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

Area of Science:

• Population Genetics
• Genomic Epidemiology
• Statistical Genetics

Background:

• Estimating heritability from local ancestry in admixed populations is vital for understanding complex trait genetics.
• Existing methods for heritability estimation can be biased by ancestral population structure.

Approach:

• We introduce Heritability estimation from Admixture Mapping Summary STAtistics (HAMSTA), a new method using admixture mapping summary statistics.
• HAMSTA infers local ancestry heritability while correcting for ancestral stratification biases.
• Simulations confirm HAMSTA provides unbiased estimates robust to stratification.

Key Points:

• HAMSTA-derived sampling schemes ensure calibrated family-wise error rates (~5%) for admixture mapping under stratification.
• Applied to 20 quantitative phenotypes in African Americans (PAGE study), HAMSTA revealed heritability estimates ranging from 0.062 to 0.85.
• Analysis showed minimal inflation (mean factor 0.99) in admixture mapping studies due to ancestral stratification.

Conclusions:

• HAMSTA offers a fast and powerful tool for genome-wide heritability estimation in admixed populations.
• The method effectively evaluates biases in admixture mapping test statistics.
• Findings suggest current admixture mapping studies have low inflation from ancestral stratification.