Here’s another ADMIXTURE analysis of my heritage that uses the same K=12 but with different sample populations for comparison. You can compare this to my prior post to see how my admixture has ‘changed’. This is a good illustration of the limitations of this method: it is highly dependent on what you put in it!
My Asian admixture is gone, and my Arab/Near Eastern admixture is greatly reduced. But not only that, look at what has happened with even my European admixture! It seems to split three ways now: Western (typified by the Spanish Basque population), Southern (Sardinian, which is an island off of mainland Italy), and North/eastern (closest match appears to be equally Russian and the islands off the northern coast of Scotland). I still have the 1% Amerindian; that doesn’t seem to be going anywhere as long as the Pima are in the mix. I’d be interested to see what happened if other, more northern Native American sample populations are included. (Perhaps I should learn how to run ADMIXTURE myself, but I don’t have a computer that runs Linux, and I’m loathe to install a dual-boot.) At any rate, this more closely resembles how I had previously thought of my heritage, but is it actually closer to the truth? It’s hard to say for certain.
Edit: Another interesting thing is what I seem to lack in comparison to your average white American, which is the Berber (North African, with lots of immigration to Europe; think Othello and Morgan Freeman in Robin Hood) genetic input in bright pink. Huh.