We corrected for multiple hypotheses testing in this study using the false discovery rate (FDR) less than alpha = 0.05. In different expression analyses, we perform FDR correction across all cell types tested. Similarly for gene set enrichment analysis and AUCell pseudobulk analyses, we perform FDR correction across all pathways and differentially expressed genes across cell types. We computed the FDR correction with the swfdr R package which increases power by leveraging Storey’s q-value and modeling the relationship of null P-values and independent variables such as the average logCPM or the number of genes in a pathway149,150. In differential gene expression analyses, we calculated the FDR within each cell type comparison (adj.P.Val.Within) and between all cell type comparisons (adj.P.Val.Between). We report the more conservative correction between cell types in the main text and figures, adj.P.Val.Between. Measured metrics (gene expression counts, quality control metrics) or derived metrics (AUCell gene activity scores) are repeated measurements of the same cohort of human individuals or rhesus macaque subjects. Metrics that are differentially expressed across multiple cell type conditions are reported and summarized at the
