Raw reads were aligned to human genome 19 (hg19) using STAR aligner (version 2.5.3.a)16. We used QC tools RSeQC (http://code.google.com/p/rseqc/) and Picard (https://broadinstitute.github.io/picard/) to evaluate RNA sequence quality including the %GC, %duplicates, gene body coverage, unsupervised clustering, and the library complexity. We used the Picard “MarkDuplicates” option to flag and remove duplicate reads. Gene quantification was performed with featureCounts (SUBREAD package; release 1.6.0)17 using Gencode annotations (Release 19 (GRCh37.p13)).
