To make more sense of STRUCTURE results i've made a quick STRUCTURE analysis (of the same dataset, but this time on the whole set without prior grouping of chromosomes). And, voila now we have two perfect Baltic clusters. The first one is somewhat similiar to the EUROGENES project's Southeast-Baltic (or, Balto-Slavs) and second one, - that reaches its peak in Finns (Northern or Nordic :) Baltic?)
Isn't it kinda odd that CEU (which is normally identified as being of mostly English and smaller Danish ancestry and tends to overlap with North Sea and maybe South Baltic peoples in Europe) and Romanian are so extremely identical? Haven't you duplicated some data with two different labels?
ReplyDeleteI strongly suspect that you have duplicated the Romanian-average sample under the CEU label because CEU should be close to the English samples in fact (50% purple, 30% blue, 10% red maybe).
Also I'd be curious about the origins of V184 and the other few people who display such a big light blue cluster. To a lesser extent, V157 could give info on the origin of the orange cluster.
How do you decide where to stop with the K runs, or, in other words why is not more informative K=5 than K=6, etc.?