NYU scientists have created an AI tool using crystallographic data to accurately predict stable tautomers in drug-like molecules.
Key Details
- 1A graph neural network model was trained on over 1.1 million tautomeric states from the Cambridge Structural Database.
- 2The AI predicts hydrogen atom positions to resolve tautomers, addressing a challenge unmet by X-ray crystallography or quantum mechanics.
- 3The model corrected about 2.5% of tautomer assignments in a Protein Data Bank ligand dataset, improving chemical accuracy.
- 4The open-source tool (Tautomer-Predictor) analyzes ~4.6 million compounds in 3.2 hours on a single GPU node.
- 5Accurate tautomer prediction influences ligand-protein interaction studies and downstream drug discovery workflows.
Why It Matters
Reliable tautomer assignment is crucial for modeling ligand-protein interactions in drug discovery, which is closely linked to molecular imaging and radiopharmaceutical development. Advances in AI-based molecular modeling can inform radiology researchers working on precision diagnostics and new tracer design.
Source
EurekAlert
Related News
•EurekAlert
Photonic Encoder Transforms Medical Signals for Ultrafast AI Diagnosis
Researchers created a fiber-compatible, all-optical biomimetic spike encoder enabling rapid, precise conversion of medical signals for neuromorphic analysis.
•EurekAlert
Sydney Team Reveals Fractal Limits to Predictability in AI Training
University of Sydney researchers discover that deep learning models may remain unpredictable due to deep-rooted mathematical properties, even under controlled training conditions.
•EurekAlert
Machine Learning Tool Maps Alzheimer’s Risk on fMRI Spectrum
Researchers developed an AI framework using resting-state fMRI to assess Alzheimer's risk along a continuum from SuperAgers to AD patients.