Fig. 4: Detection capabilities for m6A among sequence motifs.

a, b Relative recall rates (a) and precision (b) for 12 motifs bearing the consensus sequence RRACH. Relative recall rates/precision were calculated as the ratio of the recall/precision for the sites in a specific motif to the overall mean (across all sites). A m6A profile created from miCLIP results was used for validation. c Current intensity of modified and unmodified sites in GGACA or GAACA motifs. The m6A candidate sites (methylation rate of WT–KO > 0.8) were identified using m6A-REF-seq and marked as “Mod” (modified, sites in WT sample) or “UNM” (unmodified, sites in KO sample). This dataset was also used in d, e. For boxplots in c, d, the upper and lower limits represent the 75th and 25th percentiles, respectively, while the center line represents the median; upper and lower whiskers indicate ±1.5× the interquartile range (IQR). The number of reads and events aligned to the sites are indicated in parentheses. d A comparison of mismatch and deletion frequencies for modified and unmodified sites in GGACA and GAACA motifs. The number of sites is indicated in parentheses. e IGV snapshots of two specific sites in different motifs (GGACA and GAACA). Modification status (modified or unmodified) was determined by m6A-REF-seq. Sites with allele frequency ≥0.1 are indicated with corresponding colors. Coverage for each site is indicated on the left of each short sequence.