Reproduce “Graph AI generates neurological hypotheses validated in molecular, organoid, and clinical systems” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ayush Noori, Joaquín Polonuer, Katharina Meyer, Bogdan Budnik, Shad Morton, Xinyuan Wang, Sumaiya Nazeen, Yingnan He and 21 more
A novel graph AI model, Proton, generates and prospectively validates hypotheses for neurological disorders by integrating multi-scale biological data. This approach identified five drugs associated with a reduced 7-year dementia risk in a clinical cohort of over 600,000 patients, repurposed calcitriol for bipolar disorder organoids, and predicted genetic links and environmental toxins in Parkinson's disease.
Key insights
- Graph AI models, when trained on context-rich knowledge graphs like NeuroKG, can learn deep, aligned representations of biological entities across multiple scales.
- Rigorous prospective validation across diverse biological and clinical systems is essential to bridge the gap between AI predictions and real-world biological/clinical utility.
- AI can generate specific, testable hypotheses that lead to new mechanistic insights, drug repurposing candidates, and clinically actionable findings in neurological diseases.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.