Clusters Galore
The Clusters Galore can be found in the spreadsheet. After inspection of the 23 clusters inferred with 21 dimensions, they could be described as:
- Mordvin
- East Slavic
- Polish-Ukrainian
- East Balkan
- Vologda Russians
- Lithuanian
- Central European (combining many groups with small sample sizes)
- A couple of related (?) individuals
- Anatolian
- Greek
- Chuvash
- Ossetian
- A couple of related individuals
- A couple of related individuals
- Balkar
- A couple of related individuals
- Chechen
- Kumyk
- A couple of related individuals
- Adygei
- Lezgin #1 (main)
- Lezgin #2
- Lezgin #3
Inter-Population IBD
Results for Dodecad Participants
Results can be found in the spreadsheet.
If you have joined the Project, please consider leaving a comment in the Information about Project samples thread. That will help others make better sense of their results, e.g., if you find that you belong in the same cluster with some other individual, you might want to know something about their origins.
UPDATE: I have added the IBD sharing matrix.See here on how to use it.
Is it possible to get the full IBD sharing matrix? Thanks.
ReplyDeleteSee updated post.
ReplyDeleteI can't understand how Bulgarians_D could be low score (light blue) with themselves on the matrix, while high score with Ukrainians_Y.
ReplyDeleteIt could be certain individuals have higher scores with other ethnicities or Bulgarians_Y to be somewhat different to Bulgarian_D due to the small sample size but not matching themselves?!
Bulgarians are quite different to Ukrainians, common could be some Ashkenazi admixture, which biases the results. I think Bulgarian DOD815 has some Askenazi ancestry, this is the reason he/she falls in the Greek cluster.
It is not necessary that a population will have their lowest IBD matching with themselves. This depends on how homogeneous a population is. The overall pattern of sharing is such that MCLUST places (most) Bulgarians in different clusters than Greeks and Ukrainians.
ReplyDeleteA simple way to understand this is to think of three individuals: A1 and A2 belonging to one population, and B belonging to another. Now, imagine a triangle formed by these three individuals, whose sides represent their pairwise distances.
If A1 and A2 are very close to each other (homogeneous population), then their distance will be near 0, and will be much smaller than the distance of either of them to B.
If A1 and A2 are very distant from each other (heterogeneous population), then it is possible that the triangle will be very obtuse, and if B is equidistant to A1 and A2, then at the limit, the average of BA1 and BA2 will be half the distance of A1 with A2.
In a small sample, the occurrence of such triplets is possible, and if the sample is truly homogeneous, then increased sample sizes will decrease such effects. However, I suspect that this will not go completely away. For example, the Slavic groups at the top of the heat map have a higher proportion of IBD sharing with Lithuanians, which is probably an indication that they are quite heterogeneous in comparison to Lithuanians, which makes sense, since these are big groups encompassing tens of millions of people.
A different way to see this, is that if Bulgarians always had higher IBD sharing with other Bulgarians (i.e., the Bulgarians were a super-homogeneous population like e.g., Lithuanians or Basques), then they would form their own specific cluster. Such a cluster may yet arise if there are more participants, or it may not.
Finally, lower intra-population IBD could also be the result of the presence of foreign admixture in some individuals as you are suggesting.
This may also be relevant as an explanation for why Bulgarians have high IBD sharing with Ukrranians
ReplyDeletehttp://en.wikipedia.org/wiki/File:Pontic_steppe_region_around_650_AD.png
Well, maybe Bulgarians are not close to each other, however they can't be closest to Ukrainians due to old Bulgars.
ReplyDeleteThe reason is Pontic steppe was devastated several times during different Asian invasions and population completely replaced from 18c on with settlers from Western Ukraine and Russia. So this could be more Slavic, than Bulgar connection.
The latest steppe people there, Crimean Tartars and Nogays emigrated to the Ottoman empire, but due to being Moslims they did not assimilate among Bulgarians, but among Turks. For that reason those that remained did not mix with Christian Ukrainians, either.
Do you have an explanation for the sudden split in some Balkan groups which looked homogeneous until now? What would be your age estimate for the IBD sharing that apparently exists all across Eastern Europe, even reaching the Balkans?
ReplyDeleteCluster 7 looks like Central European Celtic.
ReplyDeleteOne of the most interesting results of fastIBD calculations for me is the connection/clustering of Alpine groups (German_D, South_Tyrol_D, North_Italian) with Romanian, Bulgarian, Hungarian and Tuscan (see also http://dodecad.blogspot.com/2012/01/fastibd-analysis-of-iberia-france-italy.html). And it is simliar with the connection/clustering I noticed in some of Dodecad K12a Oracle calculations.
ReplyDeletea) Is this clustering confirmed also in other calculation models?
b) What can be the background connecting this groups of populations? Has it primarily to do with Anatolian/Tuscans, eastern Celts/Gauls, Slavs, Avars and Magyars/Hungarians as I suspect? Which one could be the dominant admixture connecting these modern european people?
A very interested greeting from the central Alps ;-)
No comments/explanations ? :(
ReplyDeleteI'm sure participants would appreciate some help with interpreting their results.
I think the closeness of Balkan and North Italian population is "pseudo".
ReplyDeleteWe look close as we are the same mixture old Anatolian/Caucasian/Near Eastern neolithic stock with comparatively new invasions of Slavs and Germans.
Since those tribes could not reach the very South of the Balkans and Italy, Greeks and South Italians remain somewhat different to own peninsular, but close to each other.
However Balkanians and Italians are rarely matched on Relative Finder or Family Finder, unless they have family history of more recent connection. This shows the link is very old, that's why I call it pseudo.