Home

Goal : Editing Evaluation of functional annotation‐informed and ancestry‐specific polygenic risk scores for ischemic stroke

Ischemic stroke (IS) is a leading cause of morbidity and mortality worldwide, and predicting incident events remains challenging. Polygenic Risk Scores (PRS) aggregate common genetic variants for risk prediction, but performance varies ancestries and remains modest, particulary in non-European populations.

Overview of study design, PRS development and optimization, and validation

Ancestry-specific (EA and AA) summary statistics from the GIGASTROKE consortium were used to construct PRS using three methods: PRS-CS with the HM3 reference panel (PRS-CS-HM3), PRS-CS with the TagIt reference panel (PRS-CS-TagIt), and SBayesRC. For PRS-CS, global shrinkage parameters (phi, ɸ) of 1.0E-06, 1.0E-04, 1.0E-02, and 1 were evaluated, and the parameter that maximized the liability-scale R2, had the strongest parameter estimate (odds ratio), and the largest PRS AUC, was selected as the optimized PRS (ɸ =1.0E-02). For SBayesRC, the optimal PRS is output directly. Optimized PRS were evaluated in three independent cohorts (no participant overlaps with the discovery GWAS used for PRS construction or with the REGARDS optimization cohort). Abbreviations: AA-African ancestry; EA-European ancestry; IS-ischemic stroke; EHR-electronic health records; AUC-area under the curve.