Extended Data Fig. 5: Characters of intronic 3’-end splicing variants.
(a) Significant splicing variants are significantly closer to the 3’ss. Identity of two distributions was examined by two-sided Kolmogorov-Smirnov test. (b) Mutation effect on splicing efficiency of variants of various pathogenic levels. n = 247 variants. (c) Genome-wide association between the number of intronic splicing regulatory elements near the 5’ss and the intron length, stratified by exon length, related to Fig. 2f. n = 704,953 introns. (d) Enrichment of add-AG variants in the significant splicing variants. (e) Differential exon-intron GC content for non-significant and significant add-AG variants, stratified by intron length, related to Fig. 3f. n = 281 add AG variants. **: P-value of two-sided Wilcoxon test between the non-significant and HC significant variants 9 × 10−4; * 0.01. The boxes in box plots represent medians (central line) and interquartile ranges (IQR; 25th to 75th percentile). The whiskers indicate ±1.5 × IQR from the box or the last data point within that and the dots show the outliers (b,c,e). (f) Preference of 3’ss with various 3’ss strengths was examined by 3’ss swap** assay with splicing minigenes. A representative experiment of three repeats is presented.