Fig. 10: Single-molecule resolution detection of m⁶A reveals co-occurring m⁶A sites and those sites associated with alternative mRNA splicing.

a Schematic of a co-occurring pair of m⁶A sites. The reads mapped to a pair of m⁶A sites were classified into four groups (modified at both, only at the first, only at the second, and in neither of the sites), and the co-occurrence was examined by Fisher’s exact test. b An example of a co-occurring pair of m⁶A sites in SHQ1. The positions of the significantly co-occurring m⁶A sites are shown in red (top left) with m⁶A modification scores of individual reads, categorized into four groups as in a (bottom left). The fraction of reads belonging to each group and the fraction of modified reads at each site are shown (bottom left). The numbers of reads classified based on modification status at the two co-occurring sites are shown (right) with the q-value from Fisher’s exact test with Benjamini-Hochberg correction. c Schematic of an exon skipping-associated m⁶A site. Each read was classified into either exon-included (yellow) or exon-skipped (green) group, and subsequently classified into either modified or unmodified group. The association between m⁶A and exon skipping was examined by Fisher’s exact test. d An example of an exon skipping-associated m⁶A site in TPT1. The read depth of the two types of isoforms generated by exon skipping is shown with the positions of m⁶A sites in red (top). The transcript structures are shown in the middle genomic track with the per-read m⁶A modification scores of individual reads, categorized as in (c) (bottom). 50 reads were randomly subsampled from each group, either exon-included (yellow) or exon-skipped (green), for visualization. The fractions of modified and unmodified reads in each type of isoforms are shown with the q-value from Fisher’s exact test with Benjamini-Hochberg correction.