Abstract
Over 75 Alzheimer's disease (AD) and dementia-associated variants have been identified through genome-wide association studies, but the utility of polygenic risk scores (PRS) for predicting AD and dementia in diverse and admixed populations remains unclear. We compared how PRS approaches differing in p -value thresholds, variant weights, and source ancestry perform in predicting dementia in 6,338 African American, Chinese, Hispanic, and White individuals from the Multi-Ethnic Study of Atherosclerosis. We tested clumping and thresholding (C+T) methods with varying parameters against Bayesian approaches (PRS-CS, PRS-CSx). We compared the ability of each method to predict incident dementia in all participants and in groups stratified by self-reported race/ethnicity. We additionally analyzed performance across groups stratified by estimated proportion of non-Finnish European (NFE)-like ancestry. Including more variants does not improve performance. The PRS based on C+T method with only 15 SNPs is more strongly associated with dementia (HR (5e-08) = 1.21, 95% CI: 1.11-1.31) than PRS derived from Bayesian models that include >800,000 SNPs (HR (CSx) = 1.13, 95% CI: 1.04-1.23), even in populations genetically dissimilar from the source data (HR (lowNFE) _ (5e-08) = 1.26, 95% CI: 1.08-1.47; HR (lowNFE) _ (CSx) = 1.13, 95% CI: 0.96-1.32). More selective PRS models using fewer SNPs may offer better AD prediction across diverse populations.