Fig. 4. Distinct characteristics of methylation canyons categorized by methylation concurrence.

Average profiles of methylation concurrence, average methylation, DNase I hypersensitive sites, H3K4me3, and H3K27me3 in methylation canyons in CD3+ T cell. ‘aCanyons’ (red) are low-concurrence canyons. ‘pCanyons’ (blue) are high-concurrence canyons. The X-axis indicates the distance to canyon borders. b Expression of aCanyon target genes (red, n = 1212), pCanyon target genes (blue, n = 438), control group which features a similar expression distribution as aCanyon Targets (purple, n = 1212), and control group of randomly selected lowly expressed genes (gold, n = 1,000). The line in the box center refers to the median, the limits of box refer to the 25th and 75th percentiles and whiskers are plotted at the highest and lowest points within the 1.5 times interquartile range. Lists of aCanyons and pCanyons are in Supplementary Data 3, together with their target genes. c Functional enrichment analysis of aCanyon target genes, pCanyon target genes, and control groups. Enriched gene counts in each group are indicated on the left side. P-values were measured by two-tailed Fisher’s Exact Test and adjusted by the Benjamini–Hochberg method. d Spearman correlation between gene expression and average methylation of 100 bp-bin in gene regions. In total, 120 bins from −2 kb to 10 kb were measured. e The profiles of average methylation for each gene group. The standard deviations are indicated by the width of the shaded area. f The Canyon distribution around the TSS. In each position from TSS−10 kb to TSS+10 kb, the percentage of aCanyon (red) and pCanyon (blue) target genes covered by Canyon is shown on the Y-axis. g Relationship between fold enrichment and 5mC preference of TF motifs. Each dot represents a motif. Y-axis indicates the fold change (log2) between enrichment at aCanyon and enrichment at pCanyon of the same motif (see Methods). The X-axis shows the 5mC preference of motifs measured by the SELEX technique. ‘methyl-plus’ TFs prefer to bind methylated sequences, while binding of ‘methyl-minus’ TFs are not favored by 5mC. The list of ‘methyl-plus’ and ‘methyl-minus’ TFs are in Supplementary Data 4. Spearman’s rank correlation was used. P values were calculated by the two-tailed correlation test for Spearman’s correlation. A linear model was plotted to describe the relationships between variables (indicated by the dashed line).