In general, there are two major types of designs that are commonly used in association research: population-based and family-based studies. As the most common population-based approach, the case-control setup compares an unrelated healthy control group and affected case group. The genotyped SNPs are investigated to identify the allele frequency differences between these two groups. The study then determines whether the SNPs are associated with the genetic trait or disease based on the statistical significance of the differences. The independent samples are typically easier to obtain in a case-control study than family samples. However, many case-control samples select independent cases from existing family data that were originally used in linkage analysis. Because cases can be over-sampled from groups with higher disease prevalence, the differences of allele frequencies in an admixture of ethnic groups may produce spurious associations. Therefore, although case-control studies have shown advantages in identifying association between the disease susceptibility and markers in a candidate gene, the results may reflect type I errors (false-positive) due to unaccounted confounding factors [9]–[11] such as population stratification [12]–[16].
