>>3842051
>we can see clearly that different racial groups cluster together
The scientific problem with "racial groups" is that, as you agreed with the other anon, there's no consistent concept of "race". Raciology had centuries to come up with one and it never did. The very PCA you posted validates existence of only one pronounced race - negroid, that is, while neither "Europeoids" nor "Mongoloids" form anything approaching a coherent exclusive cluster. And that's with those insanely uneven components (that's the percentage values next to the PC#). And that's before we even consider the issue that it's a PCA plot. And you don't understand PCA and why it's not really a credible analytical method. I can take an open human dataset, fire up RStudio and very easily get you a plot that clusters blacks squarely with Irish, decisively proving once and for all that Irish are the niggers of Europe. See picrel. It's just the way PCA or any other dimensionality reduction works, actual studies find clustering using other methods (usually Bayesian clustering for a series of priors with MCMC), and then use one of the more similar PCAs on the same dataset as a visual aid, because they are much pretties and easier to read than Bayesian plots (as evident from online haplofuhreres really loving PCA plots). And then, clustering is one of the weakest evidences for, basically anything, from phenotypical traits to ancestry. And we didn't even being touching the sampling issues.