The previously mentioned candidate gene studies are hypothesis-driven and therefore only analyze genes with known connections to the phenotype of interest. This inherent bias can result in many relevant genes being overlooked. An unbiased method for identifying genetic variants associated with a phenotype is the genome-wide association study (GWAS), in which statistical analyses are performed on a large number of polymorphisms across the entire genome. One of the first GWAS of OUD compared the frequencies of 10,000 SNPs between 104 heroin dependent patients of European descent and 101 controls [36]. No significant associations were found after multiple testing correction, most likely due to the small cohort size and lack of statistical power. The same group published a larger study containing an analysis of 100,000 variants in 325 methadone-maintained heroin addicts and 250 controls [37]. This newer study included subjects of both African (case n = 125, control n = 100) and European (case n = 200, control n = 150) ancestry. A single intergenic variant (rs10494334) was significant in the patients of European descent after multiple testing correction (p =
