Systematic Profiling of Poly(A)+ Transcripts Modulated by Core 3’ End Processing and Splicing Factors Reveals Regulatory Rules of Alternative Cleavage and Polyadenylation
Fig 5
Detailed analysis of five C/P factors.
(A) Schematic of the experimental design. Proliferating C2C12 cells were harvested 32 hr after knockdown (KD), and both total and nuclear RNAs were extracted for 3’READS analysis. (B) Western Blot analysis of protein expression after 32 hr of KD. The percent of expression in KD cells compared to siCtrl cells for each KD is indicated. (C) Normalized number of genes with regulated 3’UTR-APA in each sample as examined by GAAP. See Fig 2D for details of the plot. Both total and nuclear RNA data are shown. (D) Cluster analysis of 3’UTR-APA regulation by the five factors. RED scores using the top two most abundant APA isoforms (based on the number of poly(A) site-supporting reads in all samples) of each gene were used for this analysis. Only pA isoforms with read number ≥5 in all samples were used. A RED score is difference in relative expression of distal pA isoform vs. proximal pA isoform between KD and siCtrl cells, as illustrated in Fig 2A. RED scores are represented in a heatmap using the color scheme shown in the graph. Positive and negative RED scores indicate lengthened and shortened 3’UTRs, respectively. RED scores for APA events were set to 0 when q-value > 0.05 (SAAP). Pearson correlation was used as metric for hierarchical clustering. (E) Venn diagrams comparing genes with significant 3’UTR-APA regulations by the five factors using total RNA (left) or nuclear RNA (right) data. (F) Relationship between extent of 3’UTR-APA regulation and aUTR size. Genes were divided into five bins based on the aUTR size (distance between the pAs of top two most abundant APA isoforms). The aUTR size range for each bin is shown in the graph. The extent of 3’UTR-APA regulation is represented by average RED scores, based on the data shown in (D). Only genes with ≥20 PASS reads (proximal and distal pAs combined) in both KD and siCtrl samples were used for RED calculation. RED scores of genes in bin #1 were compared with those in bin #5 by the Wilcoxon rank sum test for each sample, and p-values are shown.