molclaw-quickvina-docking

GPU-accelerated molecular docking computes protein-ligand binding affinities and generates docking pose files using QuickVina2-GPU, requiring prepared receptor and ligand PDBQT inputs along with defined pocket center and box sizes.

Protein-ligand docking requires receptor and ligand PDBQT inputs, or upstream conversion from PDB and SMILES formats, along with specified pocket center coordinates and box sizes with a minimum of 25 Å per axis.

Failed docking poses with invalid binding affinities trigger negative affinity validation and progressive box enlargement retries, ensuring mandatory pose file retrieval for downstream rescoring and hit prioritization.

QuickVina docking requires a minimum box size of 25 Å per axis to ensure adequate space for ligand conformational sampling and accurate binding affinity estimation during virtual screening.

The docking workflow integrates with protein preparation and pocket detection by orchestrating protein retrieval, chain extraction, structure repair, pocket detection, and format conversions into a single reproducible pipeline.

SMILES-derived ligands can be used for high-throughput virtual screening by converting them to PDBQT format and docking them against prepared protein structures to return affinity values and save pose PDBQT files.