chem-docking-void

To dock ligands into porous materials, you need a valid SMILES string for the ligand and a CIF file for the host structure. The tool generates RDKit conformers, samples the host's pore space, and outputs ranked docked CIF poses.

Voronoi clustering identifies accessible pore space within a CIF host structure to sample physically plausible insertion sites. It places guest molecules throughout these voids, applying collision-aware distance checks to filter steric clashes between host and guest.

You need a SMILES ligand, a CIF host, and an environment with RDKit, pymatgen, and VOID installed. The process generates 3D conformers optimizes them with MMFF94, and selects the lowest-energy candidates for ranking.

Yes, you can use RDKit and pymatgen to prepare zeolite inclusion poses for DFT evaluation. The workflow exports ranked docked complexes as CIF files and a docking_results.json summary containing pose metadata for downstream DFT or MLIP steps.

The best way to sample spatially diverse binding poses inside a MOF is using Voronoi-based porous sampling. This method covers the accessible pore space by placing multiple guest conformers and applying physics-informed minimum-distance fitness functions to accept valid poses.

A key limitation is the requirement for specific dependencies like VOID, RDKit, and pymatgen in the execution environment. Additionally, the tool filters collisions using physics-informed minimum-distance fitness functions, which may reject poses if the host pores are too small.