Enrichment analysis of nominally significant GWAS variants (p < 0.05) that were also eQTLs (p < 5 × 10−8) for genes in the thistle2 module showed significant under enrichment in the two-tail Fisher test (Table 2). The under-enrichment remained significant even after 100,000 permutations. This might be due to the small size of this module (N = 72 genes). Although some of the differntialy expressed genes were significant in the gene-based tests performed in UKBB-AC and PGC-AD datasets using MAGMA (CHRNA6, CHRNA2, MN1 and HAPLN1). In the calcium signaling network (Fig. 3c), a few genes that were not part of the thistle2 module, but were essential to create network connections, were also found to be significant (3.4 × 10−2 ≤ P ≥ 4.8 × 10−2) in the gene-based tests (circled in red; Fig. 4c). This suggests possible gene–environment (alcohol exposure) interactions in the etiology of alcohol dependence. This also reinforces the need for multi-omics data to understand a complex disorder like alcoholism. eQTLs for genes in the brown4 module (N = 726 genes) were significantly enriched for GWAS signals (P
