After correcting for technical effects and unmeasured confounding variables, we found that thousands of gene expression traits were significantly associated with demographic variables. At a 5% false discovery rate (FDR), 769, 336, and 3,110 genes were significantly associated with ancestry, sex, and age, respectively within the UC livers. Genes significantly affected by sex or age (FDR<5%, Figure 1, Figure S2, examples displayed in Figure S3) have a marked enrichment for small p-values in both replication samples (Figure 1A, 1D). To lessen the influence of differential statistical power among the three studies (n = 206, 60, 266), we defined ‘replication’ as having a nominally significant p-value in the independent sample (p-value<0.05) and having a concordant effect direction (i.e., is YFG more highly expressed in males or females?). 29.9% and 32.1% of genes significantly affected by sex (UC sex t-test FDR<5%) replicated in the UW and Merck studies, respectively (Figure 1B). At more stringent thresholds, validation rates exceeded 80%, albeit with fewer included genes (Figure 1B). We also note that the sex-associated gene set was strongly enriched for genes on the X
