Many companies are announcing AI-powered virtual screening for drug discovery. Very few are publishing peer-reviewed research showing how the compounds those systems identify actually perform in human disease models.
A study led by Xinnan Wang's lab at Stanford University and published in Cell Chemical Biology documents how BIOPTIC B1 was used to identify novel neuroprotective compounds for Friedreich's ataxia, a rare hereditary disease affecting the nervous system and heart for which there is no cure.
The screen required only the SMILES of a single known Miro1 ligand as input. No 3D protein structure. BIOPTIC B1 searched 3.2 billion small molecules and returned 62 chemically novel candidates, 53 of which were synthesized and tested across two independent experimental pipelines. Two top hits matched or outperformed the lead compound at doses as low as 1 micromolar in FA patient-derived sensory neurons, with results confirmed across three independent patient lines.
This is the second peer-reviewed publication demonstrating BIOPTIC B1 in a drug discovery application, following earlier work on LRRK2 as a Parkinson's disease target. A patent on the identified compounds has been filed.