Supplementary Figure 3: Ethnicity inference using PCA (HapMap samples).
From: Comprehensive analysis of cancer-associated somatic mutations in class I HLA genes
Ethnicities of 132 of 133 HapMap samples were inferred correctly based on their projection in the 2-dimensional space defined by the first two principal components. The colored icons show the clustering of the 1,398 training samples belonging to four different ethnic groups. The black icons depict the projection of 132 HapMap samples in this space. (NA12878 was removed from the PCA step as an outlier.) The success rate for attributing the correct ethnicity to each sample was 100%.
