A phylogeny-aware GWAS framework to correct for heritable pathogen effects on infectious disease traits

Abstract

Infectious diseases are a unique challenge for genome-wide association studies (GWAS) because pathogen, host, and environmental factors can all affect disease traits. Previous GWAS have successfully identified several human genetic variants associated with HIV-1 set point viral load (spVL), among other important infectious disease traits. However, these GWAS do not account for potentially confounding or extraneous pathogen effects that are heritable from donor to recipient in transmission chains. We propose a new method to consider the full genome of each patients infecting pathogen strain, remove strain-specific effects on a trait based on the pathogen phylogeny, and thus better estimate the effect of human genetic variants on infectious disease traits. In simulations, we show our method can increase GWAS power to detect truly associated host variants when pathogen effects are highly heritable, with strong phylogenetic correlations. When we apply our method to HIV-1 subtype B data from the Swiss HIV Cohort Study, we recover slightly weaker but qualitatively similar signals of association between spVL and human genetic variants in the CCR5 and major histocompatibility complex (MHC) gene regions compared to standard GWAS. Our simulation study confirms that based on the estimated heritability and selection parameters for HIV-1 subtype B spVL, standard GWAS are robust to pathogen effects. Our framework may improve GWAS for other diseases if pathogen effects are even more phylogenetically correlated amongst individuals in a cohort.