Assessment of mis-splicing was performed as described previously17. The alignment was performed with the STAR aligner (v.2.5.2a) against the hg19 human genome. Resulting bam files from the STAR alignment were indexed with samtools for use by rMATS, as described elsewhere18,19. Briefly, rMATS pipeline used RNA sequencing reads which were mapped to different splice variants to estimate the isoform proportion, and a hierarchical framework was employed to simultaneously account for estimation uncertainty in individual replicates and variability among replicates. This software package has intrinsic features that focus on the interreplicate variability to identify underrepresented or overrepresented samples which are accomplished by counting the level of inclusion and exclusion of an event. FDR cutoff <0.05 was used to define statistical significance20,21.
