No single nucleotide polymorphisms (SNPs) resulting in CpG site gain or loss were identified within the assayed regions. In addition to analysis using average percent methylation across the 20 CpG sites in our region of interest, we applied principal component analysis (PCA) to these 20 CpG sites in both the Discovery and Replication cohorts. The unrotated correlation matrix was analyzed to output principal component scores. An eigenvalue greater than 1 indicates that PCs account for more variance than accounted by one of the original variables in standardized data. PCA resulted in five PCs with eigenvalues>1 in both samples. The first PC exhibited effects similar to those associated with SLC6A4 promoter methylation values averaged across all 20 CpG sites in both the Discovery and the Replication cohort (Supplementary Tables 2, 3, 5 and 6; Supplementary Fig. 1). In light of the smaller number of datapoints (n = 34 final sample), no PCA was performed in the postmortem cohort, where CpG sites were analyzed individually.
