Merge pull request #2298 from ncfrey/docking-dev

from deepchem.utils.vina_utils import write_vina_conf

from deepchem.utils.vina_utils import load_docked_ligands

from deepchem.utils.vina_utils import prepare_inputs

from deepchem.utils.voxel_utils import convert_atom_to_voxel

from deepchem.utils.voxel_utils import convert_atom_pair_to_voxel

      xyz = rdkit_utils.get_xyz_from_mol(ligand)

      assert score < 0  # This is a binding free energy

      assert np.count_nonzero(xyz) > 0

  def test_prepare_inputs(self):

    pdbid = '3cyx'

    ligand_smiles = 'CC(C)(C)NC(O)C1CC2CCCCC2C[NH+]1CC(O)C(CC1CCCCC1)NC(O)C(CC(N)O)NC(O)C1CCC2CCCCC2N1'

    protein, ligand = vina_utils.prepare_inputs(

        pdbid, ligand_smiles, pdb_name=pdbid)

    assert np.isclose(protein.GetNumAtoms(), 1415, atol=3)

    assert np.isclose(ligand.GetNumAtoms(), 124, atol=3)

    protein, ligand = vina_utils.prepare_inputs(pdbid + '.pdb',

                                                'ligand_' + pdbid + '.pdb')

    assert np.isclose(protein.GetNumAtoms(), 1415, atol=3)

    assert np.isclose(ligand.GetNumAtoms(), 124, atol=3)

    os.remove(pdbid + '.pdb')

    os.remove('ligand_' + pdbid + '.pdb')

    os.remove('tmp.pdb')

    mol = Chem.MolFromPDBBlock(str(pdb_block), sanitize=False, removeHs=False)

    molecules.append(mol)

  return molecules, scores

def prepare_inputs(protein: str,

                   ligand: str,

                   replace_nonstandard_residues: bool = True,

                   remove_heterogens: bool = True,

                   remove_water: bool = True,

                   add_hydrogens: bool = True,

                   pH: float = 7.0,

                   optimize_ligand: bool = True,

                   pdb_name: Optional[str] = None) -> Tuple[RDKitMol, RDKitMol]:

  """This prepares protein-ligand complexes for docking.

  Autodock Vina requires PDB files for proteins and ligands with

  sensible inputs. This function uses PDBFixer and RDKit to ensure

  that inputs are reasonable and ready for docking. Default values

  are given for convenience, but fixing PDB files is complicated and

  human judgement is required to produce protein structures suitable

  for docking. Always inspect the results carefully before trying to 

  perform docking.

  Parameters

  ----------

  protein: str

    Filename for protein PDB file or a PDBID.

  ligand: str

    Either a filename for a ligand PDB file or a SMILES string.

  replace_nonstandard_residues: bool (default True)

    Replace nonstandard residues with standard residues.

  remove_heterogens: bool (default True)

    Removes residues that are not standard amino acids or nucleotides.

  remove_water: bool (default True)

    Remove water molecules.

  add_hydrogens: bool (default True)

    Add missing hydrogens at the protonation state given by `pH`.

  pH: float (default 7.0)

    Most common form of each residue at given `pH` value is used.

  optimize_ligand: bool (default True)

    If True, optimize ligand with RDKit. Required for SMILES inputs.

  pdb_name: Optional[str]

    If given, write sanitized protein and ligand to files called

    "pdb_name.pdb" and "ligand_pdb_name.pdb"

  Returns

  -------

  Tuple[RDKitMol, RDKitMol]

    Tuple of `protein_molecule, ligand_molecule` with 3D information.

  Note

  ----

  This function requires RDKit and OpenMM to be installed.

  Read more about PDBFixer here: https://github.com/openmm/pdbfixer.

  Examples

  --------

  >>> p, m = prepare_inputs('3cyx', 'CCC')

  >>> p.GetNumAtoms()

  1415

  >>> m.GetNumAtoms()

  11

  >>> p, m = prepare_inputs('3cyx', 'CCC', remove_heterogens=False)

  >>> p.GetNumAtoms()

  1720

  """

  try:

    from rdkit import Chem

    from pdbfixer import PDBFixer

    from simtk.openmm.app import PDBFile

  except ModuleNotFoundError:

    raise ImportError(

        "This function requires RDKit and OpenMM to be installed.")

  if protein.endswith('.pdb'):

    fixer = PDBFixer(protein)

  else:

    fixer = PDBFixer(url='https://files.rcsb.org/download/%s.pdb' % (protein))

  if ligand.endswith('.pdb'):

    m = Chem.MolFromPDBFile(ligand)

  else:

    m = Chem.MolFromSmiles(ligand, sanitize=True)

  # Apply common fixes to PDB files

  if replace_nonstandard_residues:

    fixer.findMissingResidues()

    fixer.findNonstandardResidues()

    fixer.replaceNonstandardResidues()

  if remove_heterogens and not remove_water:

    fixer.removeHeterogens(True)

  if remove_heterogens and remove_water:

    fixer.removeHeterogens(False)

  if add_hydrogens:

    fixer.addMissingHydrogens(pH)

  PDBFile.writeFile(fixer.topology, fixer.positions, open('tmp.pdb', 'w'))

  p = Chem.MolFromPDBFile('tmp.pdb', sanitize=True)

  # Optimize ligand

  if optimize_ligand:

    m = Chem.AddHs(m)  # need hydrogens for optimization

    Chem.AllChem.EmbedMolecule(m)

    Chem.AllChem.MMFFOptimizeMolecule(m)

  if pdb_name:

    Chem.rdmolfiles.MolToPDBFile(p, '%s.pdb' % (pdb_name))

    Chem.rdmolfiles.MolToPDBFile(m, 'ligand_%s.pdb' % (pdb_name))

  return (p, m)

Debug Utilities

---------------

Docking Utilities

-----------------

These utilities assist in file preparation and processing for molecular

docking.

.. autofunction:: deepchem.utils.vina_utils.write_vina_conf

.. autofunction:: deepchem.utils.vina_utils.load_docked_ligands

.. autofunction:: deepchem.utils.vina_utils.prepare_inputs

Print Threshold

^^^^^^^^^^^^^^^