Fig. 4

Localization of differentially hydroxymethylated regions. a EWAS results of the high 5hmC CpGs (n = 37,173) comparing tumor and non-tumor tissue from an age and sex adjusted linear model (limma). Volcano plot of differentially hydroxymethylated CpGs, plotting the difference of log2 fold change in beta value and the respective negative log10 of the unadjusted P value. Blue points represent 726 differentially hypohydroxymethylated regions (DHMR) with an adjusted P value (FDR) < 0.1. b Forest plot demonstrates enrichment of differentially hydroxymethylated regions within the high 5hmC loci. c Oxidative bisulfite derived mean 5mC beta values at these DHMRs stratified by tumor type demonstrates significant hypermethylation at these loci in tumors as compared to controls. However, bisulfite treatment alone only cannot distinguish between 5mC and 5hmC and demonstrated no change or hypomethylation. d We leverage the locus overlap analysis package (LOLA) to determine the significance of overlap of hypohydroxymethylated sites with binding sites of specific transcription factors profiled by ENCODE. The top 12 enriched transcription factors discovered by LOLA analysis are represented above. TFs are plotted on the x axis and the y axis represents the -log10 q value (corrected for multiple hypothesis testing). The size of the bubbles represents the odds ratio and the color represents the cell line. The dotted line represents a q value of 0.05. CTCF is the only transcription factor that meets statistical significance