OCDocker - open consensus docker
This project proposes the development, improvement, and validation of a molecular docking system that employs consensus scoring to assist in identifying compounds with therapeutic potential. To this end, a machine learning model was trained using molecular descriptors of ligands and receptors from specialized databases, alongside data derived from docking simulations performed with various molecular docking programs. The consensus approach demonstrated a competitive-and often superior-ability to distinguish active compounds from inactive ones, based on classifications from the DUDEz and PDBbind databases.
Team: Artur Duque Rossi, Guilherme Ian Spelta, Maria Clara Monachesi