PoseBusters: Plausibility checks for generated molecule poses.

PoseBusters is a command line tool and Python library to check the output of molecular conformation generators, docking programs, or other programs that generate molecules in 3D.

Quick start

PoseBusters can be installed from PyPI.

>>> pip install posebusters

Use the bust command to check molecules, docked ligands or generated molecules conditioned on a protein.

>>> bust redocked_ligand.sdf -l crystal_ligand.sdf -p protein.pdb --outfmt short
redocked_ligand.sdf redocked_ligand  passes (25 / 25)
...

>>> bust redocked_ligand.sdf -l crystal_ligand.sdf -p protein.pdb --outfmt long
Long summary for redocked_ligand.sdf redocked_ligand
MOL_PRED loaded                          .   
MOL_TRUE loaded                          .   
MOL_COND loaded                          .   
Sanitization                             .   
All atoms connected                      .   
Molecular formula                        .   
Molecular bonds                          .   
Double bond stereochemistry              .   
Tetrahedral chirality                    .   
Bond lengths                             .   
Bond angles                              .   
...

>>> bust redocked_ligand.sdf -l crystal_ligand.sdf -p protein.pdb --outfmt csv
file,molecule,mol_pred_loaded,mol_true_loaded,mol_cond_loaded,sanitization,all_atoms_connected,molecular_formula,molecular_bonds,double_bond_stereochemistry,tetrahedral_chirality,bond_lengths,bond_angles,internal_steric_clash,aromatic_ring_flatness,double_bond_flatness,internal_energy,protein-ligand_maximum_distance,minimum_distance_to_protein,minimum_distance_to_organic_cofactors,minimum_distance_to_inorganic_cofactors,minimum_distance_to_waters,volume_overlap_with_protein,volume_overlap_with_organic_cofactors,volume_overlap_with_inorganic_cofactors,volume_overlap_with_waters,rmsd_≤_2å
redocked_ligand.sdf,redocked_ligand,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True,True
...

For more usage examples, bulk processing, and the Python API see the documentation for the command line tool and the Python library.

Docking methods comparison

For more detailed information about the tests and for a study using PoseBusters to compare docking methods, refer to our preprint:

@online{buttenschoen2023posebusters,
   title = {{{PoseBusters}}: {{AI-based}} Docking Methods Fail to Generate Physically Valid Poses or Generalise to Novel Sequences},
   shorttitle = {{{PoseBusters}}},
   author = {Buttenschoen, Martin and Morris, Garrett M. and Deane, Charlotte M.},
   date = {2023-08-10},
   eprint = {2308.05777},
   eprinttype = {arxiv}
}