Using Protein Signatures to Uncover Alzheimer's Disease Subtypes and Predict Disease Progression

This project aims to use protein measurements from biofluid to identify distinct biological subtypes of Alzheimer’s disease and predict how each patient’s disease will progress over time.

Alzheimer’s disease is the leading cause of dementia, but patients differ greatly in their biology and in how quickly the disease progresses, making treatment and prediction difficult. Our team has generated large-scale CSF protein data from 1,104 participants and built models that can predict disease onset and progression, showing the power of proteomics to capture this complexity. In this project, we will define CSF-based biological subtypes of ADRD and test whether blood-based proteins can identify the same subtypes and trajectories, offering a scalable and non-invasive path to diagnosis.

This proposal leverages one of the deepest proteomic datasets in Alzheimer’s research, profiling 2,492 proteins across diverse biological pathways in over 1,100 participants, to move beyond the traditional focus on amyloid and tau. We uniquely combine CSF-based molecular subtyping with a targeted plasma protein panel, using machine learning to determine whether a simple blood test can capture the same biological signatures as spinal fluid. Together, this approach offers a path toward patient stratification and personalized prognosis that is both biologically rich and clinically practical.

Identifying distinct biological subtypes of Alzheimer’s disease could transform clinical trials by ensuring the right patients are enrolled in the right studies, accelerating the development of targeted therapies. If plasma proteomics can reliably mirror CSF-derived signatures, it would provide a minimally invasive and scalable tool for early detection, disease monitoring, and prognosis that could be widely deployed in clinical settings. More broadly, this work will deepen our understanding of the diverse biological mechanisms driving Alzheimer’s, opening new therapeutic targets beyond amyloid and tau for the millions of patients currently without effective treatment options.