Fig. 6: Stop codon TGA may favor m6A deposition at and adjacent to Stop codon.
From: Deep learning modeling m6A deposition reveals the importance of downstream cis-element sequences
a Positional plot of average modeled m6A probability around the stop codon, the position 0 was the T nucleotide for the stop codons. The red, green, and blue lines represented genes with TAA, TAG, or TGA as its stop codon respectively. Up panel: regions 500 nt upstream and downstream from 0 position. Bottom panel: regions 10 nt upstream and downstream from 0 position. b Positional plot for stop codon swap. First panel: illustrator of stop codon swap. Second panel: positional plot for TGA to TAA or TAG. Third panel: positional plot for TAA to TGA. Fourth panel: positional plot for TAG to TGA. c Cumulative distribution function (CDF) plot of modeled probability for TGA to TAA or TAG. The p-values were calculated by the Kolmogorov–Smirnov test (KS-test). (the left, middle, and right panel for NRACTRR, TRACN, and TRRACN motifs). d Cumulative distribution function (CDF) plot of modeled probability for TAA or TAG to TGA. The p-values were calculated by the Kolmogorov–Smirnov test (KS-test). (the left, middle and right panel for NRACTRR, TRACN, and TRRACN motifs). e The m6A sites were categorized into two groups (m6A or non-m6A) based on its probability value (the cutoff = 0.05), donut plot of percentage of stop codon for m6A sites and non-m6A sites (Left panel). Bar plot of log2(odd ratio, m6A sites over non-m6A sites) of percentage of stop codon (Right panel). The p-value was calculated by the Fisher’s exact test. f Box plot of conservation score of stop codons with or without m6A sites (n = 1311 for TGA with m6A sites, n = 1562 for TGA without m6A sites, n = 726 for non-TGA with m6A sites, n = 1125 for TGA without m6A sites). Median and interquartile ranges are presented for the box plot. The p-values were calculated by the one-sided Student’s t-test (Significance: **P < 0.01, ***P < 0.001).
