Figure 1

5hmC profiling of mouse liver DNA. (a) An 11 kb promoter array region split into six indicated regions for epigenetic mapping analysis. (b) 5hmC and 5mC enrichment peaks in liver DNA map largely to intra-genic regions: left, distribution of all array probes; right, 5hmC and 5mC enrichment peaks. Chi² values indicate significance of the peak distributions compared to distribution of all array probes. (c) EpiMark qPCR of hmCpG (purple), 5mCpG (red) and non-modified CpG (green) levels over loci in control livers (n = 2). Percentage scores represent frequency of each CpG state over a single MspI site. '5hmC +ve', 5hmC-positive regions; '5hmC -ve', 5hmC-negative regions. Error bars represent standard errors. (d) Box plot showing levels of 5hmC (purple) and 5mC (red) over 1 kb long enhancer and promoter regions. Asterisk denotes significant difference in signal levels (P < 0.001). (e) Sliding window analysis of average 5hmC profiles centered at genes' TSS regions based on their transcriptional activities. 5hmC levels differ over the TSS and flanking regions in a transcription-dependant manner. Highly transcribed genes (green) contain less 5hmC directly over the TSS and greater levels at flanking regions than medium (blue) and lowly expressed (red) genes. (f) 5hmC-enriched TSS regions are largely associated with intermediate CpG content sequences (ICP; red). The CpG density of all TSS regions (left plot) reveals a skew towards high CpG content sequences (HCP; blue) over most promoters. In contrast, 5hmC-marked TSS regions tend to contain ICP promoters. LCP denotes regions of low CpG content. (g) Left: distribution of all probes associated with a CGI (n = 87,234). Right: distribution of a small number of CGI probes that overlap with 5hmC probes (n = 601). Chi₂ values represent the significance of the 5hmC CGI peak distributions compared to distribution of all CGI probes.