Network Medicine Approach Unravels Endophenotype Signature in Alzheimer's Disease through Large-Scale Comparative...

Alzheimer's disease (AD) is the most common neurodegenerative disease burdening public health. We proposed a network-based infrastructure to identify protein signatures for five AD pathological endophenotypes: amyloidosis, tauopathy, vascular dysfunction, lysosomal dysfunction, and neuroinflammation. We analyzed 23 proteomic data sets from AD patients and transgenic mouse models, using network proximity to measure associations between endophenotype modules and differentially expressed proteins (DEPs) in the integrated AD proteome. We focused on the vascular dysfunction signature with 21 DEPs by integrating RNA-seq, single-cell transcriptomics, GWAS, and literature. Experiments on APP/PS1 and MCAO models highlighted three proteins (SEPT5, SNAP25, STXBP1) as novel AD biomarker candidates. This study demonstrates a network medicine framework for deciphering endophenotype signatures in AD.