Originally Posted by
Petalpusher
It depends what the components are based on, they are a bit different than the K15 calculator, and seeing the spreadsheet now there is no very high peak specially among the euro components (30-50% ish) so i can see how Eurogenes improved that part a bit on K13/K15 (EUtest is just a pre-version of K15). It can create more deviation in the admixture matches for some people but it still puts almost exactly the same than the K15 Pca with the better matches of the multi pop / mixmode. I tried to place the same intermediates and it got me on the same point than on the K15 pca, a few pixels off. It just creates some "visual effect" in the single pop list i think.
The thing is the more specific Euro components you create, you are getting more prone to create artificial differences. For example if i was creating 4-5 components peaking in all corners of UK i would probably create great "distances" between all of the people of the isles, while you are obviously all very close. So there s always a sweet spot to try to decipher population that are close without getting into too much artificial differences that don't really exist.
From what i ve seen people in NW France usually get FR, W-C German, Cornish first in different orders, comes down to individuals.