Joint ancestry and association testing in admixed individuals

Area of Science:

• Genetics
• Statistical Genomics
• Population Genetics

Background:

• Admixed individuals allow for testing both ancestry and genotype effects.
• Existing methods for joint analysis of admixture and association mapping have limitations.

Purpose of the Study:

• To introduce BMIX, a novel joint test combining admixture and association mapping.
• To demonstrate BMIX's superior power compared to existing methods.
• To identify genetic loci influencing fasting plasma glucose in admixed African Americans.

Main Methods:

• Developed BMIX, a joint test integrating local ancestry and stratified regression.
• Utilized generalized linear models for flexible phenotype analysis.
• Employed autocorrelation for empirical estimation of testing burdens.
• Incorporated admixture mapping posterior probabilities as priors for association mapping.

Main Results:

• Simulations show BMIX is orders of magnitude more powerful than the MIX score.
• Analysis of fasting plasma glucose in admixed African Americans identified significant loci at 1q24 and 6q26.
• The 1q24 signal was resolved, revealing potential regulatory elements near FAM78B linked to type 2 diabetes.

Conclusions:

• BMIX offers a substantial power increase for genetic discovery in admixed populations.
• The identified loci provide insights into the genetic architecture of fasting plasma glucose.
• This method can advance understanding of complex trait genetics in diverse populations.