fastIBD analysis of Iberia, France, Italy, Balkans, Anatolia and European Jews
The inter-population IBD study is rather interesting IMO as it clearly defines patterns in SW Europe even though I don't get everything to be honest : Heatmap. Interestingly enough, the French Basques seem to share more with Aragonese people than with fellow Spanish Basque people, which is pretty relative though but it might hint to a structure within "Vasconian" lands which has yet to be properly analyzed.
My own results : DOD133
French Basque : 4.93
Pais_Vasco_1KG : 1.84
Aragon : 1.66
Murcia : 0.93
Cantabria : 0.71
Castilla la Mancha : 0.63
Valencia : 0.53
Cataluña : 0.34
French : 0.09
Galicia : -0.05
My own results : DOD133
French Basque : 4.93
Pais_Vasco_1KG : 1.84
Aragon : 1.66
Murcia : 0.93
Cantabria : 0.71
Castilla la Mancha : 0.63
Valencia : 0.53
Cataluña : 0.34
French : 0.09
Galicia : -0.05
I do not understand well the 'heatmap', it would seem that the redder (and hence negative) values are the ones indicating a closest relationship (as the reddest colors usually correspond to internal relations within a population).
ReplyDeleteDoes that mean that you show low affinity with Basques and high affinity with Galicians or viceversa? I would think that vice versa but the values displayed in Dienekes' "heatmap" indicate higher affinity for negative values, so I'm puzzled.
Looks like he reverted the numbers for his heatmap as a closer looks at the results would mean I would then be closer to Cypriots than to neighbouring populations which is not logical. In the comments, Kadu appears to be closer to Cantabrians then to the French Basques (but his results are all below 2 and only 1.05 with fellow Portuguese people).
Delete4.93 is indeed quite high IMO, if I get what this data means, it hints to the fact that Béarnais and French Basque people do share very recent common ancestors.
... "it hints to the fact that Béarnais and French Basque people do share very recent common ancestors".
ReplyDeleteIt's the most logical thing to expect, right? I'd say of all Gascons but very specially Bearnais and South Landais (like Dax).
A bit more surprising is that the affinity between North and South Basques is not too much (per your data and per the preliminary Dodecad clustering exercise in which most North Basques formed their own cluster). Nevertheless in that exercise all "French" clustered on their own bloc, except one (that I assumed it was you) who clustered with Catalans and Aragonese (not North Basques). How would you explain these differences?
I'm indeed that French guy clustering with Aragonese/Catalan people. My idea is that there are not enough Gascon people so I'm assigned to a proxy cluster made of Central Pyreneans as French Basque genetic variation might be too specific for me to totally fit. In previous runs, the algorithm was unable to assign me to a cluster or I defined my own cluster.
ReplyDeleteThe not no high affinity between North and South Basques is indeed surprising though it's somehow still on par with the affinity between 2 Portuguese people : it may mean there is a regional variation within Basque lands which might be quite old, I cannot say for sure (with North Basques being more "Pyrenean" than fellow South Basques). The Murcian results are also rather intriguing.
All in all, there are patterns but some of them are a bit contradictory.
To be honest I do not understand the methods Dienekes uses sometimes. I'm familiar with Admixture and similar algorithms, the rest is a mystery to me and therefore I look at all with suspicion. Why has not he run an unsupervised Admixture run with the IBD+ sample? Beats me but looks strange, because it seems the first thing to do, right?
ReplyDelete