‘genome-wide’ levels of statistical significance (P-value = 8.9 × 10−8) (23). The above results cannot be directly combined, mainly because of incomplete reporting (P-values but no effect estimate) and heterogeneity of populations and alcohol phenotypes presented. However, consistency of the direction of association across all studies as well as P-values from the largest studies offer reassurance and strengthen the evidence supporting a role of this SNP in influencing alcohol consumption (sample results from the above cited studies are presented in Supplementary Material, Table S6 available online).
