Gene-based Rare Variant Association Tests for Ancestry-matched Case-control Data

摘要

With an increasingly large amount of human sequencing data available, analysis incorporating external controls becomes a popular and cost-effective approach to boost statistical power in disease association studies. To prevent spurious association due to population stratification, it is important to carefully match the ancestry backgrounds of cases and external controls. However, popular rare variant association tests based on logistic regression models, including the burden test, sequence kernel association test (SKAT) and the mixed effects score test (MiST), which is a hybrid version of burden and SKAT, are conservative for matched case-control samples when all matched strata have the same case-control ratio and might become anti-conservative when case-control ratio varies across strata. To account for the matching structure, we propose gene-based tests based on a conditional logistic regression (CLR) model, namely CLR-burden, CLR-SKAT, and CLR-MiST. We show that the CLR model coupled with ancestry matching is a general approach to control for population stratification. Through extensive simulations of population stratification and matching schemes, we demonstrate that both CLR-burden and CLR-SKAT are more powerful than standard burden test and SKAT respectively in ancestry-matched data while robustly controlling the type 1 error, and CLR-MiST is most powerful under a wide range of scenarios. Furthermore, because CLR-based tests allow for different case-control ratios across strata, a full-matching scheme can be employed to fully utilize available cases and controls to accelerate the discovery of disease genes.