AI scans 4.6 million compounds in hours to predict hydrogen positions in drug-like molecules

Researchers at New York University developed an AI model that predicts hydrogen atom positions in drug-like molecules by learning chemical patterns linked to stability, enabling rapid identification of stable tautomers. The model leverages a large dataset from the Cambridge Structural Database and improves tautomer assignments in biomolecular complexes where experimental data is insufficient.

This Wire brief sits within Fusion42's coverage of Drug Discovery, and 3 sources have reported it between 23 Sep 2026 and 24 Sep 2026.