def generate_vacuum_hybrid_topology(mol_name="naphthalene", ref_mol_name="benzene"):
from topology_proposal import SmallMoleculeSetProposalEngine, TopologyProposal
import simtk.openmm.app as app
from openmoltools import forcefield_generators
from perses.tests.utils import createOEMolFromIUPAC, createSystemFromIUPAC, get_data_filename
m, unsolv_old_system, pos_old, top_old = createSystemFromIUPAC(mol_name)
refmol = createOEMolFromIUPAC(ref_mol_name)
initial_smiles = oechem.OEMolToSmiles(m)
final_smiles = oechem.OEMolToSmiles(refmol)
gaff_xml_filename = get_data_filename("data/gaff.xml")
forcefield = app.ForceField(gaff_xml_filename, 'tip3p.xml')
forcefield.registerTemplateGenerator(forcefield_generators.gaffTemplateGenerator)
solvated_system = forcefield.createSystem(top_old)
gaff_filename = get_data_filename('data/gaff.xml')
system_generator = SystemGenerator([gaff_filename, 'amber99sbildn.xml', 'tip3p.xml'])
geometry_engine = FFAllAngleGeometryEngine()
proposal_engine = SmallMoleculeSetProposalEngine(
[initial_smiles, final_smiles], system_generator, residue_name=mol_name)
#generate topology proposal
topology_proposal = proposal_engine.propose(solvated_system, top_old)
#generate new positions with geometry engine
new_positions, _ = geometry_engine.propose(topology_proposal, pos_old, beta)
return topology_proposal, pos_old, new_positions
def AtomPathLength(ifs, ofs, itf, atm1, atm2):
for mol in ifs.GetOEGraphMols():
oechem.OETriposAtomNames(mol)
a1 = None
a2 = None
for atm in mol.GetAtoms():
if atm.GetName() == atm1:
a1 = atm
if atm.GetName() == atm2:
a2 = atm
if a1 is not None and a2 is not None:
break
if a1 is None or a2 is None:
oechem.OEThrow.Warning(
"Failed to find atoms %s and %s in molecule" % (atm1, atm2))
continue
pathlen = oechem.OEGetPathLength(a1, a2)
if itf.GetBool("-verbose") or not itf.HasString("-o"):
print("Path length: %s in %s" %
(pathlen, oechem.OEMolToSmiles(mol)))
spath = oechem.OEShortestPath(a1, a2)
spathmol = oechem.OEGraphMol()
adjustHCount = True
oechem.OESubsetMol(spathmol, mol, oechem.OEIsAtomMember(spath),
adjustHCount)
spathsmiles = oechem.OEMolToSmiles(spathmol)
if itf.HasString("-o"):
oechem.OEWriteMolecule(ofs, spathmol)
elif itf.GetBool("-verbose"):
print(spathsmiles)
def extract_transformation(run, compound_microstates, project):
import perses
import openmoltools
import numpy as np
from openeye import oechem
from fah_xchem.schema import Transformation
try:
npz = np.load(f'{project}/RUNS/RUN{run}/htf.npz', allow_pickle=True)
x = npz['arr_0']
htf = x.item()
old_smiles = oechem.OEMolToSmiles(
htf._topology_proposal.old_topology.residue_oemol)
new_smiles = oechem.OEMolToSmiles(
htf._topology_proposal.new_topology.residue_oemol)
if (old_smiles not in compound_microstates):
print(f'{old_smiles} not found')
return None
elif (new_smiles not in compound_microstates):
print(f'{new_smiles} not found')
return None
#print(run, old_smiles, new_smiles, compound_microstates[old_smiles].microstate_id, compound_microstates[new_smiles].microstate_id)
transformation = Transformation(
run_id=run,
xchem_fragment_id=xchem_fragment_id,
initial_microstate=compound_microstates[old_smiles],
final_microstate=compound_microstates[new_smiles])
except Exception as e:
print(e)
return None
return transformation
def generate_solvated_hybrid_test_topology(current_mol_name="naphthalene", proposed_mol_name="benzene"):
"""
Generate a test solvated topology proposal, current positions, and new positions triplet
from two IUPAC molecule names.
Parameters
----------
current_mol_name : str, optional
name of the first molecule
proposed_mol_name : str, optional
name of the second molecule
Returns
-------
topology_proposal : perses.rjmc.topology_proposal
The topology proposal representing the transformation
current_positions : np.array, unit-bearing
The positions of the initial system
new_positions : np.array, unit-bearing
The positions of the new system
"""
import simtk.openmm.app as app
from openmoltools import forcefield_generators
from perses.tests.utils import createOEMolFromIUPAC, createSystemFromIUPAC, get_data_filename
current_mol, unsolv_old_system, pos_old, top_old = createSystemFromIUPAC(current_mol_name)
proposed_mol = createOEMolFromIUPAC(proposed_mol_name)
initial_smiles = oechem.OEMolToSmiles(current_mol)
final_smiles = oechem.OEMolToSmiles(proposed_mol)
gaff_xml_filename = get_data_filename("data/gaff.xml")
forcefield = app.ForceField(gaff_xml_filename, 'tip3p.xml')
forcefield.registerTemplateGenerator(forcefield_generators.gaffTemplateGenerator)
modeller = app.Modeller(top_old, pos_old)
modeller.addSolvent(forcefield, model='tip3p', padding=9.0*unit.angstrom)
solvated_topology = modeller.getTopology()
solvated_positions = modeller.getPositions()
solvated_system = forcefield.createSystem(solvated_topology, nonbondedMethod=app.PME, removeCMMotion=False)
barostat = openmm.MonteCarloBarostat(1.0*unit.atmosphere, temperature, 50)
solvated_system.addForce(barostat)
gaff_filename = get_data_filename('data/gaff.xml')
system_generator = SystemGenerator([gaff_filename, 'amber99sbildn.xml', 'tip3p.xml'], barostat=barostat, forcefield_kwargs={'removeCMMotion': False, 'nonbondedMethod': app.PME})
geometry_engine = geometry.FFAllAngleGeometryEngine()
proposal_engine = SmallMoleculeSetProposalEngine(
[initial_smiles, final_smiles], system_generator, residue_name=current_mol_name)
#generate topology proposal
topology_proposal = proposal_engine.propose(solvated_system, solvated_topology)
#generate new positions with geometry engine
new_positions, _ = geometry_engine.propose(topology_proposal, solvated_positions, beta)
return topology_proposal, solvated_positions, new_positions
def SmartsPathLength(ifs, ofs, itf, ss1, ss2):
for mol in ifs.GetOEGraphMols():
oechem.OEPrepareSearch(mol, ss1)
oechem.OEPrepareSearch(mol, ss2)
if not (ss1.SingleMatch(mol) and ss2.SingleMatch(mol)):
oechem.OEThrow.Warning(
"Unable to find SMARTS matches in %s, skipping" %
mol.GetTitle())
continue
unique = True
allminlen = sys.maxsize
for match1 in ss1.Match(mol, unique):
for match2 in ss2.Match(mol, unique):
minlen = sys.maxsize
for atom1 in match1.GetTargetAtoms():
for atom2 in match2.GetTargetAtoms():
pathlen = oechem.OEGetPathLength(atom1, atom2)
if minlen > pathlen:
minlen = pathlen
atompairs = []
atompairs.append([atom1, atom2])
elif minlen == pathlen:
atompairs.append([atom1, atom2])
if minlen < allminlen:
allminlen = minlen
allatompairs = atompairs[:]
elif minlen == allminlen:
allatompairs += atompairs[:]
if itf.GetBool("-verbose") or not itf.HasString("-o"):
print("Shortest path length: %s in %s" %
(allminlen, oechem.OEMolToSmiles(mol)))
spathlist = set()
for satom1, satom2, in allatompairs:
spath = oechem.OEShortestPath(satom1, satom2)
spathmol = oechem.OEGraphMol()
oechem.OESubsetMol(spathmol, mol, oechem.OEIsAtomMember(spath))
spathsmiles = oechem.OEMolToSmiles(spathmol)
if spathsmiles in spathlist:
continue
spathlist.add(spathsmiles)
if itf.HasString("-o"):
oechem.OEWriteMolecule(ofs, spathmol)
elif itf.GetBool("-verbose"):
print(spathsmiles)
return
def test_remove_atom_map():
from openeye import oechem
mapped_smiles = '[H:5][C:1]([H:6])([C:2]([H:7])([H:8])[O:4][H:10])[O:3][H:9]'
mapped_mol = oechem.OEMol()
oechem.OESmilesToMol(mapped_mol, mapped_smiles)
chemi.remove_map(mapped_mol)
assert oechem.OEMolToSmiles(mapped_mol) == 'C(CO)O'
chemi.restore_map(mapped_mol)
assert oechem.OEMolToSmiles(mapped_mol) == mapped_smiles
def generate_poses(receptor, refmol, target_molecules, output_filename):
"""
Parameters
----------
receptor : openeye.oechem.OEGraphMol
Receptor (already prepped for docking) for identifying optimal pose
refmol : openeye.oechem.OEGraphMol
Reference molecule which shares some part in common with the proposed molecule
target_molecules : list of OEMol
List of molecules to build
output_filename : str
Output filename for generated conformers
"""
# Expand uncertain stereochemistry
print('Expanding uncertain stereochemistry...')
target_molecules = expand_stereochemistry(target_molecules)
print(f' There are {len(target_molecules)} target molecules')
# TODO: Expand protonation states
# Identify optimal conformer for each molecule
with oechem.oemolostream(output_filename) as ofs:
from rich.progress import track
from multiprocessing import Pool
from tqdm import tqdm
pool = Pool()
args = [(receptor, refmol, mol) for mol in target_molecules]
for pose in track(pool.imap_unordered(
generate_restricted_conformers_star, args),
total=len(args),
description='Enumerating conformers...'):
#for pose in map(generate_restricted_conformers_star, args): # DEBUG
if pose is not None:
# DEBUG
if 'EDJ-MED-e4b030d8-2' in pose.GetTitle():
msg = 'Writing docked pose for: '
msg += pose.GetTitle() + '\n'
msg += f'{"":5s} ' + oechem.OEMolToSmiles(pose) + '\n'
print(msg)
oechem.OEWriteMolecule(ofs, pose)
if 'EDJ-MED-e4b030d8-2' in pose.GetTitle():
msg = 'Wrote docked pose for: '
msg += pose.GetTitle() + '\n'
msg += f'{"":5s} ' + oechem.OEMolToSmiles(pose) + '\n'
print(msg)
pool.close()
pool.join()
def test_keep_track_of_map():
from openeye import oechem
smiles = 'c1ccc(cc1)Nc2ncccn2'
mapped_mol = oechem.OEMol()
oechem.OESmilesToMol(mapped_mol, smiles)
frags = fragmenter.fragment.CombinatorialFragmenter(mapped_mol)
frags.fragment()
#frags._fragment_all_bonds_not_in_ring_systems()
#frags._combine_fragments(min_rotors=1, max_rotors=frags.n_rotors+1, restore_maps=True)
keys = list(frags.fragments.keys())
assert oechem.OEMolToSmiles(frags.fragments[keys[0]][0]) == '[H:14][c:1]1[c:2]([c:4]([c:9]([c:5]([c:3]1[H:16])[H:18])[NH:13][H:22])[H:17])[H:15]'
assert oechem.OEMolToSmiles(frags.fragments[keys[1]][0]) == '[H:19][c:6]1[c:7]([n:11][c:10]([n:12][c:8]1[H:21])[NH:13][H:22])[H:20]'
def test_smiles_to_oemol():
from openeye import oechem
mol = chemi.smiles_to_oemol('CCCC')
assert isinstance(mol, oechem.OEMol)
assert oechem.OEMolToSmiles(mol) == 'CCCC'
assert mol.GetTitle() == 'butane'
mol = chemi.smiles_to_oemol('CCCC', normalize=False)
assert mol.GetTitle() == ''
mol = chemi.smiles_to_oemol('CCCC', add_atom_map=True)
assert oechem.OEMolToSmiles(
mol
) == '[H:5][C:1]([H:6])([H:7])[C:3]([H:11])([H:12])[C:4]([H:13])([H:14])[C:2]([H:8])([H:9])[H:10]'
def generate_vacuum_topology_proposal(current_mol_name="benzene", proposed_mol_name="toluene"):
"""
Generate a test vacuum topology proposal, current positions, and new positions triplet
from two IUPAC molecule names.
Parameters
----------
current_mol_name : str, optional
name of the first molecule
proposed_mol_name : str, optional
name of the second molecule
Returns
-------
topology_proposal : perses.rjmc.topology_proposal
The topology proposal representing the transformation
current_positions : np.array, unit-bearing
The positions of the initial system
new_positions : np.array, unit-bearing
The positions of the new system
"""
from openmoltools import forcefield_generators
from perses.tests.utils import createOEMolFromIUPAC, createSystemFromIUPAC, get_data_filename
current_mol, unsolv_old_system, pos_old, top_old = createSystemFromIUPAC(current_mol_name)
proposed_mol = createOEMolFromIUPAC(proposed_mol_name)
initial_smiles = oechem.OEMolToSmiles(current_mol)
final_smiles = oechem.OEMolToSmiles(proposed_mol)
gaff_xml_filename = get_data_filename("data/gaff.xml")
forcefield = app.ForceField(gaff_xml_filename, 'tip3p.xml')
forcefield.registerTemplateGenerator(forcefield_generators.gaffTemplateGenerator)
solvated_system = forcefield.createSystem(top_old, removeCMMotion=False)
gaff_filename = get_data_filename('data/gaff.xml')
system_generator = SystemGenerator([gaff_filename, 'amber99sbildn.xml', 'tip3p.xml'], forcefield_kwargs={'removeCMMotion': False, 'nonbondedMethod': app.NoCutoff})
geometry_engine = geometry.FFAllAngleGeometryEngine()
proposal_engine = SmallMoleculeSetProposalEngine(
[initial_smiles, final_smiles], system_generator, residue_name=current_mol_name)
#generate topology proposal
topology_proposal = proposal_engine.propose(solvated_system, top_old, current_mol=current_mol, proposed_mol=proposed_mol)
#generate new positions with geometry engine
new_positions, _ = geometry_engine.propose(topology_proposal, pos_old, beta)
return topology_proposal, pos_old, new_positions
def expand_stereochemistry(mols):
"""Expand stereochemistry when uncertain
Parameters
----------
mols : openeye.oechem.OEGraphMol
Molecules to be expanded
Returns
-------
expanded_mols : openeye.oechem.OEMol
Expanded molecules
"""
expanded_mols = list()
from openeye import oechem, oeomega
omegaOpts = oeomega.OEOmegaOptions()
omega = oeomega.OEOmega(omegaOpts)
maxcenters = 12
forceFlip = False
enumNitrogen = False
warts = True # add suffix for stereoisomers
for mol in mols:
compound_title = mol.GetTitle()
compound_smiles = oechem.OEMolToSmiles(mol)
enantiomers = list()
for enantiomer in oeomega.OEFlipper(mol, maxcenters, forceFlip,
enumNitrogen, warts):
enantiomer = oechem.OEMol(enantiomer)
enantiomer_smiles = oechem.OEMolToSmiles(enantiomer)
oechem.OESetSDData(enantiomer, 'compound', compound_title)
oechem.OESetSDData(enantiomer, 'compound_smiles', compound_smiles)
oechem.OESetSDData(enantiomer, 'enantiomer_smiles',
enantiomer_smiles)
enantiomers.append(enantiomer)
expanded_mols += enantiomers
# DEBUG
if 'EDJ-MED-e4b030d8-2' in mol.GetTitle():
msg = 'Enumerated microstates for compound: '
msg += mol.GetTitle() + '\n'
msg += f'{"":5s} ' + oechem.OEMolToSmiles(mol) + '\n'
for index, m in enumerate(enantiomers):
msg += f'{index:5d} : ' + oechem.OEMolToSmiles(m) + '\n'
print(msg)
return expanded_mols
def calculate_t142_central_wbo(mol: oechem.OEMol,
params: Dict[str, List[List[int]]]) -> float:
"""Calculates the WBO between the central atoms in the t142 param in the molecule.
(WBO is Wiberg Bond Order.)
The `params` argument contains the parameters of the molecule (see
`calculate_mol_params`).
Returns -1 if the calculation fails.
"""
# Only use first occurrence of the parameter.
indices = params['t142'][0]
# For torsion parameters such as t142, the central atoms should be at the
# second and third index.
central_indices = [indices[1], indices[2]]
# Generate molecule conformer.
oechem.OEAddExplicitHydrogens(mol)
omega = oeomega.OEOmega()
omega.SetMaxConfs(1)
omega.SetCanonOrder(False)
omega.SetSampleHydrogens(True)
omega.SetEnergyWindow(15.0) #unit?
omega.SetRMSThreshold(1.0)
# Don't generate random stereoisomer if not specified.
omega.SetStrictStereo(True)
status = omega(mol)
if status is False:
omega.SetStrictStereo(False)
new_status = omega(mol)
if new_status is False:
logger.error("Failed to generate conformer for %s",
oechem.OEMolToSmiles(mol))
return -1
# Calculate the WBO between the two central atoms.
conf = next(iter(mol.GetConfs()))
charged_copy = oechem.OEMol(conf)
results = oequacpac.OEAM1Results()
if not AM1_CALCULATOR.CalcAM1(results, charged_copy):
logger.error("Failed to assign partial charges to %s",
oechem.OEMolToSmiles(mol))
return -2
return results.GetBondOrder(central_indices[0], central_indices[1])
def gen_canonical_isomeric_smiles(oemol):
# 1. Create an OpenFF molecule from the OpenEye molecule, guessing the
# stereochemistry if needed.
oe_molecule = oechem.OEMol(oemol)
try:
molecule = Molecule.from_openeye(oe_molecule)
except:
molecule = Molecule.from_openeye(oe_molecule,
allow_undefined_stereo=True)
stereoisomers = molecule.enumerate_stereoisomers(undefined_only=True,
max_isomers=1)
if len(stereoisomers) > 0:
molecule = stereoisomers[0]
# 2. Canonically order the molecule
molecule = molecule.canonical_order_atoms()
# 3. Figure out which atoms in the canonical molecule should be tagged.
mapped_smiles = oechem.OEMolToSmiles(oe_molecule)
torsion_match = molecule.chemical_environment_matches(mapped_smiles)[0]
# 4. Generate a canonical isomeric mapped smiles
molecule.properties["atom_map"] = {
j: i + 1
for i, j in enumerate(torsion_match)
}
center_bond = set(molecule.properties["atom_map"].keys())
canonical_isomeric_smiles = molecule.to_smiles(isomeric=True,
explicit_hydrogens=True,
mapped=False)
return molecule, canonical_isomeric_smiles, center_bond
def get_mols_from_frags(this_smiles, old_smiles=None):
if old_smiles is None:
old_smiles = []
fragfunc = GetFragmentationFunction()
mol = oechem.OEGraphMol()
oechem.OESmilesToMol(mol, this_smiles)
frags = [f for f in fragfunc(mol)]
len_frags = len(frags)
for smile in old_smiles:
mol2 = oechem.OEGraphMol()
oechem.OESmilesToMol(mol2, smile)
frags += [f for f in fragfunc(mol2)]
oechem.OEThrow.Info("%d number of fragments generated" % len(frags))
fragcombs = GetFragmentCombinations(mol, frags, frag_number=len_frags)
oechem.OEThrow.Info("%d number of fragment combinations generated" %
len(fragcombs))
smiles = set()
for frag in fragcombs:
if oechem.OEDetermineComponents(frag)[0] == 1:
smiles = smiles.union(oechem.OEMolToSmiles(frag))
return smiles
def test_n_valence():
from openeye import oechem
json_molecule = {
'symbols': [
'C', 'C', 'C', 'C', 'C', 'C', 'N', 'N', 'N', 'N', 'N', 'H', 'H',
'H', 'H'
],
'geometry':
np.array([
11.02088236, 0.30802536, 2.96687012, 10.37270642, 2.8383686,
2.75522059, 9.32012957, -1.48532476, 2.09948562, 8.06346176,
3.48843435, 1.68941515, 6.98820713, -0.772898, 1.02801107,
5.21186447, -2.73065435, 0.12850138, 5.70508328, -5.1797392,
0.28345893, 6.45152507, 1.7536658, 0.86549457, 2.97820833,
-2.31491455, -0.90706852, 3.71709131, -6.31357514, -0.68408084,
2.05980154, -4.57124733, -1.40784597, 12.76887939, -0.24566439,
3.77189345, 11.61992628, 4.26322222, 3.39583795, 9.76610505,
-3.43174262, 2.23743576, 7.53811768, 5.41217579, 1.50989122
]),
'connectivity':
[[0, 1, 1], [0, 2, 2], [0, 11, 1], [1, 3, 2], [1, 12, 1], [2, 4, 1],
[2, 13, 1], [3, 7, 1], [3, 14, 1], [4, 5, 1], [4, 7, 2], [5, 6, 1],
[5, 8, 2], [6, 9, 1], [8, 10, 1], [9, 10, 2]]
}
mol = utils.load_molecule(json_molecule)
assert utils.has_explicit_hydrogen(mol)
assert oechem.OEMolToSmiles(mol) == 'c1ccnc(c1)c2[n-]nnn2'
def create_mapped_smiles(mol):
"""
Generate an index-tagged explicit hydrogen SMILES.
Exmaple:
SMILES string for carbon monoxide "CO"
With index-tagged explicit hydrogen SMILES this becomes
'[H:3][C:1]([H:4])([H:5])[O:2][H:6]'
Parameters
----------
mol: OEMOl
Returns
-------
index-tagged explicit hydrogen SMILES str
"""
# Check if molecule already has explicit hydrogens
HAS_HYDROGENS = oechem.OEHasExplicitHydrogens(mol)
if not HAS_HYDROGENS:
# Add explicit hydrogens
oechem.OEAddExplicitHydrogens(mol)
for atom in mol.GetAtoms():
atom.SetMapIdx(atom.GetIdx() + 1)
return oechem.OEMolToSmiles(mol)
def convert_extension(infile, outfile, canonical=False):
"""
Convert one molecule file format into another using OpenEye tools.
The user may also assign canonical smiles as name before writing output.
"""
# open input file
mols = reader.read_mols(infile)
# open output file
ofs = oechem.oemolostream()
if not ofs.open(outfile):
oechem.OEThrow.Fatal("Unable to open %s for writing" % outfile)
# write to output
for mol in mols:
if canonical:
smi = oechem.OEMolToSmiles(mol)
for conf in mol.GetConfs():
if canonical:
conf.SetTitle(smi)
oechem.OEWriteConstMolecule(ofs, conf)
# close filestreams
ofs.close()
def test_small_molecule_proposals():
"""
Make sure the small molecule proposal engine generates molecules
"""
list_of_smiles = ['CCCC','CCCCC','CCCCCC']
list_of_mols = []
for smi in list_of_smiles:
mol = smiles_to_oemol(smi)
list_of_mols.append(mol)
molecules = [Molecule.from_openeye(mol) for mol in list_of_mols]
stats_dict = defaultdict(lambda: 0)
system_generator = SystemGenerator(forcefields = forcefield_files, barostat=barostat, forcefield_kwargs=forcefield_kwargs, nonperiodic_forcefield_kwargs=nonperiodic_forcefield_kwargs,
small_molecule_forcefield = small_molecule_forcefield, molecules=molecules, cache=None)
proposal_engine = topology_proposal.SmallMoleculeSetProposalEngine(list_of_mols, system_generator)
initial_system, initial_positions, initial_topology, = OEMol_to_omm_ff(list_of_mols[0], system_generator)
proposal = proposal_engine.propose(initial_system, initial_topology)
for i in range(50):
#positions are ignored here, and we don't want to run the geometry engine
new_proposal = proposal_engine.propose(proposal.old_system, proposal.old_topology)
stats_dict[new_proposal.new_chemical_state_key] += 1
#check that the molecule it generated is actually the smiles we expect
matching_molecules = [res for res in proposal.new_topology.residues() if res.name=='MOL']
if len(matching_molecules) != 1:
raise ValueError("More than one residue with the same name!")
mol_res = matching_molecules[0]
oemol = generateOEMolFromTopologyResidue(mol_res)
smiles = SmallMoleculeSetProposalEngine.canonicalize_smiles(oechem.OEMolToSmiles(oemol))
assert smiles == proposal.new_chemical_state_key
proposal = new_proposal
def get_result(self, query_id):
cur_rank = list()
url = self.args.url + "/queries/{}/".format(query_id)
response = None
tries = 0
while response == None or data["status"]["job"] != "COMPLETED":
time.sleep(60 * tries)
tries += 1
response = requests.get(url)
data = response.json()
results_url = data["results"]
results_data = requests.get(self.args.url + results_url)
with tempfile.NamedTemporaryFile(suffix='.oeb',
mode='wb',
delete=False) as temp:
temp.write(results_data.content)
temp.flush()
with oechem.oemolistream(temp.name) as results:
for mol in results.GetOEGraphMols():
cur_rank.append(
(oechem.OEMolToSmiles(mol), mol.GetTitle(),
float(oechem.OEGetSDData(mol, 'TanimotoCombo')),
self.baitset[0], False))
os.remove(temp.name)
return cur_rank
def update_ranking(self, mol, max_tanimoto, ka_tag):
index = 0
if len(self.ranking) >= self.args.topn and max_tanimoto < self.ranking[
len(self.ranking) - 1][2]:
pass
else:
for top_mol in self.ranking:
if max_tanimoto < top_mol[2]:
index = self.ranking.index(top_mol) + 1
else:
break
upper = self.ranking[:index]
lower = self.ranking[index:]
self.ranking = upper + [(oechem.OEMolToSmiles(mol), mol.GetTitle(),
max_tanimoto, self.baitset[0], ka_tag)
] + lower
i = self.args.topn - 1
while i < len(self.ranking) - 1:
if self.ranking[i][2] != self.ranking[i + 1][2]:
self.ranking = self.ranking[:i + 1]
break
else:
i += 1
def test_select_with_longer_fingerprints(pipeline_test_files):
(smiles_file, filter_output_oeb, fingerprint_output_oeb, smiles_dataset_file,
sorted_by_fingerprint_oeb) = pipeline_test_files
smiles_file.write_text("\n".join(TEST_SMILES))
dp = DancePipeline("SMILES", smiles_file)
dp.filter(_relevant_always, filter_output_oeb)
# Each fingerprint consists of (1, 1, num_atoms), to ensure that fingerprints
# are sorted correctly when the fingerprint is longer.
dp.assign_fingerprint(lambda mol: (1, 1, mol.NumAtoms()), fingerprint_output_oeb)
dp.select(3, "SMILES", smiles_dataset_file, sorted_by_fingerprint_oeb, in_memory_sorting_threshold=3)
# Check that the molecules are sorted by fingerprint.
outputted_smiles = \
[oechem.OEMolToSmiles(mol) for mol in utils.get_mols_from_oeb(sorted_by_fingerprint_oeb)]
assert outputted_smiles == \
utils.get_list_of_canonical_isomeric_smiles(["N", "N#N", "C#N", "O=C=O"]) or \
outputted_smiles == \
utils.get_list_of_canonical_isomeric_smiles(["N", "C#N", "N#N", "O=C=O"])
# Check that the correct molecules were selected.
utils.assert_smiles_in_file_are_equal(smiles_dataset_file,
utils.get_list_of_canonical_isomeric_smiles(["N", "O=C=O"]))
def test_small_molecule_proposals():
"""
Make sure the small molecule proposal engine generates molecules
"""
from perses.rjmc import topology_proposal
from openmoltools import forcefield_generators
import openeye.oechem as oechem
list_of_smiles = ['CCCC','CCCCC','CCCCCC']
gaff_xml_filename = get_data_filename('data/gaff.xml')
stats_dict = {smiles : 0 for smiles in list_of_smiles}
system_generator = topology_proposal.SystemGenerator([gaff_xml_filename])
proposal_engine = topology_proposal.SmallMoleculeSetProposalEngine(list_of_smiles, system_generator)
initial_molecule = generate_initial_molecule('CCCC')
initial_system, initial_positions, initial_topology = oemol_to_omm_ff(initial_molecule, "MOL")
proposal = proposal_engine.propose(initial_system, initial_topology)
for i in range(50):
#positions are ignored here, and we don't want to run the geometry engine
new_proposal = proposal_engine.propose(proposal.old_system, proposal.old_topology)
stats_dict[new_proposal.new_chemical_state_key] += 1
#check that the molecule it generated is actually the smiles we expect
matching_molecules = [res for res in proposal.new_topology.residues() if res.name=='MOL']
if len(matching_molecules) != 1:
raise ValueError("More than one residue with the same name!")
mol_res = matching_molecules[0]
oemol = forcefield_generators.generateOEMolFromTopologyResidue(mol_res)
assert oechem.OEMolToSmiles(oemol) == proposal.new_chemical_state_key
proposal = new_proposal
def UniqMol(ifs, ofs):
smiles = {}
for mol in ifs.GetOEMols():
smi = oechem.OEMolToSmiles(mol)
if smi not in smiles:
smiles[smi] = True
oechem.OEWriteMolecule(ofs, mol)
def xyz_to_smiles(filename):
"""
Parse an .xyz file, heuristically perceive chemistry, and return canonical isomeric SMILES
"""
from openeye import oechem
ifs = oechem.oemolistream(filename)
for oemol in ifs.GetOEGraphMols():
return oechem.OEMolToSmiles(oemol)
def test_build_fragment():
from openeye import oechem
smiles = 'CCCCCC'
mol = chemi.smiles_to_oemol(smiles)
f = fragmenter.fragment.WBOFragmenter(mol)
f.calculate_wbo()
f._get_rotor_wbo()
setattr(f, 'threshold', 0.05)
for bond in f.rotors_wbo:
f._build_fragment(bond)
assert len(f.fragments) == 3
remove_atom_map(f.fragments[(3, 5)])
assert oechem.OEMolToSmiles(f.fragments[(3, 5)]) == 'CCCCC'
remove_atom_map(f.fragments[(4, 6)])
assert oechem.OEMolToSmiles(f.fragments[(4, 6)]) == 'CCCCC'
remove_atom_map(f.fragments[(5, 6)])
assert oechem.OEMolToSmiles(f.fragments[(5, 6)]) == 'CCCCCC'
def test_selects_just_one_smallest_molecule(self, tmpdir):
output_file = str(tmpdir / "select-final.smi")
self.create_smiles_in_tmpdir(tmpdir)
self.invoke_select_final(1, str(tmpdir), output_file)
ifs = oechem.oemolistream(output_file)
mol = oechem.OEMol()
while oechem.OEReadMolecule(ifs, mol):
assert oechem.OEMolToSmiles(mol) == "N"
def get_smiles(self):
"""
Returns
-------
smiles: str
SMILES string for the molecule
"""
smiles = oechem.OEMolToSmiles(self.mol)
return smiles
def test_load_molecule(toolkit):
"""Test load molecules"""
mol = utils.load_molecule(
'[H]C([H])([H])C([H])([H])C([H])([H])C([H])([H])[H]', toolkit=toolkit)
if toolkit == 'openeye':
from openeye import oechem
assert oechem.OEMolToSmiles(mol) == 'CCCC'
if toolkit == 'rdkit':
from rdkit import Chem
assert Chem.MolToSmiles(mol) == 'CCCC'
def get_new_action_set(self, aligner=None):
with self.logger("get_new_action_set") as logger:
if aligner is not None:
self.set_mole_aligner(aligner)
mols = self.fastrocs_query(self.mol_aligner,
self.config.space_size,
self.config.host)
# mols = [self.mol_aligner_conformers.from_oemol(mol) for mol in mols]
# mols = list(filter(lambda x : x is not None, mols))
smiles = [oechem.OEMolToSmiles(mol) for mol in mols]
return mols, smiles
def _oe_stream_from_file(file_path: str, as_smiles=False): # pragma: no cover
from openeye import oechem
from openff.toolkit.topology import Molecule
input_molecule_stream = oechem.oemolistream()
input_molecule_stream.open(file_path)
for oe_molecule in input_molecule_stream.GetOEMols():
yield (oechem.OEMolToSmiles(oe_molecule) if as_smiles else
Molecule.from_openeye(oe_molecule, allow_undefined_stereo=True))
