Figure 2

Heat map of distributions of reads acquired from next-generation bisulfite sequencing. Three independent preparations of ESC mitochondrial nucleic acid (mtNA) (E1, E2 and E3), two independent preparations of brain mtNA (B1 and B2), two independent preparations of liver mtNA (L1 and L2) and a preparation of synthetic mtDNA (S1) were subjected to bisulfite conversion for 5, 15, 40, 60 and 90 min (5, 15, 40, 60 and 90) and a total of 24 bisulfite-converted samples (Table 1) were deep-sequenced. Reads were sorted to corresponding samples with distinctions between L stands (L) and H strands (H) of mtDNA (a), and between plus strands (p) and minus strands (m) of spiked λDNA^−mC (b). Unconversion rates of cytosines in each read were calculated from statuses (converted or unconverted) of all cytosines in the read, and were expressed as percentages. Reads were then categorised with 10% increments of unconversion rates, and distributions of the reads in each increment block were calculated as percentages and are presented in heat maps; colour changes from white to red indicate from 0% to 100%. Calculated percentages of read distributions in increment blocks for each read assembly are shown in Supplementary Fig. S2e,f.