Fig. 3: Trial of direct nanopore sequencing of 5hmC-immunoprecipitated (IP) DNA.

a Nanopore sequenced hMeDIP peaks overlapping the Kcnj11 gene (chr7:45,746,000-45,751,000). For each repeat, tracks show the proportion of direct 5hmC calls at each CpG site (top) as well as sequencing coverage depth and the corresponding MACS2 narrow peak (bottom). Lowest tracks show WGS data from input. Produced using pyGenomeTracks (v3.8)⁹³. b Bar plot of primary genomic context overlapped by peak regions, comparing the direct nanopore hMeDIP (N = 3) with public hMeDIP-seq data (N = 3). Genomic background, based on mm39, provided as reference. Error bars indicate 1 standard deviation. c Bar plot showing modification states as a percentage of all CpG-context cytosine base-calls within the PromethION WGS data and hMeDIP peaks (N = 3). Error bars indicate 1 standard deviation. d Density plot underlay in shades of blue shows 5hmC Z-Score for matched 500 bp windows of WGS data from nanopore and TAB-seq. Z-scores are calculated using the arcsine transformed proportion of all CpG base-calls (merged across replicates) enclosed in a window detected as 5hmC. Peaks from all nanopore hMeDIP-seq replicates are overlaid onto the window they intersect as a scatterplot, with size and hue proportional to fold enrichment over input. Dotted line indicates sample mean 5hmC Z-Score for the (x) TAB-seq and (y) nanopore datasets. *p < 0.05; **p < 0.01; ***p < 0.001.