Clinically Validated
Polygenic Risk Scores

GWASs look for genetic differences between disease cases and healthy controls. Each GWAS examines millions of variants. If variants are much more or less common in the case group they may be associated with the disease. Each variant’s effect size is an estimate of its contribution to disease risk. Aggregating the effects of the disease-associated variants yields PRSs [R, R].

Polygenic Risk Scores (PRS) sum up the impacts of many inherited variants on the risk of developing a specific disease. They can identify high-risk individuals to target for early therapeutic intervention [R]. Most PRS models include thousands of variants with small effects on disease risk. These variants are identified by Genome-Wide Association Studies (GWAS).

PRS models built from larger or even multiple GWASs tend to be more predictive. We generate and test a great number of PRS models. Once we build a PRS model, we check the risk predictions in another validation dataset. These datasets include genetic information and case-control status in a new independent group.

To check that our PRS are accurate in diverse patient populations, we create models for different ancestries. We then test their performance in multiple populations. We also apply techniques to decrease the impact of biased associations. These practices significantly increase the effectiveness of our PRS reports in clinical settings.