One of the earlier GWAS of AUD illustrates the kinds of manipulations that are being undertaken in the absence of genome-wide significant results to make sense of the wealth of data. Treutlein and colleagues [73] performed GWAS in 487 male inpatient alcoholics and 1358 controls and identified 121 SNPs with p values < 10−4. These SNPs were genotyped in a follow-up sample of 1024 male inpatient alcoholics and 996 controls. A total of 15 SNPs showed significant association with the same allele as in the GWAS; nine of these were located in genes including CDH13 and ADH1C, previously associated with AUD. In the combined analysis, two closely linked intergenic SNPs met criteria for genome-wide significance. Using max drinks as an excessive consumption phenotype, Kapoor and colleagues [74] performed one GWAS in over 2300 individuals from the COGA dataset and another GWAS in 2600 individuals derived from the Study of Addiction: Genes and Environment (SAGE) dataset. Neither study had genome-wide significant results but the investigators noted that among the top SNPs in each dataset, far more showed the same direction of
