Bayesian association mapping for quantitative traits in a mixture of two populations

Area of Science:

• Genetics
• Statistical Genetics
• Bioinformatics

Background:

• Population structure significantly confounds genetic association studies.
• Accurate estimation of population stratification is crucial for identifying true genetic associations.
• Existing methods may not adequately address simultaneous mapping and structure estimation in admixed populations.

Purpose of the Study:

• To develop a novel Bayesian approach for simultaneous quantitative trait loci (QTL) mapping and population structure estimation.
• To analyze unrelated individuals from a mixture of two unknown populations without assuming admixture.
• To account for both population stratification and genetic heterogeneity.

Main Methods:

• Utilized Bayesian hierarchical modeling and Markov chain Monte Carlo (MCMC) estimation.
• Employed "grouping" markers with differing allele frequencies between populations to estimate structure.
• Treated the number and positions of QTL as random variables, obtaining posterior distributions.

Main Results:

• Successfully estimated and accounted for population structure concurrently with association mapping.
• The Bayesian model allowed for flexible estimation of QTL number and location.
• Candidate and grouping markers were selected based on preliminary SOLAR analysis results.

Conclusions:

• The proposed Bayesian method offers a robust framework for association mapping in structured populations.
• This approach effectively handles population stratification and genetic heterogeneity.
• The simultaneous estimation provides more accurate QTL identification in complex genetic studies.