def align2d(file1, file2):
atomexpr = oechem.OEExprOpts_AtomicNumber | oechem.OEExprOpts_RingMember
bondexpr = oechem.OEExprOpts_RingMember
ifs1 = oechem.oemolistream(file1)
ifs2 = oechem.oemolistream(file2)
ifs1.SetConfTest(oechem.OEAbsCanonicalConfTest())
ifs2.SetConfTest(oechem.OEAbsCanonicalConfTest())
popts, dopts, report = prep_pdf_writer()
for mol1, mol2 in zip(ifs1.GetOEMols(), ifs2.GetOEMols()):
oechem.OESuppressHydrogens(mol1)
oechem.OESuppressHydrogens(mol2)
oechem.OEGenerate2DCoordinates(mol2)
ss = oechem.OESubSearch(mol2, atomexpr, bondexpr)
oechem.OEPrepareSearch(mol1, ss)
alignres = oedepict.OEPrepareAlignedDepiction(mol1, ss)
if not alignres.IsValid():
oechem.OEThrow.Error(
"Substructure is not found in input molecule!")
cell1 = report.NewCell()
cell2 = report.NewCell()
oedepict.OEPrepareDepiction(mol1, popts)
oedepict.OEPrepareDepiction(mol2, popts)
disp1 = oedepict.OE2DMolDisplay(mol1, dopts)
disp2 = oedepict.OE2DMolDisplay(mol2, dopts)
oedepict.OERenderMolecule(cell1, disp1)
oedepict.OERenderMolecule(cell2, disp2)
ofs = oechem.oeofstream()
if not ofs.open('output.pdf'):
oechem.OEThrow.Fatal("Cannot open output file!")
oedepict.OEWriteReport(ofs, "pdf", report)
def find_string_tag(infile):
# read input file
ifs = oechem.oemolistream()
ifs.SetConfTest(oechem.OEAbsCanonicalConfTest())
if not ifs.open(infile):
oechem.OEThrow.Warning("Unable to open input file for reading")
# loop through and evaluate tags
for mol in ifs.GetOEMols():
for conf in mol.GetConfs():
mytag = oechem.OEGetSDData(conf, 'SMILES QCArchive')
count1 = mytag.count('S')
count2 = mytag.count('P')
count3 = mytag.count('C#N')
count4 = mytag.count('N/N')
print(f"{conf.GetTitle()}\t{count1}\t{count2}\t{count3}\t{count4}")
def main(infile):
# open multi-molecule, multi-conformer file
ifs = oechem.oemolistream()
ifs.SetConfTest(oechem.OEAbsCanonicalConfTest())
if not ifs.open(infile):
raise FileNotFoundError(f"Unable to open {infile} for reading")
mols = ifs.GetOEMols()
for i, mol in enumerate(mols):
# perceive stereochemistry for mol
oechem.OEPerceiveChiral(mol)
oechem.OEAssignAromaticFlags(mol, oechem.OEAroModel_MDL)
# assign charges to copy of mol
# note that chg_mol does NOT have conformers
try:
chg_mol = charge_mol(mol)
except RuntimeError:
# perceive stereochem
#find_unspecified_stereochem(mol)
oechem.OE3DToInternalStereo(mol)
# reset perceived and call OE3DToBondStereo, since it may be missed
# by OE3DToInternalStereo if it thinks mol is flat
mol.ResetPerceived()
oechem.OE3DToBondStereo(mol)
try:
chg_mol = charge_mol(mol)
print(f'fixed stereo: {mol.GetTitle()}')
except RuntimeError:
find_unspecified_stereochem(mol)
title = mol.GetTitle()
smilabel = oechem.OEGetSDData(mol, "SMILES QCArchive")
print(' >>> Charge assignment failed due to unspecified '
f'stereochemistry {title} {smilabel}')
continue
def main(infile, ffxml):
# open multi-molecule, multi-conformer file
ifs = oechem.oemolistream()
ifs.SetConfTest(oechem.OEAbsCanonicalConfTest())
if not ifs.open(infile):
raise FileNotFoundError(f"Unable to open {infile} for reading")
mols = ifs.GetOEMols()
for i, mol in enumerate(mols):
# perceive stereochemistry for mol
oechem.OEPerceiveChiral(mol)
oechem.OEAssignAromaticFlags(mol, oechem.OEAroModel_MDL)
for j, conf in enumerate(mol.GetConfs()):
# perceive sterochemistry for conf coordinates
oechem.OE3DToInternalStereo(conf)
min_ffxml(conf, ffxml)
def master():
mol = oechem.OEMol()
ifs = oechem.oemolistream(input_smiles_file)
ifs.SetConfTest(oechem.OEAbsCanonicalConfTest())
for pos, mol in enumerate(ifs.GetOEMols()):
smiles = oechem.OEMol(mol)
ligand_name = smiles.GetTitle()
status = MPI.Status()
comm.recv(source=MPI.ANY_SOURCE, tag=WORKTAG, status=status)
rank_from = status.Get_source()
data = (pos, smiles, ligand_name)
comm.send(data, dest=rank_from, tag=WORKTAG)
if args.v == 1 and pos % 1000 == 0:
print("sent", pos, "jobs", flush=True)
for i in range(1, world_size):
comm.send([], dest=i, tag=DIETAG)
comm.Barrier()
def read_mols(in_file, mol_slice=None):
"""
Open a molecule file and return molecules and conformers as OEMols.
Provide option to slice the mols to return only a chunk from the
specified indices.
Parameters
----------
in_file : string
name of input file with molecules
mol_slice : numpy slice object
The resulting integers are numerically sorted and duplicates removed.
e.g., slices = np.s_[0, 3:5, 6::3] would be parsed to return
[0, 3, 4, 6, 9, 12, 15, 18, ...]
Can also parse from end: [-3:] gets the last 3 molecules, and
[-2:-1] is the same as [-2] to get just next to last molecule.
Returns
-------
mols : OEMols
"""
def flatten(x):
# https://stackoverflow.com/questions/2158395/flatten-an-irregular-list-of-lists
if isinstance(x, collections.Iterable):
return [a for i in x for a in flatten(i)]
else:
return [x]
ifs = oechem.oemolistream()
ifs.SetConfTest(oechem.OEAbsCanonicalConfTest())
if not ifs.open(in_file):
raise FileNotFoundError(f"Unable to open {in_file} for reading")
mols = ifs.GetOEMols()
if mol_slice is None:
return mols
# set max number of molecules for decoding slices
# TODO: how to get num_mols without re-reading file and loading all mols
ifs2 = oechem.oemolistream()
ifs2.SetConfTest(oechem.OEAbsCanonicalConfTest())
ifs2.open(in_file)
mols2 = ifs2.GetOEMols()
num_mols = len(list(mols2))
# parse mol_slice for multiple slice definitions provided
# e.g., (1, 4, 8) for second, fifth, and ninth molecules
# e.g., (0, slice(3, 5, None), slice(6, None, 3)) for example in docs
if isinstance(mol_slice, tuple) or isinstance(mol_slice, list):
idx_to_keep = []
for s in mol_slice:
# parse the slice object
if isinstance(s, slice):
idx_to_keep.append(list(range(num_mols))[s])
# else decode the negative int to positive int
elif isinstance(s, int) and s < 0:
idx_to_keep.append(s + num_mols)
# else just append the positive int
elif isinstance(s, int):
idx_to_keep.append(s)
else:
raise ValueError(
f"ERROR in parsing 'mol_slice' from {mol_slice}"
f" due to {s} being neither slice nor int")
# flatten to 1d, use set to remove duplicates, then sort list
idx_to_keep = list(set(flatten(idx_to_keep)))
idx_to_keep.sort()
#print(idx_to_keep)
elif isinstance(mol_slice, slice):
# parse the slice object
idx_to_keep = list(range(num_mols))[mol_slice]
# else just store the single value in a list
elif isinstance(mol_slice, int):
if mol_slice < 0:
mol_slice = mol_slice + num_mols
idx_to_keep = list(mol_slice)
else:
raise ValueError(f"ERROR in parsing 'mol_slice' from {mol_slice}")
# go through the generator and retrive the specified slices
mlist = []
for i, m in enumerate(mols):
if i in idx_to_keep:
# append a copy else still linked to orig generator
mlist.append(copy.copy(m))
# if this index is the last one in idx_to_keep, finish now
if i == idx_to_keep[-1]:
return mlist
return mlist
def main(infile, outfile, ffxml, minimizer):
# open multi-molecule, multi-conformer file
ifs = oechem.oemolistream()
ifs.SetConfTest(oechem.OEAbsCanonicalConfTest())
if not ifs.open(infile):
raise FileNotFoundError(f"Unable to open {infile} for reading")
mols = ifs.GetOEMols()
# open an outstream file
ofs = oechem.oemolostream()
if os.path.exists(outfile):
raise FileExistsError("Output file {} already exists in {}".format(
outfile, os.getcwd()))
if not ofs.open(outfile):
oechem.OEThrow.Fatal("Unable to open %s for writing" % outfile)
# minimize with openforcefield ffxml file
for i, mol in enumerate(mols):
# perceive stereochemistry for mol
oechem.OEPerceiveChiral(mol)
oechem.OEAssignAromaticFlags(mol, oechem.OEAroModel_MDL)
# assign charges to copy of mol
# note that chg_mol does NOT have conformers
try:
chg_mol = charge_mol(mol)
except RuntimeError:
# perceive stereochem
#find_unspecified_stereochem(mol)
oechem.OE3DToInternalStereo(mol)
# reset perceived and call OE3DToBondStereo, since it may be missed
# by OE3DToInternalStereo if it thinks mol is flat
mol.ResetPerceived()
oechem.OE3DToBondStereo(mol)
try:
chg_mol = charge_mol(mol)
print(f'fixed stereo: {mol.GetTitle()}')
except RuntimeError:
title = mol.GetTitle()
smilabel = oechem.OEGetSDData(mol, "SMILES QCArchive")
print(' >>> Charge assignment failed due to unspecified '
f'stereochemistry {title} {smilabel}')
continue
for j, conf in enumerate(mol.GetConfs()):
# perceive sterochemistry for conf coordinates
oechem.OE3DToInternalStereo(conf)
# assign charges to the conf itself
chg_conf = charge_conf(chg_mol, conf)
if minimizer == 'ffxml':
# minimize with parsley (charges set by ff not used from conf)
min_ffxml(chg_conf, ofs, ffxml)
if minimizer == 'mmff94':
# minimize with mmff94
min_mmff94x(chg_conf, ofs, mmff94s=False)
if minimizer == 'mmff94s':
# minimize with mmff94S
min_mmff94x(chg_conf, ofs, mmff94s=True)
if minimizer == 'gaff':
# minimize with gaff
min_gaffx(chg_conf, ofs, gaff2=False)
if minimizer == 'gaff2':
# minimize with gaff2
min_gaffx(chg_conf, ofs, gaff2=True)
ifs.close()
ofs.close()