Addressing Overlapping Sample Challenges in Genome-Wide Association Studies: Meta-Reductive Approach

Abstract

AbstractPolygenic risk scores (PRS) are instrumental in genetics, offering insights into an individual level genetic risk to a range of diseases based on accumulated genetic variations. These scores rely on Genome-Wide Association Studies (GWAS). However, precision in PRS is often challenged by the requirement of extensive sample sizes and the potential for overlapping datasets that can inflate PRS calculations. In this study, we present a novel methodology, Meta-Reductive Approach (MRA), that was derived algebraically to adjust GWAS results, aiming to neutralize the influence of select cohorts. Our approach recalibrates summary statistics using algebraic derivations. Validating our technique with datasets from Alzheimer’s disease studies, we showed perfect correlation between summary statistics of proposed approach and “leave-one-out” strategy. This innovative method offers a promising avenue for enhancing the accuracy of PRS, especially when derived from meta-analyzed GWAS data.Availability and implementationMRA script is freely available athttps://github.com/hihg-um/MRAas an R function.