Machine learning reveals heterogeneous associations between environmental factors and cardiometabolic diseases across polygenic risk scores

Abstract

Although polygenic risk scores (PRSs) are expected to be helpful in precision medicine, it remains unclear whether high-PRS groups are more likely to benefit from preventive interventions for diseases. Recent methodological advancements enable us to predict treatment effects at the individual level. We employed causal forest to explore the relationship between PRSs and individual risk of diseases associated with certain environmental factors. Following simulations illustrating its performance, we applied our approach to investigate the individual risk of cardiometabolic diseases, including coronary artery diseases (CAD) and type 2 diabetes (T2D), associated with obesity and smoking among individuals from UK Biobank (UKB; n = 369,942) and BioBank Japan (BBJ; n = 149,421). Here we find the heterogeneous association of obesity and smoking with diseases across PRS values, complicated by the multi-dimensional combination of individual characteristics such as age and sex. The highest positive correlations of PRSs and the exposure-related disease risks are observed between obesity and T2D in UKB and between smoking and CAD in BBJ (Spearman’s ρ = 0.61 and 0.32, respectively). However, most relationships are weak or negative, suggesting that high-PRS groups will not necessarily benefit most from environmental factor prevention. Our study highlights the importance of individual-level prediction of disease risks associated with target exposure in precision medicine. This study aimed to understand if people with a high genetic risk for certain diseases benefit more from preventive strategies. Using a machine-learning-based method, we analyzed data from large groups of people in the UK and Japan. We examined the risk of heart and metabolic diseases in relation to obesity and smoking. The results showed that the link between genetic risk and disease is complex and varies widely among individuals. Our results suggested that those with a high genetic risk for disease may not always benefit more from the prevention of obesity and smoking. This finding suggests that we need to consider more than risk in decisions on how to prevent diseases in individuals. Naito and Inoue, et al. apply machine learning to reveal heterogeneous associations between environmental factors and diseases across polygenic risk scores. Focusing on cardiometabolic diseases shows that those with high genetic disease susceptibility may not necessarily benefit the most from the reduction of corresponding disease risk factors.