The Alzheimer’s Disease Data Initiative is pleased to announce the winners of the Global Neurodegeneration Proteomics Consortium (GNPC) Data Challenge. Launched in October 2025, the challenge invited researchers around the world to analyze the GNPC v1 Harmonized Dataset — the largest proteomics dataset for neurodegenerative diseases, with 250 million protein measurements from 35,000 biosample analyses — and to answer priority research questions in Alzheimer's disease and related dementias.
All analyses were conducted on AD Workbench, AD Data Initiative’s flagship data sharing and analytics platform for Alzheimer’s research. Alongside the GNPC v1 Harmonized Dataset, participants also had the opportunity to draw on datasets from GAP Bio-Hermes-001, Answer ALS, Emory Goizueta Alzheimer’s Disease Research Center (ADRC), University of Kansas ADRC, and Mayo Clinic Study of Aging (MCSA).
The winning projects span detection, progression, prognosis, subtyping, prevention, and the biology of plasma proteins, demonstrating what becomes possible when proteomic data at this scale is openly available to the global research community.
In total, 10 prizes were awarded across six priority research questions, recognizing projects that either replicated known findings in GNPC or surfaced preliminarily novel results that warrant further follow-up and replication.
Question 1 — Protein dynamics over time: Do longitudinal protein changes track healthy aging and disease progression, including preclinical to post-diagnosis AD?
- First place: Wenyu Zhou, Teal Rise, Inc., Protein Dynamics Over Time: Tracking Healthy Aging and Disease Progression from Preclinical to Post-Diagnosis AD — revealed distinct longitudinal protein trajectory patterns separating healthy aging from AD progression, identifying early vulnerability and later neurodegenerative signatures.
- Second place: Juan Shu, University of Pennsylvania, Dynamic Protein Profiles and Aging Trajectories in Alzheimer's Disease — showed that AD is characterized by accelerated, age-dependent divergence in protein trajectories, and developed dynamic biomarkers that outperform static measurements for risk prediction.
Question 2 — Plasma markers and progression: Do plasma changes influence/reflect progression speed? Any promising therapeutic targets?
- First place: Junyoung Park, Stanford University, Uncovering Causal Drivers of Alzheimer's Disease Progression Rate through Longitudinal Plasma Proteomics and Mendelian Randomization — prioritized 11 proteins with convergent longitudinal and genetic evidence as potential causal regulators of AD progression, spanning vascular, immune, and metabolic biology.
- Second place: Evan Boyle, Octave Bioscience, Levodopa-Stratified Plasma Proteomics for Parkinson's Disease Progression — demonstrated that plasma AOC3-based medication stratification enables cross-sectional inference of Parkinson's disease progression and uncovers neuroinflammatory progression markers.
Question 3 — Prognostic markers: Can we predict cognitive decline and neurodegeneration?
- First place: Andre Altmann, University College London, Data-Driven CSF Proteomic Event-Based Modeling for Alzheimer's Disease Staging — built a data-driven CSF proteomic staging model that orders biomarker changes temporally and predicts cognitive decline and MCI-to-dementia conversion beyond Aβ and p-Tau.
- Second place: Felix Breach, University of Cambridge, AI-Omics Markers for Precise Dementia Prediction and Prognosis — delivered interpretable machine learning models that accurately distinguish dementia types and generate individualized prognostic scores correlated with future cognitive decline.
Question 4 — Subtyping AD: Can biomarker patterns define meaningful clinical subtypes?
- First place: Yann Le Guen, Stanford University, Plasma Proteomics Reveals Continuous, Region-Specific Vulnerability Axes in Alzheimer's Disease — established that AD heterogeneity is best captured by continuous, brain-region–specific plasma proteomic vulnerability gradients rather than discrete molecular subtypes.
Question 5 — Delaying onset: Can proteomics inform prevention?
- First place: Andrew Phipps, University of Tasmania, Can Proteomics Inform Prevention? Modifiable Risk and Molecular Signatures — demonstrated that higher modifiable dementia risk burden is reflected in distinct plasma proteomic signatures, implicating RNA processing and stress-related pathways.
- Second place: Yann Le Guen, Stanford University, Proteomic Signatures of APOE*ε4 Resilience Reveal Causal Pathways That Delay Alzheimer's Disease Onset — identified genetically supported plasma proteins associated with delayed AD onset in ε4 carriers, highlighting actionable resilience pathways in neurotransmission, metabolism, immune signaling, and myelination.
Question 6 — How do the cellular origins and biological pathways of plasma proteins reflect and influence Alzheimer’s disease processes, and what insights could this provide for developing diagnostic or therapeutic strategies?
- First place: Kuan-Lin Huang, Icahn School of Medicine at Mount Sinai, TRACE: Tracing Cellular Origins of Alzheimer's Disease Circulating Biomarkers — localized AD-associated plasma proteins to specific neuronal and glial cell types, linking circulating biomarkers to synaptic, immune, and metabolic brain pathways.
Thank you to every researcher who took on a priority research question, to the GNPC member cohorts and partners who made this dataset possible, and to the broader community working to accelerate discoveries in Alzheimer's and related dementias.
Interested in exploring the data and continuing the work?
Learn more about the GNPC Data Challenge, explore the GNPC v1 Harmonized Dataset, and sign up for AD Workbench.