Saturday, November 6, 2010

ADMIXTURE analysis with Dodecad Populations

Thanks to our project members, I am now able to repeat the ADMIXTURE analysis which launched this product, with 5 new populations, composed entirely of Project members. I have made a spreadsheet where you can find admixture proportions for the original 36 populations, as well as the 5 new ones:

The new populations are: Assyrians, Scandinavians, Greeks, Finns, and South Italians/Sicilians. As more people sign up to the project, more populations will pass the 5-person threshold, and will be included in the admixture analysis, and the admixture proportions of existing ones will be fine-tuned. Moreover, some of the populations, such as Scandinavians, may be split into e.g., Swedes, Danes, and Norwegians.

Stay tuned for a notice about future opportunities to submit your data.


  1. I just noticed something interesting in the last samples analysed. Individuals 168 and 169 who seem to be from NorthWest Africa (Tunisia ?) show up to 8% North European ancestry, so much higher than what is found in Cypriots and similar to Turks, for exemple. If I remember, this component was almost not found in the sample from Morocco you analysed previously.

    So, do you think it can be linked to Vandals invasions in Tunisia in the 5th century or is it pre-historic ?

    1. In the recent beylical Tunisia (1530 - 1950) the mamluks played an important role, they were slaves brought from europe and then became rulers. Corsairs were also very active around the coast and they brought slaves from europe, a lot of families had to settle in Tunisia.

  2. We can't arrive at any conclusions based on 2 samples, but that's an interesting observation. There are certainly links between North Africa and Europe, including the Vandals, so that might be an explanation. Then there are Coon's blond Riffians. Who knows what is going on; I'll be sure to include North Africans in my next call for samples.

  3. Perhaps the north European (orange component) amongst Greeks corresponds to Centum Indo-European speaking (see that component amongst Uyghurs with Centum ie Tocharian legacy)
    And since Greek is a centum language but in the same times classified under the central ie branch together with Indo-Iranian and Armenian (for common morphologica features as well as some isoglosses) it could be that Greek is the result of amalgamation of Western Asian(light blue) Satem IE speaking+North European(orange) Centum IE speaking.

    The South European(light green)component perhaps would correspond to the pre IE "Pelasgians" and account for the substratum words of Greek that does not have and IE etymology

    The South-Western European(light red)component perhaps would correspond to the Phoenicians (see legend of Kadmos and Europe as well as toponyms of Semitic origin such as Salamis and Thebes)and will explain the (ancient)Greek words that have Semitic etymology(see Chantraine)

  4. Ashraf, why do you associate the Western Asian component with "satem languages" and the Northern European one with "centum languages" when the Northern European component culminates in "satem speaking" Slavs and Balts and non IE-speaking Finns?

    I don't think any of the components here can be associated with Indo-European languages. If anything the West-Asian component seems to be related to the spread of agriculture, and the Northern-European one to indigenous Northwestern Eurasian paleolithic hunter-gatherers...
    If anything, I think the proto-Indo-European probably carried those two components in more or less equal amount.

    It would explain why non-IE Basques and Georgians have (almost for the latter) only one of the two components, while surrounding IE-speaking populations usually have at least a bit of both.

  5. carthage and the algerian coast were settled by people from italy and other parts of the north western med. during roman times as well.

  6. Balto-Slavic despite being a Satem branch shares common morphological features with the western IE languages but have underwent dramatic developpements grace of its symbiosis with various Iranic speaking folks in the Pontic steppes as well as-probably-southeastern European actual Bulgaria wich is the home when old Slavic developped and was first written around the 9 th century(Scythians,Sarmatians,Cimmerians)and as you know Balto-Slavic is younger than the highly diversified Iranic branch of Indo-Iranian .
    Northwestern European hunters gatherers can not account for IE language since IE originated in Western Asia.
    Northern Europe (wich I do think originated in Western Anatolia=>çatal höyük culture) seems to be populated essentially populated by Anatolian farmers (R1b+R1a)and I do think that it's the southern European component(wich is most probably the result of the south European refugium during LGM)one wich do accounts for the pre IE languages (Pelasgian,Aquitanian,Raetian,Iberian,Tartessian,Ligurian,Pelasgian...)
    But of course all those are just speculations.
    And it could be that the north European component is also "indigenous" and did spoke the language connected with a substratum layer in Saami before shifting to the language of the( bronze age dated)advancing Western Asian Indo-Europeans(since the north European component is nearly absent from the Indo-Iranian speaking south Indians and Brahmins!)
    Please see the thread below

  7. Good observations, Ashraf. Thanks for your answer.
    Thread in the link was a bit of a let down, though, I've seen these scythomaniac arguments over and over again on these boards.

    I have made some maps using the importance of each components from Dienekes' spread sheets, and the Southern European component overlaps quite well with the spread of Centum languages (as seen on the map on wikipedia about them). Suggesting Centum might have been the result of a substratum from these pIE people you mention.

    Also, do you consider the Northern European component to have been present in the Basque (and other pIE SW/W Europeans) since their ethnogenesis, or is it due to a more recent admixture (postdating the appearance of early forms of the Basque language for instance), according to you?

  8. Dod168 is my sample and Dod169 is my wife sample, we are both from Msaken, Tunisia
    The relatively hight percentage of North european component (8%) in my sample , yes it can be explained by teh presence of Vandals, I was wondering about that when I found the mitochondria haplogroup is T2b3 which is rare in Tunisia but more common in Europe,
    it can aslo be explained by the slave trade, the presence of mamelouks in Tunisia.
    I noted also that North europen coponent is more less or absent in the other north african samples (algeria, morocco)

  9. One more fellow from Msaken, Tunisia has tested through FAMILY TREE DNA Family Finder
    He has sent his raw data I think to Dienekes
    I hope to see his analys result in Dodeca sheet soon
    He is matching with me and with my wife in and we have the Y DNA Haplogroup
    J2a4h2f1, sharing a common Paternel ancestor not more than 800-1000 Years before present.