Here, the authors analyze polygenic scores for 457 phenotypes, finding 49 with robust polygenic score predictive models ( > 0.6 AUC). They find that population diversity and environmental factor integration are key considerations to improving the model’s performance.