Internship in Polygenic Scores for Subtypes of Diseases

Complex common diseases, like stroke, cancer or Alzheimer disease, have a multi-faceted etiology including the genetic background as well as environmental and lifestyle factors. The aim of polygenic scores (PGS) is to quantify the genetic influences for each individual as accurately as possible. The vision of PGS is to enable personalized preventive risk assessment and therapy development. The performance of PGS for individualized prediction, however, has so far been limited. One reason for that is that diseases often consist of several subtypes which have differing genetic backgrounds but polygenic scores are typically derived and calibrated in very large datasets where information on subtype are missing or ignored. In the proposed project we will investigate if and to what extent the performance of PGS could be improved by generating PGS for subtypes. We will simulate subtypes with varying degrees of genetic overlap to elucidate this question as well as developing statistical methods for tuning polygenic risk scores in the presence of genetic-heterogeneity.