Functional Variants Identify Sex Specific Genes and Pathways in Alzheimer’s Disease

Abstract

AbstractBackgroundAlzheimer’s Disease (AD) incidence is almost double in women, suggesting sex‐specific risk genes for AD remain to be uncovered.MethodTo identify AD risk genes in smaller cohorts of only male or only female subjects we designed a machine learning case‐control study of genome variant differences that included a functional importance score as a feature. The method robustly finds genes in smaller cohorts, enabling us to analyze men and women separately.ResultWe found AD‐associated risk genes men‐only, women‐only, or combined subjects. These genes were consistent with a) prior AD GWAS; b) expression in AD patients brain regions; c) transcriptome analyses; d) and in vivo experiments in fly models of Amyloid beta and Tau neurodegeneration. Notably, the sex‐specific genes implicate pathways that are differentially altered in each sex – a stress response pathway unique to AD affected men and cell‐cycle/DNA quality control pathways unique to AD affected women.ConclusionA machine learning combined with an evolution‐based training feature for variant impact identified common as well as sex‐specific AD‐associated genes and mechanisms of potential value for pre‐symptomatic risk assessment and drug targeting.