A Phylogeny-aware GWAS Framework to Correct for Heritable Pathogen Effects on Infectious Disease Traits. (3rd August 2022)

• Record Type: Journal Article
• Authors: Nadeau, Sarah; Thorball, Christian W; Kouyos, Roger; Günthard, Huldrych F; Böni, Jürg; Yerly, Sabine; Perreau, Matthieu; Klimkait, Thomas; Rauch, Andri; Hirsch, Hans H; Cavassini, Matthias; Vernazza, Pietro; Bernasconi, Enos; Fellay, Jacques; Mitov, Venelin; Stadler, Tanja
• Editors: Crandall, Keith
• Is Part Of: Molecular biology and evolution. Volume 39:Number 8(2022)
• Publication Date: 2022-08-03
• Subjects: genome-wide association study -- infectious disease -- phylogenetic mixed model -- heritability; Molecular biology -- Periodicals; Molecular evolution -- Periodicals; Evolution, Molecular -- Periodicals; Molecular Biology -- Periodicals; 572.8
• DOI: 10.1093/molbev/msac163
• Languages: English
• ISSNs: 0737-4038

Abstract

Abstract: Infectious diseases are particularly challenging for genome-wide association studies (GWAS) because genetic effects from two organisms (pathogen and host) can influence a trait. Traditional GWAS assume individual samples are independent observations. However, pathogen effects on a trait can be heritable from donor to recipient in transmission chains. Thus, residuals in GWAS association tests for host genetic effects may not be independent due to shared pathogen ancestry. We propose a new method to estimate and remove heritable pathogen effects on a trait based on the pathogen phylogeny prior to host GWAS, thus restoring independence of samples. In simulations, we show this additional step can increase GWAS power to detect truly associated host variants when pathogen effects are highly heritable, with strong phylogenetic correlations. We applied our framework to data from two different host–pathogen systems, HIV in humans and X. arboricola in A. thaliana . In both systems, the heritability and thus phylogenetic correlations turn out to be low enough such that qualitative results of GWAS do not change when accounting for the pathogen shared ancestry through a correction step. This means that previous GWAS results applied to these two systems should not be biased due to shared pathogen ancestry. In summary, our framework provides additional information on the evolutionary dynamics of traits in pathogen populations and may improve GWAS if pathogen effects are highly phylogenetically correlated amongst individuals in a cohort.