def get_modified_inchi_key(mol, atoms):
"""
Generates InChIKey for the input molecule.
Passed atoms of the molecule are mutated (atomic number changed)
based on the mapping defined in the function.
@param mol:
@param atoms:
@type mol: OEGraphMol
@type atoms: list[OEAtombase]
@return: str
"""
copy_mol = oechem.OEGraphMol(mol)
atom_map = {
oechem.OEElemNo_C: oechem.OEElemNo_Pb,
oechem.OEElemNo_N: oechem.OEElemNo_Bi,
oechem.OEElemNo_O: oechem.OEElemNo_Po,
oechem.OEElemNo_F: oechem.OEElemNo_At,
oechem.OEElemNo_S: oechem.OEElemNo_Te,
oechem.OEElemNo_Cl: oechem.OEElemNo_I,
oechem.OEElemNo_P: oechem.OEElemNo_Sb,
oechem.OEElemNo_Br: 117,
oechem.OEElemNo_I: 118,
}
for ref_atom in atoms:
copy_atom = copy_mol.GetAtom(oechem.OEHasAtomIdx(ref_atom.GetIdx()))
if copy_atom is None:
raise Exception("Null atom found")
copy_atom.SetAtomicNum(atom_map[copy_atom.GetAtomicNum()])
return oechem.OECreateInChIKey(copy_mol)
def get_specific_dihedral_inchi_key(mol):
"""
generates unique dihedral inchi key by mutating all four dihedral atoms
"""
try:
a, b, c, d = get_torsion_oeatom_list(mol)
modified_inchi = get_modified_inchi_key(mol, [a, b, c, d])
inchiKey = oechem.OECreateInChIKey(mol) + modified_inchi
return inchiKey
except Exception as e:
logging.warning(e)
return None
def get_generic_dihedral_inchi_key(mol):
"""
generates dihedral inchi key after mutating only the central two atoms
"""
try:
_, b, c, _ = get_torsion_oeatom_list(mol)
modified_inchi = get_modified_inchi_key(mol, [b, c])
inchiKey = oechem.OECreateInChIKey(mol) + modified_inchi
return inchiKey
except Exception as e:
logging.warning(e)
return None
def __makeChemCompDescriptorCategory(self, ccId, oeMol):
"""
loop_
_pdbx_chem_comp_descriptor.comp_id
_pdbx_chem_comp_descriptor.type
_pdbx_chem_comp_descriptor.program
_pdbx_chem_comp_descriptor.program_version
_pdbx_chem_comp_descriptor.descriptor
ARG SMILES ACDLabs 10.04 "O=C(O)C(N)CCCNC(=[NH2+])N"
ARG SMILES_CANONICAL CACTVS 3.341 "N[C@@H](CCCNC(N)=[NH2+])C(O)=O"
ARG SMILES CACTVS 3.341 "N[CH](CCCNC(N)=[NH2+])C(O)=O"
ARG SMILES_CANONICAL "OpenEye OEToolkits" 1.5.0 "C(C[C@@H](C(=O)O)N)CNC(=[NH2+])N"
ARG SMILES "OpenEye OEToolkits" 1.5.0 "C(CC(C(=O)O)N)CNC(=[NH2+])N"
ARG InChI InChI 1.03 "InChI=1S/C6H14N4O2/c7-4(5(11)12)2-1-3-1..... "
ARG InChIKey InChI 1.03 ODKSFYDXXFIFQN-BYPYZUCNSA-O
#
"""
rowL = []
#
aRow = {}
aRow["comp_id"] = ccId
aRow["type"] = "SMILES_CANONICAL"
aRow["program"] = "OpenEye OEToolkits"
aRow["program_version"] = self.__oeVersion
aRow["descriptor"] = oechem.OECreateIsoSmiString(oeMol)
rowL.append(aRow)
#
aRow = {}
aRow["comp_id"] = ccId
aRow["type"] = "SMILES"
aRow["program"] = "OpenEye OEToolkits"
aRow["program_version"] = self.__oeVersion
aRow["descriptor"] = oechem.OECreateCanSmiString(oeMol)
rowL.append(aRow)
#
aRow = {}
aRow["comp_id"] = ccId
aRow["type"] = "InChI"
aRow["program"] = "OpenEye OEToolkits"
aRow["program_version"] = self.__oeVersion
aRow["descriptor"] = oechem.OECreateInChI(oeMol)
rowL.append(aRow)
#
aRow = {}
aRow["comp_id"] = ccId
aRow["type"] = "InChIKey"
aRow["program"] = "OpenEye OEToolkits"
aRow["program_version"] = self.__oeVersion
aRow["descriptor"] = oechem.OECreateInChIKey(oeMol)
rowL.append(aRow)
#
return rowL
def extract_molecule_torsion_data(parent_mol, frag_mols=None):
"""
extract dihedral angle associated with each torsion motif in the input molecule
Torsion motifs are represented using generic modified inchi (central two atoms)
and specific modified inchi (4 torsion atoms)
@param parent_mol:
@type parent_mol: oechem.OEGraphMol
@return: tuple(str, dict[str, list[float]])
"""
if frag_mols is None:
frag_mols = get_molecule_torsion_fragments(parent_mol)
torsion_data = collections.defaultdict(list)
for frag_mol in frag_mols:
inchi_key = oechem.OECreateInChIKey(frag_mol)
atom_map = get_fragment_to_parent_atom_mapping(parent_mol, frag_mol)
try:
_, b, c, _ = get_torsion_oeatom_list(frag_mol)
for a in b.GetAtoms(oechem.OEIsHeavy()):
for d in c.GetAtoms(oechem.OEIsHeavy()):
if a.GetIdx() == c.GetIdx() or d.GetIdx() == b.GetIdx():
continue
ap = atom_map[a]
bp = atom_map[b]
cp = atom_map[c]
dp = atom_map[d]
if (a.GetAtomicNum() == ap.GetAtomicNum()
and b.GetAtomicNum() == bp.GetAtomicNum()
and c.GetAtomicNum() == cp.GetAtomicNum()
and d.GetAtomicNum() == dp.GetAtomicNum()):
angle = (
oechem.OEGetTorsion(parent_mol, ap, bp, cp, dp) *
oechem.Rad2Deg)
torsion_inchi = inchi_key + get_modified_inchi_key(
frag_mol, [a, b, c, d])
torsion_data[torsion_inchi].append(
(ap.GetIdx(), bp.GetIdx(), cp.GetIdx(),
dp.GetIdx(), angle))
except Exception as e:
logging.warning(e)
continue
parent_inchi = get_modified_molecule_inchi(parent_mol)
return (parent_inchi, torsion_data)
def get_dihedral_inchi_key(mol):
try:
a, b, c, d = get_torsion_oeatom_list(mol)
ad = a.GetAtomicNum() * d.GetAtomicNum()
bc = b.GetAtomicNum() * c.GetAtomicNum()
adAro = int(a.IsAromatic()) + int(d.IsAromatic())
bcAro = int(b.IsAromatic()) + int(c.IsAromatic())
count1 = len(list(oechem.OEGetSubtree(b, c)))
count2 = len(list(oechem.OEGetSubtree(c, b)))
count = count1 * count2
inchiKey = oechem.OECreateInChIKey(mol)
inchiKey = inchiKey + str(ad) + str(bc) + str(adAro) + str(
bcAro) + str(count)
return inchiKey
except Exception as e:
logging.warning(e)
return None
def get_modified_molecule_inchi(mol):
title = mol.GetTitle().replace(",", "")
inchi = oechem.OECreateInChIKey(mol)
return inchi + title
def construct_dihedral_energy_profile(torsion_conformers, num_points=24):
angle_list = np.array([360 * i / num_points for i in range(num_points)])
num_confs = 0
profile = np.full(num_points, np.nan)
for mol in torsion_conformers:
if not mol:
continue
num_confs += 1
conf = mol.GetActive()
conf_title = get_sd_data(conf, "CONFORMER_LABEL")
tor_atoms = get_sd_data(mol, "TORSION_ATOMS_ParentMol").split()
parent_name = conf_title[:-3]
dih_label = "_".join(str(x) for x in tor_atoms)
fragment_label = parent_name + "_" + dih_label
angle_idx = int(conf_title[-2:])
profile[angle_idx] = np.float(get_sd_data(conf, "PSI4_ENERGY"))
logging.debug("angle_idx: %d", angle_idx)
logging.debug("Psi4 Energy: %f",
float(get_sd_data(conf, "PSI4_ENERGY")))
# check for angles where no energies are available
for angle in angle_list[np.all(np.isnan(profile))]:
logging.warning(
"Warning: No energies found for angle {:.1f} for fragment: {}".
format(angle, fragment_label))
# calculate relative energies
min_energy = np.nanmin(profile)
profile -= min_energy
profile[np.isnan(profile)] = -1 # set nans to -1
torsional_strain = np.column_stack((angle_list, profile))
# combine conformers
output_conformers = oechem.OEMol(torsion_conformers[0])
output_conformers.DeleteConfs()
title = fragment_label
output_conformers.SetTitle(title)
# setup normalization
torsion_tag = "TORSION_ATOMS_FRAGMENT"
torsion_atoms_in_fragment = get_sd_data(mol, torsion_tag).split()
print(torsion_atoms_in_fragment)
dihedral_atom_indices = [int(x) - 1 for x in torsion_atoms_in_fragment]
dih, _ = get_dihedral(output_conformers, dihedral_atom_indices)
for old_conf in torsion_conformers:
if old_conf:
new_conf = output_conformers.NewConf(old_conf)
normalize_coordinates(new_conf, dih)
oechem.OEClearSDData(new_conf)
for dp in oechem.OEGetSDDataPairs(old_conf.GetActive()):
if dp.GetTag() not in ["OEConfTitle", "CONFORMER_LABEL"]:
oechem.OESetSDData(new_conf, dp.GetTag(), dp.GetValue())
torsion_angle = get_sd_data(old_conf, "TORSION_ANGLE")
title = fragment_label + ": Angle " + torsion_angle
new_conf.SetTitle(title)
write_energy_profile_to_sddata(output_conformers, torsional_strain.copy())
# Calculate all possible torsion inchi keys for this fragment
torsion_inchi_list = []
inchi_key = oechem.OECreateInChIKey(output_conformers)
_, b, c, _ = get_torsion_oeatom_list(output_conformers)
for a in b.GetAtoms(oechem.OEIsHeavy()):
for d in c.GetAtoms(oechem.OEIsHeavy()):
if a.GetIdx() == c.GetIdx() or d.GetIdx() == b.GetIdx():
continue
torsion_inchi = inchi_key + get_modified_inchi_key(
output_conformers, [a, b, c, d])
torsion_inchi_list.append(torsion_inchi)
return output_conformers, torsional_strain, torsion_inchi_list