Estimating heritability explained by local ancestry and evaluating stratification bias in admixture mapping from...

Area of Science:

• Population Genetics
• Statistical Genetics
• Genomics

Background:

• Heritability explained by local ancestry (h_γ²) is key to understanding complex traits in admixed populations.
• Estimating h_γ² can be biased by ancestral population structure.
• Existing methods struggle to correct for ancestral stratification.

Purpose of the Study:

• Introduce Heritability Estimation from Admixture Mapping Summary Statistics (HAMSTA).
• Develop a method to infer h_γ² while adjusting for ancestral stratification biases.
• Provide a calibrated family-wise error rate (FWER) for admixture mapping.

Main Methods:

• Utilize summary statistics from admixture mapping.
• Employ a HAMSTA-derived sampling scheme for FWER estimation.
• Apply the HAMSTA approach to quantitative phenotypes in African American individuals.

Main Results:

• HAMSTA estimates of h_γ² are approximately unbiased and robust to ancestral stratification.
• Demonstrated calibrated FWER of ~5% for admixture mapping under stratification.
• Observed h_γ² ranging from 0.0025 to 0.033 across 20 phenotypes.
• Found minimal inflation due to ancestral stratification (mean inflation factor ~0.99).

Conclusions:

• HAMSTA offers a fast and powerful method for genome-wide heritability estimation.
• The approach effectively evaluates biases in admixture mapping test statistics.
• HAMSTA improves the accuracy of genetic architecture studies in admixed populations.