Fig. 2

Nonlinear principal component analysis (PCA) on 24 gene variants used as variables, to investigate multivariate patterns of variation among the gene variants within the PGx–CNS sample (501 individuals). a PC1,2 (representing a partial amount of total sample variance of 28.01%) showed correlations among gene variants, distinctively forming four groups of individuals. b PC3,4 representing a partial amount of total sample variance of 11.6%, revealed the presence of two specimens with extreme scores (outliers)