Testing for population subdivision and association in four case-control studies

Area of Science:

• Population Genetics
• Genetic Epidemiology

Background:

• Replication of disease-marker associations is a challenge in genetic studies.
• Population structure is a suspected cause of unreplicated findings.
• Few case-control studies have been rigorously evaluated for population stratification.

Purpose of the Study:

• To assess the presence of detectable population subdivision in moderate-sized case-control samples.
• To investigate the impact of population structure on the reliability of genetic association studies.
• To provide insights into factors affecting the replication of genetic association findings.

Main Methods:

• Examined four case-control samples (N=3,472) including U.S. whites and African Americans with hypertension, and U.S. whites and Polish whites with type 2 diabetes.
• Utilized sum of case-control allele frequency chi-squared statistics for 9 STR and 35 SNP markers to test for population structure.
• Replicated the PPARg Pro12Ala polymorphism association in type 2 diabetes populations.

Main Results:

• Weak evidence of population structure was detected only in the African American sample.
• Refining the African American sample to include only individuals with U.S.-born parents and grandparents eliminated the detected stratification.
• The study suggests that carefully matched, moderate-sized case-control samples in cosmopolitan populations are unlikely to have significant population structure leading to false-positive associations.

Conclusions:

• Population structure is unlikely to be a major driver of unreplicated genetic associations in well-matched, moderate-sized case-control studies from cosmopolitan populations.
• Factors such as differences in study power due to sample size and allele frequencies may play a more significant role in the replication problem.
• Careful sample matching and characterization are crucial for robust genetic association studies.