Fig 3. The results of PCA and hierarchical clustering based on variant vectors.

(A) Biplot of principal component analysis based on variant vectors. The dots show the eight B. subtilis strains, and the upper left image is an enlarged image focused on the three Japanese strains located near (0, 0). The fist principal component features the non-Japanese strains, and the second principal component can be regarded as a feature to distinguish strain LaoA1 and the other non-Japanese strains. (B) Hierarchical clustering of the eight B. subtilis strains based on the Euclidean distance between variant scores of each strain using the furthest neighbor method. The different cluster indicates that strains have different variant score patterns. (C) Geographical location of each country.