We identified pathways that are differentially altered in OUD using gene set enrichment analyses with the molecular signatures database55 and the fgsea and msigdbr R-packages141. We included the hallmark gene pathways, curated gene sets from BioCarta, KEGG, Canonical Pathways, Reactome, and WikiPathways, and ontology gene sets. We also downloaded and included curated pathways from SynGO in enrichment analyses142. We report the full list of enriched pathways corrected for multiple hypotheses within each set of hypotheses in Supplementary Data 1–S7. We clustered redundant/related pathways using custom R-scripts and the igraph R-package creating networks of pathways connected by overlapping genes143,144 (igraph.org). We visualized the clustered pathways using igraph functions and report both clustered and singleton pathways alongside each figure in Supplementar Data 1–S7, S9, and S11. For the interaction score for calculating pathway-enrichments between female- and male-biased pathway analyses, we used the log2(fold-change) and the p-values from the OUD v. UC effect within differential analysis of males or females to calculate an interaction score: sign(log2FCF) * (-log10(p-valueF)) - sign(log2FCM) *(-log10(p-valueM)). We overlapped enriched pathways using this calculation vs. the pathways enriched in the standard OUD v. UC in females or males to further support the interpretation of female- or male- biased pathways.
