Fig. 3

Effect of DNA methylation status on the degradation and amplification biases. a Coverage of dinucleotides in WGBS datasets from unmethylated and in vitro M.CviPI-methylated TKO DNA prepared with the Heat BS-seq protocol. For direct comparison, the increase in coverage is expressed as fold difference from the genomic average and normalised to the AA dinucleotide. The dinucleotides are grouped as derived from C, G or A/T only and presented in the box-plot panel (right) as total percentage increase in coverage; crosses mark mean values and error bars represent minimum and maximum values. Statistical analysis was performed with one-way ANOVA with Dunnett’s multiple comparisons test against the AT-only dinucleotides; ****p < 0.0001. b Global methylation levels of the in vitro M.CviPI-methylated TKO DNA. The difference between the LC-MS and WGBS values is marked as a methylation ‘artefact’ as in Fig. 2. Significance between the two values was assessed with a two-tailed unpaired t-test, **p < 0.01; error bars represent standard deviation; n.s. not significant. c Read coverage dependence on the G/C composition before and after in vitro M.CviPI methylation. Cytosine content of reads was calculated over 100-bp tiles and matched to the corresponding genomic sequence and not the actual read sequence, where unmethylated cytosines are converted to thymines. Dotted black line represents the tile count distribution in G/C content. d Coverage of CG islands (CGI) in WGBS datasets of mouse WT ESCs (i.e. with similar level of methylation) compared to unmethylated DNA generated with the same library preparation protocols. Heat+KAPA and EpiGnome protocols are included only as WT values for reference, due to unavailability of corresponding unmethylated DNA datasets. Values are expressed as fold-difference from a coverage ‘no bias’ line