Polygenic risk vectors (PRV) improve genetic risk stratification for cardio-metabolic diseases

Abstract

1.AbstractAccurate disease risk stratification can lead to more precise and personalized prevention and treatment of diseases. As an important component to disease risk, genetic risk factors can be utilized as an early and stable predictor for disease onset. Recently, the polygenic risk score (PRS) method has combined the effects from hundreds to millions of single nucleotide polymorphisms (SNPs) into a score that can be used for genetic risk stratification. However, current PRS approaches only utilize the additive associations between SNPs and disease risk in a one-dimensional score. Here, we show that leveraging multiple types of genetic effects in multi-dimensional risk vectors, or a polygenic risk vector (PRV), can improve the stratification of cardio-metabolic diseases risks. Using data from UK Biobank (UKBB) and Electronic Medical Records and Genomics (eMERGE) Network biobank linked electronic health records (EHR) as development and evaluation data, we found that the combined effects between the additive PRS and the dominant PRS outperformed either one in terms of disease risk stratification, especially for the individuals in the high-risk group. Our results demonstrate that disease risks are likely to be influenced by multiple types of genetic effects, and PRV could utilize these effects for better risk stratification while retaining the simplicity of the PRS method.