This repository provides a Python implementation of two improved protocols for pose sampling: GLOW (auGmented sampLing with sOftened vdW potential) and IVES (IteratiVe Ensemble Sampling). For more details on the protocols, refer to this link. In addition, we provided new cross-docking datasets generated using GLOW and IVES containing approximately 5,000 protein-ligand cross-docking pairs, serving as invaluable resources for training and evaluating machine learning-based scoring functions. The dataset can be downloaded from here.
conda create -n glow_ives python=3.10 pip -y
conda activate glow_ives
conda install -y -c conda-forge openmm==7.7.0
conda install -y -c conda-forge openff-toolkit==0.14.0
conda install -y -c conda-forge openmmforcefields==0.11.2
conda install -y -c omnia pdbfixer==1.8.1
conda install -y -c conda-forge openbabel rdkit smina
pip install click tqdm biopandas==0.4.1 easy-parallel
To run GLOW, run the following command:
python -m glow_ives.sampling.glow glow <OUTPUT_PREFIX> <DOCKING_PROTEIN_STRUCTURE_PDBQT> <LIGAND_TO_DOCK_SDF> <REF_LIGAND_FOR_BOX_CENTER_SDF>
To run IVES, run the following command:
# python -m glow_ives.sampling.ives run_ives <OUTPUT_DIR> <DOCKING_PROTEIN_STRUCTURE_PDB> <DOCKING_PROTEIN_STRUCTURE_PDBQT> <LIGAND_TO_DOCK_SDF> <REF_LIGAND_FOR_BOX_CENTER_SDF> --num_conformations <NUM_OF_PROTEIN_CONFORMATIONS> --num_cpus <NUM_OF_CPUS> --num_iter <NUM_OF_ITERATIONS>