For comparison with the creQTL results, we also undertook an eQTL analysis of the EC dataset. Here, the microarray data were normalized to have a mean of zero and variance equal to one within each individual sample. cis eQTL analysis evaluated all SNPs within 5 MB of a gene using an analysis of variance (ANOVA) of expression level on genotype and including covariates identified with surrogate variable analysis [35]. trans eQTL was done similarly, focusing on the remaining SNPs > 5 MB from any protein-coding gene. For cis and trans eQTL analyses, p-values were adjusted separately for FDR [31].
