Skip to main content


Fig. 5 | Genome Biology

Fig. 5

From: omniCLIP: probabilistic identification of protein-RNA interactions from CLIP-seq data

Fig. 5

Binding preferences of CNBP. a Metaplot depicting the average Z-score transformed binned coverage across all genes (transcript with highest RSEM isoform percentage selected) with omniCLIP peak. Median 5’UTR (8%), CDS (78%) and 3’UTR (14%) proportions were extracted from all expressed genes in Drosophila S2 cells (TPM >0) from regular total RNA-seq experiments. Shades around solid lines indicate the standard error. b omniCLIP peak annotation grouped by strength into 10 peak SiteScore bins. (Left) Simplified annotation categories, to enable comparison to the expected annotation distribution. Here, 5’UTR contains the start codon and 3’UTR the stop codon, respectively. The expected peak annotation distribution was calculated according to the feature distribution shown in (a), for all peaks that are annotated as mature transcripts. Peaks classified as ‘other’ were ignored. (Right) Peak annotation categories grouped by peak score. Peaks annotated with start or stop codon do overlap such features. c CNBP motif calculated using HOMER2 for all peaks annotated to mature transcripts (n=29556), relative to 10x dinucleotide shuffled background sequences. d Recovery of the CNBP motif and shuffled PWM relative to peak summit of all peaks used (n = 29,433). PWM match required 80% similarity. Indicated percentages reflect peak sequences with motif hit. The next highest recovered random PWMs are variants of the identified motif

Back to article page