A subset of introns shows increased levels of retention (IR) upon Salsa depletion. (A) Volcano plot for beta distribution-based differential intron retention analysis, with differences in inclusion levels between Salsa-depleted (salsa RNAi-1) and control samples (mCherry RNAi) (ΔPIRbeta = PIRsalsa RNAi − PIRcontrol) in the x-axis and probability of differential retention [−log10(1+ − Pdiff)] in the y-axis, with 1+ = 1.001 used to avoid infinite values when Pdiff = 1. Differentially retained introns (Affected, Affected low PIR, and Less Retained), identified by Pdiff ≥ 0.9, are highlighted in red, purple and yellow, respectively. Events with an absolute ΔPIRbeta greater than 0.15 are labeled with gene name and the transcript's intron position. Control, Included, and Skipped introns are also highlighted for comparison, as well as gurken intron retention events. Differential retention of the first intron of gurken is shown in black as this intron was filtered out of the global differential intron retention analysis due to insufficient coverage in one of the samples (see Materials and Methods). (B) Diagram of simplified gurken (grk) gene structure with exons (blue) and introns (gray) and density plots (smoothed histograms) reflecting points emitted from beta distributions used to model gurken intron 1 (left panel), 2 (middle), and 3 (right) retention for each sample and PIRbeta estimates for Salsa-depleted and control conditions (dashed vertical lines). Rug plots below the density curves reflect PIR values per sample. Event identifiers and genomic coordinates of the respective alternative sequences are part of the VAST-DB dm6 annotation (Tapial et al. 2017).
