Fig. 4

Enriched RNA motifs related to the RBP groups. a The histogram shows the genomic distribution of group-related binding sites for each group (with at least two RBP components). b De novo sequence motifs and known functions reported in literatures. The p values of the motifs were calculated by binomial test against the randomly selected genome background. The false discovery rate (FDR) was calculated using HOMER (option “–fdr 1000,” so the smallest number is 0.001). c Percentages of known motifs in the binding sites co-bound by each RBP group and in the binding peaks identified from individual RBPs. The percentage number is the portion of binding sites that contain the named motif (e.g. UGUA or AAUAAA). The average and standard deviation of the percentages for individual RBPs are also shown. The binding peaks of single RBP are the overlapped ones identified by both Piranha and PARalyzer using default parameters. The enrichment was tested with Fisher’s exact test between the group and individual RBPs. Four numbers were used in the test for each motif, total number of binding sites associated to the group, total number of binding peaks associated to an individual RBP, total number of binding sites and binding peaks containing the known motif. (***p value < 0.001) d We also predicted the de novo motifs (top one is shown for each) using HOMER from the binding sequences of group 4 and individual RBPs