so.

  13. @Matt
    Is that modernization vs expected change in culture process what we would call a cargo cult mentality?

  14. @DaThang You mean like, cultural change due to copying of what modernized cultures seem to have/do?

  15. RE: Unexpected levels of Turkish inbreeding

    Maybe there’s a bit of historical lag effect going on? For the past 100 years or so during which Turkey’s been “modernized,” cousin marriage among ethnic Turks has become generally obsolete, however in the preceding centuries (pre-20th) cousin marriage was as common among ethnic Turks as it is most other Muslim MENA societies, and the past 3-4 generations haven’t quite been enough to break up all the close IBD tracts?

  16. @David Ross:

    Do yourself and the rest of us a favor and never mention that balkan land in your comments here. it just sets the dude off.

    I have learned to avoid him and not read his comments.

    I am happier.

  17. It seems, Walt is still distressed and confused after reading the neighbouring thread. Why you worry about DR writing if you do not read my comments? I remember that you left a discussion on BP without answering the question and did not come back for months. I do not write and care only about Balkan than about many other things. You are nervous because you do not have arguments. I am sure that you are very carefully reading my comments and I challenge you to find one incorrect or illogical assertion. You simply cannot handle the truth and it is visible from your reaction on some information which is questioning the official ‘Indo-European’ narrative and about what I’ve been talking for a while.

    Don’t worry, be happy! And, stay cool!

  18. @Mick
    The samples as I believe are taken from a study specializing in ALS patients and of course inbreeding puts you at a much much higher risk for ALS so it’s natural that the study would be skewed towards a such result. Inbreeding in Turkey is very low except for the Southeast of the country.

Comments are closed.