def get_conformer_rmsd(mol: RDKitMol) -> np.ndarray:
"""
Calculate conformer-conformer RMSD.
Parameters
----------
mol: rdkit.Chem.rdchem.Mol
RDKit Mol object
Returns
-------
rmsd: np.ndarray
A conformer-conformer RMSD value. The shape is `(NumConformers, NumConformers)`
"""
try:
from rdkit.Chem import AllChem
except ModuleNotFoundError:
raise ValueError("This function requires RDKit to be installed.")
rmsd = np.zeros((mol.GetNumConformers(), mol.GetNumConformers()),
dtype=float)
for i, ref_conf in enumerate(mol.GetConformers()):
for j, fit_conf in enumerate(mol.GetConformers()):
if i >= j:
continue
rmsd[i, j] = AllChem.GetBestRMS(mol, mol, ref_conf.GetId(),
fit_conf.GetId())
rmsd[j, i] = rmsd[i, j]
return rmsd
def generate_conformers(self, mol: RDKitMol) -> RDKitMol:
"""
Generate conformers for a molecule.
This function returns a copy of the original molecule with embedded
conformers.
Parameters
----------
mol: rdkit.Chem.rdchem.Mol
RDKit Mol object
Returns
-------
mol: rdkit.Chem.rdchem.Mol
A new RDKit Mol object containing the chosen conformers, sorted by
increasing energy.
"""
# initial embedding
mol = self.embed_molecule(mol)
if not mol.GetNumConformers():
msg = 'No conformers generated for molecule'
if mol.HasProp('_Name'):
name = mol.GetProp('_Name')
msg += ' "{}".'.format(name)
else:
msg += '.'
raise RuntimeError(msg)
# minimization and pruning
self.minimize_conformers(mol)
mol = self.prune_conformers(mol)
return mol
def prune_conformers(self, mol: RDKitMol) -> RDKitMol:
"""
Prune conformers from a molecule using an RMSD threshold, starting
with the lowest energy conformer.
Parameters
----------
mol: rdkit.Chem.rdchem.Mol
RDKit Mol object
Returns
-------
new_mol: rdkit.Chem.rdchem.Mol
A new rdkit.Chem.rdchem.Mol containing the chosen conformers, sorted by
increasing energy.
"""
try:
from rdkit import Chem
except ModuleNotFoundError:
raise ValueError("This function requires RDKit to be installed.")
if self.rmsd_threshold < 0 or mol.GetNumConformers() <= 1:
return mol
energies = self.get_conformer_energies(mol)
rmsd = self.get_conformer_rmsd(mol)
sort = np.argsort(energies) # sort by increasing energy
keep: List[float] = [] # always keep lowest-energy conformer
discard = []
for i in sort:
# always keep lowest-energy conformer
if len(keep) == 0:
keep.append(i)
continue
# discard conformers after max_conformers is reached
if len(keep) >= self.max_conformers:
discard.append(i)
continue
# get RMSD to selected conformers
this_rmsd = rmsd[i][np.asarray(keep, dtype=int)]
# discard conformers within the RMSD threshold
if np.all(this_rmsd >= self.rmsd_threshold):
keep.append(i)
else:
discard.append(i)
# create a new molecule to hold the chosen conformers
# this ensures proper conformer IDs and energy-based ordering
new_mol = Chem.Mol(mol)
new_mol.RemoveAllConformers()
conf_ids = [conf.GetId() for conf in mol.GetConformers()]
for i in keep:
conf = mol.GetConformer(conf_ids[i])
new_mol.AddConformer(conf, assignId=True)
return new_mol