Applying the clustering algorithm based on this eight dimensional embedding we discover 16 clusters and 3 outliers. Four of these clusters group the African American, E. Asian, Indian, and Mexican samples. These clusters conform with near perfect accuracy to self-identified ancestry; greater than 99% of the subjects in a cluster self-identified as that ancestry, and only a handful of subjects who self-identified as one of those four ancestries fall outside of the appropriate cluster.
