Comparison of polygenic risk scores for coronary artery disease highlights obstacles to overcome for clinical use

Polygenic risk scores, or PRS, are a tool to estimate individuals’ liabilities to a disease or trait measurement based solely on genetic information. One commonly discussed potential use is in the clinic to identify people who are at greater risk of developing a disease. In this paper, we compare three PRS models that incorporate a large number of genetic markers for coronary artery disease (CAD). In the UK Biobank, the cohort which was used at some point in the creation or validation of each score, we calculated the association between CAD, the scores, and population structure for the white British subset. After adjusting for geographic and socioeconomic factors, CAD was not associated with the first principal components of genetic diversity, which reflect fine-scale population structure. In contrast, all three scores were confounded by these genetic components, highlighting that PRS may be influenced by genetic factors not directly causal for CAD, thereby raising concerns about their biases in clinical application.Furthermore, we investigated the differences in risk stratification using four different UK Biobank assessment centers as separate cohorts, and tested how missing genetic data affected risk stratification through simulation. We show that missing data impact classification for extreme individuals for high- and low-risk, and quantiles of risk are sensitive to individual-level genotype missingness. Distributions of scores varied between assessment centers, revealing that thresholding based on quantiles can be problematic for consistency across centers and populations. Based on these results, we discuss potential avenues of improvements of PRS methodologies for usage in clinical practice.