Figure 2

Validation of MeDIP results. A. Quantification of methylation differences in the abuse and non-abuse groups by bisulfite pyrosequencing analysis of the SLC17A3 promoter and the PM20D1 first exon and intron. DNA methylation at 14 CpG sites in the SLC17A3 promoter and 12 and 1 CpG sites in the PM20D1 first exon and first intron, respectively, among the abuse and non-abuse groups is shown (N = 10 vs. 26 for SLC17A3; N = 9 vs 23 for PM20D1). One-sided t-tests were applied to each CpG site to test for association of methylation levels with childhood abuse, and false discovery rates were calculated for the resulting p-values in order to correct for multiple testing. All false discovery rates (FDR) were less than 0.1, indicating significant association between CpG methylation levels and childhood abuse. **: FDR < 0.025; *: FDR < 0.05; ++: FDR < 0.1; +: FDR < 0.2. The bars represent average methylation for all subjects in a group and error bars indicate the standard error of the mean. Physical maps of the regions analyzed are presented above the charts where CpG positions are indicated by balloons. The transcription start site (TSS) is indicated by a hook arrow. The positions of the primers used for pyrosequencing (Additional file 3: Table S2) are indicated by arrows. B. Replication of the quantification of the differences in methylation at PM20D1 between the abuse and non-abuse groups in an additional 27 males that were not profiled using MeDIP (N = 7 vs. 20). Pyrosequencing was applied to measure the methylation levels of 13 CpG sites in the first exon and intron of PM20D1.