Family-based samples can play an important role in genetic association studies.

Over the past 2 decades, DNA samples from thousands of families have been collected and genotyped for linkage studies of common complex diseases, such as type 2 diabetes, asthma, and prostate cancer. Unfortunately, little success has been achieved in identifying genetic susceptibility risk factors through these considerable efforts. However, significant success in identifying common disease risk-associated variants has been recently achieved from genome-wide association studies using unrelated case-control samples. These genome-wide association studies are typically done using population-based cases and controls that are ascertained irrespective of their family history for the disease of interest. Few genetic association studies have taken full advantage of the considerable resources that are available from the linkage-based family collections despite evidence showing cases that have a positive family history of disease are more likely to carry common genetic variants associated with disease susceptibility. Herein, we argue that population stratification is still a concern in case-control genetic association studies, despite the development of analytic methods designed to account for this source of confounding, for a subset of single nucleotide polymorphisms in the genome, most notably those single nucleotide polymorphisms in regions involved with natural selection. We note that current analytic approaches designed to address the issue of population stratification in case-control studies cannot definitively distinguish between true and false associations, and we argue that family-based samples can still serve an invaluable role in following up findings from case-control studies.