The covariance of genotype-phenotype associations was determined by applying sparse principal component analysis (sPCA) to the genotypic profiles (vector of association effect sizes of each SNP) of all 36 phenotypes (9 electrodes × 2 stimulus modalities × 2 sexes) for each age range. Using sPCA yields a small set of SNPs which account for the bulk of the difference in genotypic profiles between the different phenotypes at each age range. The first two components of the sPCA were sufficient to represent the phenotypic variance between both stimulus modalities and sexes with respect to their pattern of genotypic associations. Bootstrapping confirmed the reliability of these results.
