and others 2011a; Luo and others 2006]. Conversely, ALDH2 rs671 is very rare in non-Asian populations (MAF < 1% in accordance with 1,000 Genomes Project Phase 3 data; [1000 Genomes Project Consortium and others 2015]) and no informative analysis can be conducted in European-ancestry subjects. In African and Native American populations, ADH1B rs2066702 (Arg370Cys) showed protective effect similar to that observed for ADH1B rs1229984 [Ehlers and others 2012; McCarthy and others 2010]. Our meta-analysis of candidate gene studies confirmed the strong involvement of ADH1B rs1229984 in AD risk, and also for alcohol abuse and alcohol-induced diseases in multiple ethnic populations [Li and others 2011a]. The same result was also confirmed by an independent meta-analysis of genetic studies of alcohol drinking behaviors [Buhler and others 2015]. Genomic studies have demonstrated that candidate gene analysis and their meta-analyses can produce false positive results due to publication bias [Sullivan 2013]. Although the relationship between ADH1B and alcohol-related traits is on much firmer biological and statistical ground than most other such associations [Bierut and others 2012], it was also important to demonstrate the relationship in a genome-wide context. GWAS (hypothesis-free investigations) can be powerful in investigation of the genetic architecture of complex traits, such
