Preprint applies cross-trait polygenic scores to dissect Alzheimer's disease heterogeneity

A preprint posted to bioRxiv (15 May 2026) introduces a cross-trait polygenic score (PGS) analysis framework aimed at understanding the genetic basis of heterogeneity within Alzheimer's disease (AD). The authors note that AD displays substantial variation in age of onset, rate of progression, and pathological profile — variation that standard case-control GWAS approaches are poorly powered to resolve.

The strategy leverages polygenic scores pre-trained on traits that share genetic architecture with AD through pleiotropy, applying them across multiple cohorts with deep phenotypic characterisation to identify which genetic axes of variation correlate with which clinical or pathological subtypes. The approach is presented as a way to make progress without requiring very large deeply phenotyped AD cohorts, which remain scarce.

This is a preprint and has not been peer-reviewed. The work is primarily of interest to statistical geneticists and AD researchers. It contributes to the methodological literature on polygenic score applications in complex disease and illustrates how cross-trait genetic approaches can extract additional signal from existing resources. The preprint does not present clinically applicable risk tools; the polygenic scores described are research instruments applied in cohort analyses.