Given the prevailing skepticism surrounding candidate gene research, the burden of proof for the selection of a candidate gene is high. Such a rationale should be convincingly articulated in a manner specific to the phenotype and environment under study. The crux of the argument for selection of a particular gene to study lies in the statistical priors for the gene, i.e., based on prior evidence and the quality of the source of that evidence, how likely is it that this is a robust candidate. There is nothing inherently wrong with studying candidate genes, though the very idea of “candidate gene research” has fallen out of favor due to the historical issues with studying hypothesized biological candidates that have not held up in more systematic well-powered studies, as reviewed above. There are notable cases where hypothesized biological candidates have shown robust associations with outcome. For instance, rs1229984 in alcohol dehydrogenase (ADH1B) gene was one of the earliest candidate gene variants proposed in the etiology of alcoholism, particularly in Asians (A Agrawal & Bierut, 2012)5. The variant, which is rather rare in
