A0929
Title: Recent challenges and biostatistical approaches in polygenic risk score modelling
Authors: Christian Staerk - IUF-Leibniz Research Institute for Environmental Medicine (Germany) [presenting]
Hannah Klinkhammer - University of Bonn (Germany)
Carlo Maj - University of Marburg (Germany)
Andreas Mayr - University of Bonn (Germany)
Abstract: Polygenic risk scores (PRS) quantify the genetic predisposition for various traits and clinical outcomes based on genotype data. The aim is to address recent challenges in PRS modelling from a biostatistical perspective. In particular, the focus is on three important challenges: first, training PRS models on high-dimensional and large-scale genotype data with hundreds of thousands of genetic variants and individuals requires scalable and interpretable statistical learning methods. Second, the transferability of PRS models to diverse populations with different ancestries is often limited, as models are typically trained on cohorts predominantly of European ancestry. Third, the evaluation of the prediction accuracy of PRS models is complicated by different and conflicting definitions of the commonly used R-squared measure on test data. To address these challenges, recent statistical learning approaches for fitting PRS models are presented on individual-level genotype data from the UK Biobank, including scalable boosting and causal inference methods. Furthermore, open problems for future research are highlighted to enhance the precision and integration of PRS models for personalized medicine.
