The outcome of alternative splicing was quantified as percent-spliced-in (PSI, Ψ) computed by replicate Multivariate Analysis of Transcript Splicing (rMATS, version 4.0.2) [28]. We adopted the Gencode annotation (GTF, GRCh38.p13, hg38) to determine the exon skipping events. GTF annotations were converted from hg38 to hg19 using LiftOver (version 1.20.0) [29] to be consistent with the genotype data. Using the RNA-seq alignment files, junction reads supporting the inclusion or exclusion isoforms of all the annotated exon skipping events were counted. PSI was calculated as,\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Psi = \left( {\frac{I}{{L_I}}} \right)\bigg/\left( {\frac{I}{{L_I}} + \frac{S}{{L_S}}} \right)$$\end{document}Ψ=ILIILI+SLSwhere I and S are the junction read counts supporting the inclusion and exclusion (skipped exon) isoforms, respectively. LI and LS are the effective lengths of the exon inclusion and exclusion isoforms, respectively, which were automatically calculated by rMATS based on the annotation GTF.
