Population structure in genetic studies: Confounding factors and mixed models

Area of Science:

• Genetics
• Statistical genomics
• Bioinformatics

Background:

• Genome-wide association studies (GWAS) aim to find genetic variants linked to diseases.
• Population structure and relatedness can introduce false positives in genetic association studies.
• Mixed models have emerged as a computational and statistical approach to address these challenges.

Purpose of the Study:

• To characterize the problem of population structure in GWAS.
• To explain how population structure can lead to false positive associations.
• To motivate the use of mixed models for correcting population structure in association studies.

Main Methods:

• Review of existing literature on GWAS and mixed models.
• Characterization of population structure issues using laboratory mouse strains as an example.
• Description of how population structure impacts association testing.

Main Results:

• Population structure can significantly inflate false positive rates in GWAS.
• Uncorrected relatedness and population stratification are major confounding factors.
• Mixed models offer a robust framework for correcting these effects.

Conclusions:

• Accurate control of population structure is critical for reliable GWAS results.
• Mixed models represent a key advancement in mitigating confounding factors in genetic association studies.
• Further development of GWAS techniques is needed to overcome computational and statistical challenges.