PCA and ancestry

Principal component analysis run in the browser on simulated genotypes. The mathematics is ordinary PCA. The biology is what is new: the top axis explains little variance, admixed people sit between clusters, and a gradient of populations produces a smear with no natural number of groups.

Vibe-coded with ❤️ by Stephen Turner.

PCA

Individuals-
Markers used-
PC1 variance-
PC2 variance-
Estimated Fst-

The same thing in R