I imagine they do some sort of cluster analysis to find correlation along with self-identification. If so, then this is undeniably junk science based on junk-in junk-out statistical models (which is often the case with cluster-analysis).
I’m simply curious here, cluster analysis is the only method I can think of (other than guessing/categorizing arbitrarily).
So when our thread’s ancestor (pun unfortunate) says “45% Levantine” they mean they 45% of people alive today that are from that region (which includes many European immigrants). I bet this gets very messy given immigration patterns. Like which immigrants count as ancestry sample, and which don’t? For this problem I would personally use cluster analyses, however I would probably simply give up, knowing that cluster analysis would give me junk (and ultimately arbitrary) results with such noisy (and potentially skewed) data.
I don't think it's "some" -- I think that's 100% of it.