Skip to main content
Fig. 6 | Genome Biology

Fig. 6

From: Towards in silico CLIP-seq: predicting protein-RNA interaction via sequence-to-signal learning

Fig. 6

Scoring of 232 splicing mutations from MutSpliceDB along with 6087 control mutations from gnomAD taken in their vicinity, using 40 splicing-related RBPNet models. A Boxplots of RBPNet impact scores from splicing mutations and local controls per RBP. The 40 RBPs are ordered by their adjusted (Benjamini/Hochberg) P-values from Wilcoxon signed rank tests. Title in bold: RBPs with significant P-values at \(\alpha =0.05\) (22/40); orange font color indicates the spliceosomal RBPs. B Impact score distribution for splice mutations (red boxplot) and gnomAD control mutations (gray boxplots) per absolute, relative-distance bin, with impact scores obtained from RBPNet and DeepRiPe models for the spliceosomal RBP PRPF8 (being the most significant model from DeepRiPe following the Wilcoxon signed rank tests). C Scatterplot of area under the receiver operating curve (auROC) calculated from RBPNet models (y) and DeepRiPe models (x). All 30 RBPs annotated as splicing-related in common are depicted, with spliceosomal RBPs in orange. Four RBPs are specifically highlighted for being the only models showing significant P-values from the Wilcoxon signed rank tests applied on DeepRiPe models, and their adjusted P-values are reported (along the adjusted P-values for RBPNet)

Back to article page