|
Abstract:
|
GWAS and sequencing studies in ancestrally diverse populations have recently become more common due to increased interest in both identifying novel, population specific variants that underlie phenotypic diversity and generalizing associations across populations. Recently admixed populations, such U.S. Hispanics, pose special challenges for genetic association mapping of complex traits due to heterogeneous genetic backgrounds and allele frequency differentiation among populations that varies greatly across the genome; which can result in spurious associations if not properly accounted for in the association analysis. We consider the problem of association testing between a continuous trait and multiple genetic variants in a gene or genomic region in samples from ancestrally diverse populations. We propose a kernel-based association method that appropriately accounts for admixture and familial relatedness by using both fixed and random effects in a linear mixed model. We demonstrate the utility of our approach in simulation studies and a genome-wide gene-based association analysis of hematological traits in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL).
|
