RNA-seq analysis

FASTQ files were quality checked using FastQC (v0.11.4) and reads trimmed using Trim Galore! (v0.4.1) (Martin 2011). Paired-end reads were mapped to hg19 using STAR (v2.4.2) (Dobin et al. 2013). We mapped to hg19 for compatibility with previously published data sets. Blacklisted regions were removed, and uniquely mapped reads were used for processing. Therefore, our choice of hg19 over GRCh38 should not significantly affect results. PCR duplicates were removed using Picard-tools MarkDuplicates (v1.83). Mapped reads over exons were quantified using subread featureCounts (v1.6.2) (Liao et al. 2014). Statistical analysis was performed in R (R Core Team 2021) using the edgeR package (Robinson et al. 2010). DEGs were defined as FDR < 0.05. Enriched GO terms and reactome pathways were determined using PANTHER (v. 15) (Thomas et al. 2003). Published expression data was sourced from NCBI Gene Expression Omnibus (GEO; https://www.ncbi.nlm.nih.gov/geo/) accessions {"type":"entrez-geo","attrs":{"text":"GSE85988","term_id":"85988"}}GSE85988, {"type":"entrez-geo","attrs":{"text":"GSE83671","term_id":"83671"}}GSE83671, and {"type":"entrez-geo","attrs":{"text":"GSE139437","term_id":"139437"}}GSE139437 (Supplemental Table S3).