def rdkit_mmff94_xyz(smiles, **kwargs):
"""
Returns the string of the XYZ file obtained performing the MMFF94 molecular mechanics optimization of the given
SMILES using RDKit.
Writing temporary files in $MM_WORKING_DIR if defined or otherwise in /tmp
:param smiles: input_SMILES
:param max_iterations: max number of iterations (default 500)
:return : XYZ string of optimized geometry, success (whether the MM optimization was successful and the smiles has
stayed identical after optimization)
"""
working_dir = os.environ[
"MM_WORKING_DIR"] if "MM_WORKING_DIR" in os.environ else "/tmp"
# Converting the molecule to RDKit object
mol = MolFromSmiles(smiles)
smi_canon = MolToSmiles(MolFromSmiles(smiles))
# Setting paths
filename_smiles = str(os.getpid()) + "_" + smi_to_filename(smi_canon)
xyz_path = join(working_dir, filename_smiles + '.xyz')
post_MM_smi_path = join(working_dir, filename_smiles + '.smi')
# Computing geometry
try:
# Adding implicit hydrogens
mol = AddHs(mol)
# MM optimization
EmbedMolecule(mol)
value = MMFFOptimizeMolecule(mol, maxIters=kwargs["max_iterations"])
# Success if returned value is null
success_RDKIT_output = value == 0
# Computing XYZ from optimized molecule
xyz_str = MolToXYZBlock(mol)
# Writing optimized XYZ to file
with open(xyz_path, "w") as f:
f.writelines(xyz_str)
# Success if the optimization has converged and the post MM smiles is identical the pre MM smiles
success = success_RDKIT_output and check_identical_geometries(
xyz_path, smi_canon, post_MM_smi_path)
except Exception as e:
success = False
xyz_str = None
finally:
# Removing files
remove_files([post_MM_smi_path, xyz_path])
return xyz_str, success
def embed(self, mol):
success = EmbedMolecule(mol)
if success == -1:
msg = 'Failed to Embed Molecule {}'.format(mol.name)
if self.error_on_fail:
raise RuntimeError(msg)
elif self.warn_on_fail:
warnings.warn(msg)
return None
if self.add_hs:
return mol.add_hs(add_coords=True)
else:
return mol
def gen3D(mol):
mol = Chem.AddHs(mol)
if EmbedMolecule(mol) != 0:
raise ValueError("EmbedMolecule failed")
uff = False
try:
optimize(mol, MMFFOptimizeMolecule)
except ValueError:
uff = True
if uff:
optimize(mol, UFFOptimizeMolecule)
return uff, mol
# 入力する分子（thiametoxam）
smiles = 'CN1COCN(C1=N[N+](=O)[O-])CC2=CN=C(S2)Cl'
# ファイル名を決める
t = datetime.datetime.fromtimestamp(time.time())
psi4.set_output_file("{}_{}{}{}_{}{}.log".format(smiles, t.year, t.month,
t.day, t.hour, t.minute))
# SMILES から三次元構造を発生させて、粗3D構造最適化
mol = Chem.MolFromSmiles(smiles)
mol = Chem.AddHs(mol)
params = ETKDGv3()
params.randomSeed = 1
EmbedMolecule(mol, params)
# MMFF（Merck Molecular Force Field） で構造最適化する
MMFFOptimizeMolecule(mol)
#UFF（Universal Force Field）普遍力場で構造最適化したい場合は
#UFFOptimizeMolecule(mol)
conf = mol.GetConformer()
# Psi4 に入力可能な形式に変換する。
# 電荷とスピン多重度を設定（下は、電荷０、スピン多重度1)
mol_input = "0 1"
#各々の原子の座標をXYZフォーマットで記述
for atom in mol.GetAtoms():
mol_input += "\n " + atom.GetSymbol() + " " + str(conf.GetAtomPosition(atom.GetIdx()).x)\
def EnumerateStereoisomers(m,
options=StereoEnumerationOptions(),
verbose=False):
""" returns a generator that yields possible stereoisomers for a molecule
Arguments:
- m: the molecule to work with
- options: parameters controlling the enumeration
- verbose: toggles how verbose the output is
If m has stereogroups, they will be expanded
A small example with 3 chiral atoms and 1 chiral bond (16 theoretical stereoisomers):
>>> from rdkit import Chem
>>> from rdkit.Chem.EnumerateStereoisomers import EnumerateStereoisomers, StereoEnumerationOptions
>>> m = Chem.MolFromSmiles('BrC=CC1OC(C2)(F)C2(Cl)C1')
>>> isomers = tuple(EnumerateStereoisomers(m))
>>> len(isomers)
16
>>> for smi in sorted(Chem.MolToSmiles(x, isomericSmiles=True) for x in isomers):
... print(smi)
...
F[C@@]12C[C@@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@@]12C[C@@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@@]12C[C@@]1(Cl)C[C@H](/C=C/Br)O2
F[C@@]12C[C@@]1(Cl)C[C@H](/C=C\\Br)O2
F[C@@]12C[C@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@@]12C[C@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@@]12C[C@]1(Cl)C[C@H](/C=C/Br)O2
F[C@@]12C[C@]1(Cl)C[C@H](/C=C\\Br)O2
F[C@]12C[C@@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@]12C[C@@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@]12C[C@@]1(Cl)C[C@H](/C=C/Br)O2
F[C@]12C[C@@]1(Cl)C[C@H](/C=C\\Br)O2
F[C@]12C[C@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@]12C[C@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@]12C[C@]1(Cl)C[C@H](/C=C/Br)O2
F[C@]12C[C@]1(Cl)C[C@H](/C=C\\Br)O2
Because the molecule is constrained, not all of those isomers can
actually exist. We can check that:
>>> opts = StereoEnumerationOptions(tryEmbedding=True)
>>> isomers = tuple(EnumerateStereoisomers(m, options=opts))
>>> len(isomers)
8
>>> for smi in sorted(Chem.MolToSmiles(x,isomericSmiles=True) for x in isomers):
... print(smi)
...
F[C@@]12C[C@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@@]12C[C@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@@]12C[C@]1(Cl)C[C@H](/C=C/Br)O2
F[C@@]12C[C@]1(Cl)C[C@H](/C=C\\Br)O2
F[C@]12C[C@@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@]12C[C@@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@]12C[C@@]1(Cl)C[C@H](/C=C/Br)O2
F[C@]12C[C@@]1(Cl)C[C@H](/C=C\\Br)O2
Or we can force the output to only give us unique isomers:
>>> m = Chem.MolFromSmiles('FC(Cl)C=CC=CC(F)Cl')
>>> opts = StereoEnumerationOptions(unique=True)
>>> isomers = tuple(EnumerateStereoisomers(m, options=opts))
>>> len(isomers)
10
>>> for smi in sorted(Chem.MolToSmiles(x,isomericSmiles=True) for x in isomers):
... print(smi)
...
F[C@@H](Cl)/C=C/C=C/[C@@H](F)Cl
F[C@@H](Cl)/C=C/C=C/[C@H](F)Cl
F[C@@H](Cl)/C=C/C=C\\[C@H](F)Cl
F[C@@H](Cl)/C=C\\C=C/[C@@H](F)Cl
F[C@@H](Cl)/C=C\\C=C/[C@H](F)Cl
F[C@@H](Cl)/C=C\\C=C\\[C@@H](F)Cl
F[C@@H](Cl)/C=C\\C=C\\[C@H](F)Cl
F[C@H](Cl)/C=C/C=C/[C@H](F)Cl
F[C@H](Cl)/C=C\\C=C/[C@H](F)Cl
F[C@H](Cl)/C=C\\C=C\\[C@H](F)Cl
By default the code only expands unspecified stereocenters:
>>> m = Chem.MolFromSmiles('BrC=C[C@H]1OC(C2)(F)C2(Cl)C1')
>>> isomers = tuple(EnumerateStereoisomers(m))
>>> len(isomers)
8
>>> for smi in sorted(Chem.MolToSmiles(x,isomericSmiles=True) for x in isomers):
... print(smi)
...
F[C@@]12C[C@@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@@]12C[C@@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@@]12C[C@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@@]12C[C@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@]12C[C@@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@]12C[C@@]1(Cl)C[C@@H](/C=C\\Br)O2
F[C@]12C[C@]1(Cl)C[C@@H](/C=C/Br)O2
F[C@]12C[C@]1(Cl)C[C@@H](/C=C\\Br)O2
But we can change that behavior:
>>> opts = StereoEnumerationOptions(onlyUnassigned=False)
>>> isomers = tuple(EnumerateStereoisomers(m, options=opts))
>>> len(isomers)
16
Since the result is a generator, we can allow exploring at least parts of very
large result sets:
>>> m = Chem.MolFromSmiles('Br' + '[CH](Cl)' * 20 + 'F')
>>> opts = StereoEnumerationOptions(maxIsomers=0)
>>> isomers = EnumerateStereoisomers(m, options=opts)
>>> for x in range(5):
... print(Chem.MolToSmiles(next(isomers),isomericSmiles=True))
F[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)Br
F[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)Br
F[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)Br
F[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)Br
F[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)Br
Or randomly sample a small subset. Note that if we want that sampling to be consistent
across python versions we need to provide a random number seed:
>>> m = Chem.MolFromSmiles('Br' + '[CH](Cl)' * 20 + 'F')
>>> opts = StereoEnumerationOptions(maxIsomers=3,rand=0xf00d)
>>> isomers = EnumerateStereoisomers(m, options=opts)
>>> for smi in isomers: #sorted(Chem.MolToSmiles(x, isomericSmiles=True) for x in isomers):
... print(Chem.MolToSmiles(smi))
F[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)Br
F[C@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)Br
F[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)[C@@H](Cl)[C@H](Cl)[C@@H](Cl)Br
"""
tm = Chem.Mol(m)
for atom in tm.GetAtoms():
atom.ClearProp("_CIPCode")
flippers = _getFlippers(tm, options)
nCenters = len(flippers)
if not nCenters:
yield tm
return
if (options.maxIsomers == 0 or 2**nCenters <= options.maxIsomers):
bitsource = _RangeBitsGenerator(nCenters)
else:
if options.rand is None:
# deterministic random seed invariant to input atom order
seed = hash(
tuple(
sorted([(a.GetDegree(), a.GetAtomicNum())
for a in tm.GetAtoms()])))
rand = random.Random(seed)
elif isinstance(options.rand, random.Random):
# other implementations of Python random number generators
# can inherit from this class to pick up utility methods
rand = options.rand
else:
rand = random.Random(options.rand)
bitsource = _UniqueRandomBitsGenerator(nCenters, options.maxIsomers,
rand)
isomersSeen = set()
numIsomers = 0
for bitflag in bitsource:
for i in range(nCenters):
flag = bool(bitflag & (1 << i))
flippers[i].flip(flag)
isomer = Chem.Mol(tm)
Chem.SetDoubleBondNeighborDirections(isomer)
isomer.ClearComputedProps()
Chem.AssignStereochemistry(isomer,
cleanIt=True,
force=True,
flagPossibleStereoCenters=True)
if options.unique:
cansmi = Chem.MolToSmiles(isomer, isomericSmiles=True)
if cansmi in isomersSeen:
continue
isomersSeen.add(cansmi)
if options.tryEmbedding:
ntm = Chem.AddHs(isomer)
cid = EmbedMolecule(ntm, randomSeed=bitflag)
if cid >= 0:
conf = Chem.Conformer(isomer.GetNumAtoms())
for aid in range(isomer.GetNumAtoms()):
conf.SetAtomPosition(
aid,
ntm.GetConformer().GetAtomPosition(aid))
isomer.AddConformer(conf)
else:
cid = 1
if cid >= 0:
yield isomer
numIsomers += 1
if options.maxIsomers != 0 and numIsomers >= options.maxIsomers:
break
elif verbose:
print("%s failed to embed" %
(Chem.MolToSmiles(isomer, isomericSmiles=True)))
def build_molecules(starting_mol,
atom_additions=None,
stereoisomers=True,
sa_score_threshold=3.,
tryEmbedding=True):
"""Return an iterator of molecules that result from a single manipulation
(i.e., atom / bond addition) to the starting molecule
Arguments:
starting_mol: rdkit.Mol
The starting molecule as an rdkit Mol object
atom_additions: list of elements
Types of atoms that can be added. Defaults to ('C', 'N', 'O')
stereoisomers: bool
Whether to iterate over potential stereoisomers of the given molecule
as seperate molecules
sa_score_threshold: float or None
Whether to calculate the sa_score of the given molecule, and withold
molecules that have a sa_score higher than the threshold.
tryEmbedding: bool
whether to try an rdkit 3D embedding of the molecule
Yields:
rdkit.Mol, corresponding to modified input
"""
if atom_additions == None:
atom_additions = ('C', 'N', 'O')
def get_valid_partners(atom):
""" For a given atom, return other atoms it can be connected to """
return list(
set(range(starting_mol.GetNumAtoms())) -
set((neighbor.GetIdx() for neighbor in atom.GetNeighbors())) -
set(range(
atom.GetIdx())) - # Prevent duplicates by only bonding forward
set((atom.GetIdx(), ))
| set(np.arange(len(atom_additions)) + starting_mol.GetNumAtoms()))
def get_valid_bonds(atom1_idx, atom2_idx):
""" Compare free valences of two atoms to calculate valid bonds """
free_valence_1 = get_free_valence(
starting_mol.GetAtomWithIdx(atom1_idx))
if atom2_idx < starting_mol.GetNumAtoms():
free_valence_2 = get_free_valence(
starting_mol.GetAtomWithIdx(int(atom2_idx)))
else:
free_valence_2 = pt.GetDefaultValence(
atom_additions[atom2_idx - starting_mol.GetNumAtoms()])
return range(min(min(free_valence_1, free_valence_2), 3))
def add_bond(atom1_idx, atom2_idx, bond_type):
""" Given two atoms and a bond type, execute the addition using rdkit """
num_atom = starting_mol.GetNumAtoms()
rw_mol = Chem.RWMol(starting_mol)
if atom2_idx < num_atom:
rw_mol.AddBond(atom1_idx, atom2_idx, bond_orders[bond_type])
else:
rw_mol.AddAtom(Chem.Atom(atom_additions[atom2_idx - num_atom]))
rw_mol.AddBond(atom1_idx, num_atom, bond_orders[bond_type])
return rw_mol
def enumerate_stereoisomers(mol, to_use):
""" We likely want to distinguish between stereoisomers, so we do that here """
if not to_use:
# Give an easy way to pass through this function if this feature isn't used
return (mol, )
else:
opts = StereoEnumerationOptions(unique=True)
return tuple(EnumerateStereoisomers(mol, options=opts))
generated_smiles = []
# Construct the generator
for i, atom in enumerate(starting_mol.GetAtoms()):
for partner in get_valid_partners(atom):
for bond_order in get_valid_bonds(i, partner):
mol = add_bond(i, partner, bond_order)
Chem.SanitizeMol(mol)
for isomer in enumerate_stereoisomers(mol, stereoisomers):
smiles = Chem.MolToSmiles(mol)
if smiles not in generated_smiles:
generated_smiles += [smiles]
if sa_score_threshold is not None:
if sascorer.calculateScore(
isomer) >= sa_score_threshold:
continue
if tryEmbedding:
ntm = Chem.AddHs(isomer)
try:
assert EmbedMolecule(ntm) >= 0
except (AssertionError, RuntimeError):
# Failed a 3D embedding
continue
yield isomer