Abstract Body (Do not enter title and authors here): Background: As genomic data becomes more readily available, there is growing demand for tools to translate genetic risk into actionable insights for patients. We introduce a novel clinical polygenic risk score (PRS) report for cardiometabolic disease, designed to provide trait-specific risk estimates across eight major conditions.
Methods: Using genotype and clinical data from 236,393 participants in the All of Us (AoU) Research Program, we curated and integrated publicly available PRS for coronary artery disease (CAD), atrial fibrillation (AF), type 2 diabetes (T2DM), venous thromboembolism (VTE), thoracic aortic aneurysm (TAA), hypertension (HTN), hypercholesterolemia (HC), and elevated lipoprotein(a) (LPA). For each trait, we applied PRSMix, an elastic-net modeling approach that combines multiple PRS to generate a single score with enhanced predictive accuracy and stability compared to individual component scores. External validation was performed in 53,306 genotyped individuals from the Mass General Brigham Biobank (MGBB), using logistic regression models adjusted for age, sex, and population structure. We incorporated these scores into a patient-facing, clinically orderable report that summarizes individual-level genetic risk.
Results: Of the 53,306 genotyped MGBB participants, 55.6% are female, mean enrollment age is 53 ± 17 years. Each trait-specific PRSMix demonstrated robust performance, with improved predictive accuracy over individual PRS. Trait-specific risk estimates for individuals in high risk bins vs. middle quintile were: CAD (3.6 [3.3–4.0]), T2DM (3.1 [2.9–3.3]), AF (3.2 [2.9–3.5]), VTE (1.8 [1.6–1.9]), TAA (1.7 [1.5–1.9]), HTN (2.0 [1.8–2.2]), HC (4.7 [4.1–5.3]), and LPA (41.8 [26.7–65.7]). Performance was robust across ancestries, though stronger effect sizes were observed in individuals of European ancestry. Report output was structured to provide clinicians with trait-specific percentile ranks, relative risk estimates, and links to preventive lifestyle interventions. In MGBB, the 8 PRSs developed for our clinical report identified individuals with ≥3-fold elevated risk for one or more traits in 72.5% of participants.
Conclusion: We introduce a validated, cardiometabolic PRS report designed for routine clinical use. The 8 PRSs in our report demonstrate strong performance and may enable personal risk stratification and precision prevention strategies for patients with high inherited cardiometabolic risk.