Many candidate gene studies use population-based association methods—that is, methods that compare the genes of groups of people. For example, in alcoholism research such analyses would involve two samples: a group of alcoholic patients and a control group of nonalcoholic people. Ideally, the two groups would be matched with respect to numerous factors (e.g., age and ethnicity) so that they differ only in disease status. The investigators would then compare the frequencies of various alleles of a marker (e.g., an SNP) within or near the candidate gene (see figure 3A). Evidence of differences in allele frequencies between the two groups is typically interpreted as evidence that the candidate gene contributes to alcoholism susceptibility.2
