Dissecting Alzheimer's disease heterogeneity by cross-trait polygenic prediction
Mapping the genetic basis of inter-individual heterogeneity in multifactorial diseases opens the door to mechanistic insights and opportunities for targeted intervention. In Alzheimer's disease (AD), clinical and pathological heterogeneity is well recognized, but genetic dissection is limited by a lack of well-powered cohorts with deep phenotypic characterization.
Alzheimer’s disease is not one monolithic phenotype. People with the same diagnosis can differ in disease onset, progression, cognition, pathology, and treatment response. Understanding the genetic basis of that heterogeneity could point to more precise ways to interpret disease biology and stratify patients.
However, the existing approaches are not sufficient to characterize the genetic basis of disease heterogeneity. Deeply phenotyped cohorts such as ROSMAP measure cognition and neuropathology in detail, but they are not large enough for well-powered variant-level discovery across many disease features. Large-scale GWAS and biobanks have the sample size, but they usually lack the disease-specific phenotypic depth needed to ask why individuals with Alzheimer’s disease differ from one another.
Led by Bill Li, this work proposes cross-trait polygenic prediction as a way to bridge those regimes. Instead of using a single Alzheimer’s disease polygenic score for case-control prediction, the study applies a phenome-wide library of UK Biobank-derived PGS to a deeply characterized Alzheimer’s disease cohort. This turns PGS into a tool for dissecting within-disease heterogeneity and reveals multiple polygenic dimensions linked to cognitive and neuropathologic variation.
