Figure 1: MBDCap-seq identifies DMRs in discovery cohort.

A heatmap showing methylation profile of 822 hypermethylated (a) and 43 hypomethylated regions (b) across a cohort of 19 tumour and six matched normal samples in the discovery cohort. Columns are samples and rows are regions. The level of methylation (number of reads normalized with respect to fully methylated sample) is represented by a colour scale—blue for low levels and red for high levels of methylation. (c) A bar plot showing association of DMRs across functional/regulatory regions of the genome—(i) CpG islands and shores, (ii) RefSeq transcripts and (iii) Broad ChromHMM HMEC annotation. The height of the bars represents the level of enrichment measured as a ratio between the frequency of hypermethylated (pink) or hypomethylated (blue) regions overlapping a functional element over the expected frequency if such overlaps were to occur at random in the genome. Statistically significant enrichments (P value <0.05; hypergeometric test) are marked with an asterisk. (d) Sequenom validation of five hypermethylated regions—FERD3L, C9orf125, HMX2, NPY and SATB2—is shown for an independent cohort of TNBC samples (normal n=15; tumour n=33) and (e) a panel of breast cancer cell lines (normal n=3; cancer n=24). For each region, box plots displaying the distribution of methylation levels are shown in grey/blue for normal/tumour samples/cell lines. (f) A bar plot showing enrichment of genes with promoter hypermethylation in sets of genes that are up-/downregulated in the TCGA cohort of TNBC tumours as compared with matched normal samples. The height of the bars represents the level of enrichment measured as a ratio between the observed number of up-/downregulated genes with promoter hypermethylation to the expected number of such genes. (g) A Venn diagram showing overlap between genes with promoter hypermethylation, genes downregulated in TCGA TNBC cohort (hypergeometric test; FC 1.73; P value ≪0.001) and genes with two or more mutations (hypergeometric test; FC 1.92; P value ≪0.001) in TCGA breast cancer cohort.