These three examples illustrate how studies of specific responses to drugs in model organisms can be used to identify genes that can then be studied in humans. Because of the limited number of genome-wide significant results from GWAS for substance abuse, it is not surprising that the confirmation in humans has come from small, candidate gene-based studies which implicitly assume that these genes have relatively large effects on their corresponding human phenotypes. As mentioned before, if there were only a small number of genes that could influence these traits in humans, then searching for candidates in model organisms might identify a highly enriched set of genes that could be examined in humans. If, on the other hand, a large fraction of genes (say 10%) have the potential to influence the relevant human phenotypes, then enrichment using animal models will only be modestly helpful relative to randomly selecting genes to study in humans, or simply surveying all genes without prior hypotheses, as is done in human GWAS studies. If the latter situation is correct, then animal models will be most useful
