Refining the Use of Polygenic Risk Scores for Alzheimer's Disease in Diverse and Founder Populations

PROJECT SUMMARY/ABSTRACT Alzheimer’s disease (AD) is the sixth leading cause of death in the United States, affecting about 5.8 million Americans currently. Unlike other leading causes of death, deaths attributable to AD have increased immensely since 2000. As a result of population aging, the burden of AD and other chronic diseases on the healthcare

system will continue to increase in coming years. Risk for AD is multifactorial, including genetic and environmental components. Twin studies have revealed that the heritability for late-onset AD is as high as 70% but can depend on the population and the environment in which a population resides. Because of this, it is

reasonable to investigate the use of genetic risk to identify currently unaffected individuals who are at high risk of AD. One prominent means of evaluating person-level risk is through the use of polygenic risk scores (PRSs). A PRS captures the genetic risk of AD by incorporating the effects of multiple, often hundreds or thousands, of

risk and protective loci. Generally, a PRS is simply the sum of the given weight, typically derived from genome- wide association study summary statistics, for a given single nucleotide polymorphism (SNP) multiplied by the number of copies of the SNP. PRSs have been used to inform intervention, screening, and life planning for other

chronic diseases. Despite these successes, the PRS approach has several limitations, including lack of use in non-European and highly related populations, such as founder populations. Using data from the National Institute on Aging’s Alzheimer’s Disease Sequencing Project (ADSP) and the Collaborative Amish Aging & Memory

Project (CAAMP), this work will examine the possibility of using PRSs across diverse and founder populations and aim to develop a pathway-based genetic risk score for analyzing genetic risk of AD across populations. I have identified 58,925 individuals with whole genome sequencing data in the ADSP data, including 8,856

individuals with African ancestry, 4,000 Asian ancestry individuals, 9,835 Hispanic individuals. This data source represents a unique opportunity to analyze sequencing data in a large and diverse population. The CAAMP data includes over 2,100 closely related Amish individuals. The Amish, a founder population, are descendants of

Swiss Anabaptist immigrants who immigrated to the United States in the eighteenth century. They live a relatively culturally isolated lifestyle and thus, are mostly genetically and environmentally homogeneous compared to a general European ancestry cohort. The Amish are further thought to be enriched for variation that is rare in the

general US population, allowing for detection of effects that may not otherwise be captured. Each of these data sources provides a unique benefit in addressing current issues with PRS approaches. My central hypothesis is that standard PRS methods must be adapted depending on both the ancestry of the target population and

relatedness of the overall sample. The proposal will refine the use of PRSs in both clinical and research settings for non-European and founder populations. A better understanding of these effects will help to ease the burden of health disparities as PRSs are increasingly used to inform relative risk.