Table 2 presents the results from the meta-analysis of COGA and SAGE samples. An intronic variant rs4758317 in LMO1 showed evidence for association (p=7.2×10−7) with evidence coming from both SAGE (beta = 0.08, p=6.0×10−6) and COGA (beta = 0.05, p=2.9×10−2) (Table 2, supplementary figure 4). Similarly, we observed a suggestive association (p= 4.1×10−6) of maxdrinks with an intronic variant in PLCL1 (rs67031482 ), driven by both COGA (beta = −0.07, p=1.1 × 10−2) and SAGE (beta = −0.07, p=1.2×10−4) datasets. In populations of European descent, this SNP is in a region of linkage disequilibrium of ~ 200 kb on chromosome 2q that spans the entire PLCL1 gene (Supplementary figure 5). Table 2 presents the results from the meta-analysis of COGA and SAGE samples. For meta-analysis we had > 90% power to detect a genome-wide association given a SNP that explains 1% of the variance for maxdrinks.
