Polygenic risk scores and Parkinson's disease in South Africa: Moving towards ancestry-informed disease prediction

Abstract

Parkinsons disease (PD) is a complex neurodegenerative disorder with environmental and genetic influences. Using genotyping array data of 661 South African PD cases and 737 controls, a polygenic risk score (PRS) analysis was conducted using PRSice-2. Summary statistics were used from two PD association studies as base datasets. The target dataset was split into training (70%; n=979) and validation (30%; n=419) cohorts. Various clumping window sizes, linkage disequilibrium thresholds, and p-value thresholds were tested to determine the optimal combination for risk prediction. Additionally, we investigated the variance explained by different combinations of covariates. Finally, top contributing variants were identified and cross-referenced with inferred local ancestry to assess ancestry-specific effects. Overall, modest predictive performance was observed (AUC: 0.6307-0.6311). Age at recruitment was the strongest individual predictor, while sex contributed the least. Local ancestry analysis indicated that the top contributing variants were not ancestry-specific risk variants. These findings provide the first evaluation of PRS performance for PD in a highly admixed South African cohort, underscoring the importance of including underrepresented populations in genetic risk prediction.