Fig. 6

RCRUNCH results for all ENCODE eCLIP data currently available. a Cumulative distribution of the number of significant binding sites detected per experiment (FDR threshold = 0.1, in black). Cumulative distributions are also shown separately for samples corresponding to proteins with a known binding motif (gray) and to proteins for which no known motif is available in ATtRACT, but one was found by RCRUNCH (blue). b Venn diagram summarizing the motif inference in the identified peaks. We distinguished four categories of proteins: for which (1) no motif is known and also no enriched motif was identified in this study (gray), (2) a de novo motif was found for a protein for which no motif is given in ATtRACT (blue), (3) a de novo motif was found for a protein with a known motif in ATtRACT (coral), and (4) a motif is known, but none was identified de novo (sand color). c Heatmap of mean peak agreement across RBPs. Only RBPs for which an enriched motif (see “Methods”) was found are included. The agreement is calculated as the Jaccard index of the nucleotides (nts) in the peaks, where the intersection of two sets of peaks is the number of nts covered in both sets, while the union is the number of nts covered in at least one of the two sets. The color range is capped at a similarity of 0.3 to make the clusters more easily distinguishable. The top peaks are taken according to the FDR threshold (0.1), extending by 20 nts upstream and downstream from the crosslink site. Since there are multiple replicates per RBP, the mean of pairwise Jaccard indices over all combinations of sample pairs are used here. The colors on the left indicate the relative frequency of each nucleotide type averaged over all positions of the PWM. d Polar projection of the enrichment of de novo motifs inferred for individual RBPs from all peaks with FDR > 0.1 extracted from the ENCODE samples. Only RBPs for which an enriched motif (see “Methods”) was found are included. The color of the bars indicate whether the respective RBP already has a known motif in ATtRACT (coral) or not (blue)