Summary-level genetic method improves Alzheimer's risk prediction

In European participants from the UK Biobank and simulated datasets, researchers evaluated endoPRS-SS, a computational method that leverages endophenotype data to improve disease risk prediction using summary statistics and linkage disequilibrium reference panels. The framework addresses constraints of the earlier endoPRS approach, which required individual-level genotype-phenotype data for model fitting. In simulations and real data analyses, endoPRS-SS maintained the predictive accuracy of endoPRS while increasing computational efficiency up to 90-fold. The authors also extended the tool to incorporate multiple endophenotypes simultaneously, showing that multi-endophenotype models outperform single-endophenotype polygenic risk scores. Notably, simultaneously incorporating both monocyte count and executive function scores significantly improved Alzheimer's disease prediction.