Fig. 3
Quantifying and localizing epigenetic heterogeneity within normal kidney tissue. A Boxplot comparing global 5mC between three normal kidneys, 10 regions each. Each dot represents the mean methylation of 843,393 CpGs per region. B Phyloepigenetic tree showing the branching of 30 normal regions obtained from 3 normal kidneys. C Boxplot comparing the Euclidian distances within each normal kidney. Each dot represents the distance between any two samples originating from the same normal kidney. D Boxplot comparing entropy among the same normal kidney regions. Each dot represents the entropy of a sample calculated for 843,393 CpGs. E Boxplot showing the SD of each CpG within each normal kidney. Age and sex of the donors are indicated along with the number of highly variable (SD > 0.1) CpGs in each kidney. The dotted red line intersects the y axis at y = 0.1 and establishes our threshold between high and low variability. F Venn diagram showing the overlap of the highly variable CpG group between the three normal kidneys. G Bar plot showing the relative feature distribution of the highly variable CpGs of each normal kidney (n = 3) and the ccRCCs (n = 12) when normalized to the distribution of all CpGs on the EPIC array over four genomic features indicated. Enhancers and promoters are defined as in Fig. 2. The asterisk indicates a significant distribution difference of the respective feature and the EPIC array total. Y-axis—relative difference (log10) of each feature. H Ontology of enriched pathways derived from 574 and 2027 genes linked to the 368 active enhancers in the normal group and the 5,000 most variable active enhancer CpGs from the tumor group, respectively, from panel G, using IPA. The color bar represents the p-values as −log10