Post

PREPRINT (not peer-reviewed). Preparing a molecule for docking can fail before docking starts: protonation, tautomer choice, and ring geometry all shape the input. Mato and colleagues describe MolScrub, an open-source Python tool that predicts protonation states, enumerates tautomers, and generates three-dimensional structures, with a reported average runtime of 0.3–0.5 seconds per molecule per CPU core. On a test split from a roughly 3,300-molecule pKa dataset, its model reached RMSE 0.78 and R² 0.90, but training and test data came from the same mainly monoprotic source. For tautomer enumeration, chemical filters raised precision to 0.39 while recall fell to 0.50 in the NIH database benchmark. The tool is a structure-preparation aid, not evidence that downstream docking predictions are correct. Find the paper on ChemRxiv by searching the full title: MolScrub: an open-source tool to generate conformers and protonate drug-like molecules.

3:38:47 PM · Nov 22, 2026
Sign in to reply and join the conversation.