Projecting the performance of risk prediction based on polygenic analyses of genome-wide association studies

Abstract

We report a new model to project the predictive performance of polygenic models based on the number and distribution of effect sizes for the underlying susceptibility alleles and the size of the training dataset. Using estimates of effect-size distribution and heritability derived from current studies, we project that while 45% of the variance of height has been attributed to common tagging Single Nucleotide Polymorphisms (SNP), a model trained on one million people may only explain 33.4% of variance of the trait. Current studies can identify 3.0%, 1.1%, and 7.0%, of the populations who are at two-fold or higher than average risk for Type 2 diabetes, coronary artery disease and prostate cancer, respectively. Tripling of sample sizes could elevate the percentages to 18.8%, 6.1%, and 12.2%, respectively. The utility of future polygenic models will depend on achievable sample sizes, underlying genetic architecture and information on other risk-factors, including family history.