Skip to main content
Fig. 4 | Genome Biology

Fig. 4

From: Positional motif analysis reveals the extent of specificity of protein-RNA interactions observed by CLIP

Fig. 4

Differential enrichment of motif groups in eCLIP relative to in vitro methods. a Recall values for PEKA and “local” approach across 57 eCLIP datasets with available in vitro data (Additional file 8: Table S7). b Percentage of k-mers with significantly differential ranking (Welch’s t-test p < 0.01 and a fold change greater than 1.5 or less than 0.66 between each eCLIP analysis approach and in vitro data. Differential ranking was assessed on the 41 eCLIP datasets for which both mCross (raw scores) and in vitro data were available. c Swarmplot shows mean k-mer ranks across RBPs (n = 41) in each approach. K-mers which contain two or more of the same dinucleotide are colored by their most common dinucleotide and other k-mers are shown in gray. K-mer ranks in the mCross approach were assigned based on raw scores. d,e An overview of the differentially enriched (d) and depleted (e) 5-mers (see “Methods”) in eCLIP analysis approaches relative to in vitro data. In each panel, k-mer groups are arranged from top to bottom as follows: k-mers that are differentially ranked in all approaches (“all approaches”), only in raw mCross (“unique mCross”), only in PEKA (“unique PEKA”) or only in “local” approach (“unique local,” where k-mers were split into 2 clusters based on their sequence). K-mer logo is shown for each k-mer group, with nucleotides scaled proportionally to their frequency at each position in the alignment. The left line plot shows the distribution of mean motif group coverage around tXn (red) or oXn (black) within the “protein-coding gene” region across the 57 eCLIP datasets that have orthogonal in vitro data. Area around the curve shows the standard deviation. The right line plot shows the distribution of motif group coverage around tXn (red) or oXn (black) within the “protein-coding gene” region for a representative RBP that ranked highly in both eCLIP and in vitro datasets for the corresponding motif group. f, g Boxplots showing distribution of mean motif ranks across RBPs in methods used to analyze eCLIP and in vitro data for the enriched (f) and depleted (g) motif groups (shown in panels d, e and in Additional file 1: Fig. S5a,b). For each motif group, we compared the distribution of mean ranks in CLIP analysis approaches with in vitro data to assess the significance of difference with Welch’s t-test and the comparisons are marked when significant

Back to article page