RNA-seq Filtering, Processing, and Differential Expression (DE)

High-throughput sequencing (HTSeq) mRNA data were obtained for the 78 ACC patients with steroid phenotype characterization using the R package TCGAbiolinks v 2.18.0 (35–37)⁠ and SummarizedExperiment v 1.20.0 (42)⁠. The 56,457 transcripts were filtered for protein-coding genes using EnsembleDb version 101 and notated using AnnotationHub R package v 2.22.0 (43)⁠. The genes with a sum of zero counts were excluded, resulting in the inclusion of 19,169 genes. The two steroid phenotype groups underwent DE analysis using the R package DESeq2 v 1.30.0 as previously described (44)⁠. P-values were calculated using Wald’s test and adjusted with Benjamini–Hochberg estimation. The gene count expression matrix was normalized by variance-stabilizing transformation using the DESeq2 package. Three genes had duplicated symbol names (PINX1, TMSB15B, and MATR3). To differentiate these three genes, we added an underscore after the first entry in the DE results.