def enumerate_conformations(name, smiles):
"""Generate geometry and run epik."""
# Generate molecule geometry with OpenEye
print "Generating molecule {}".format(name)
oe_molecule = openeye.smiles_to_oemol(smiles)
try:
oe_molecule = openeye.get_charges(oe_molecule, keep_confs=1)
except RuntimeError as e:
traceback.print_exc()
print "Skipping molecule " + name
return
# Create output subfolder
output_basepath = os.path.join(output_dir, name)
if not os.path.isdir(output_basepath):
os.mkdir(output_basepath)
output_basepath = os.path.join(output_basepath, name)
# Save mol2 file with residue name = first three uppercase letters
print "Running epik on molecule {}".format(name)
mol2_file_path = output_basepath + '-input.mol2'
residue_name = re.sub('[^A-Za-z]+', '', name.upper())[:3]
openeye.molecule_to_mol2(oe_molecule, mol2_file_path, residue_name=residue_name)
# Run epik on mol2 file
mae_file_path = output_basepath + '-epik.mae'
schrodinger.run_epik(mol2_file_path, mae_file_path, tautomerize=True,
max_structures=32, ph_tolerance=10.0)
# Convert maestro file to sdf and mol2
schrodinger.run_structconvert(mae_file_path, output_basepath + '-epik.sdf')
schrodinger.run_structconvert(mae_file_path, output_basepath + '-epik.mol2')
def enumerate_conformations(name, smiles=None, pdbname=None):
"""Run Epik to get protonation states using PDB residue templates for naming.
Parameters
----------
name : str
Common name of molecule (used to create subdirectory)
smiles : str
Isomeric SMILES string
pdbname : str
Three-letter PDB code (e.g. 'DB8')
"""
# Create output subfolder
output_basepath = os.path.join(output_dir, name)
if not os.path.isdir(output_basepath):
os.mkdir(output_basepath)
output_basepath = os.path.join(output_basepath, name)
if pdbname:
# Make sure to only use one entry if there are mutliple
if ' ' in pdbname:
pdbnames = pdbname.split(' ')
print("Splitting '%s' into first entry only: '%s'" % (pdbname, pdbnames[0]))
pdbname = pdbnames[0]
# Retrieve PDB (for atom names)
url = 'http://ligand-expo.rcsb.org/reports/%s/%s/%s_model.pdb' % (pdbname[0], pdbname, pdbname)
pdb_filename = output_basepath + '-input.pdb'
retrieve_url(url, pdb_filename)
pdb_molecule = read_molecule(pdb_filename)
# Retrieve SDF (for everything else)
url = 'http://ligand-expo.rcsb.org/reports/%s/%s/%s_model.sdf' % (pdbname[0], pdbname, pdbname)
sdf_filename = output_basepath + '-input.sdf'
retrieve_url(url, sdf_filename)
sdf_molecule = read_molecule(sdf_filename)
# Replace atom names in SDF
for (sdf_atom, pdb_atom) in zip(sdf_molecule.GetAtoms(), pdb_molecule.GetAtoms()):
sdf_atom.SetName(pdb_atom.GetName())
# Assign Tripos atom types
oechem.OETriposAtomTypeNames(sdf_molecule)
oechem.OETriposBondTypeNames(sdf_molecule)
oe_molecule = sdf_molecule
# We already know the residue name
residue_name = pdbname
elif smiles:
# Generate molecule geometry with OpenEye
print("Generating molecule {}".format(name))
oe_molecule = openeye.smiles_to_oemol(smiles)
# Assign Tripos atom types
oechem.OETriposAtomTypeNames(oe_molecule)
oechem.OETriposBondTypeNames(oe_molecule)
try:
oe_molecule = openeye.get_charges(oe_molecule, keep_confs=1)
except RuntimeError as e:
traceback.print_exc()
print("Skipping molecule " + name)
return
residue_name = re.sub('[^A-Za-z]+', '', name.upper())[:3]
else:
raise Exception('Must provide SMILES string or pdbname')
# Save mol2 file, preserving atom names
print("Running epik on molecule {}".format(name))
mol2_file_path = output_basepath + '-input.mol2'
write_mol2_preserving_atomnames(mol2_file_path, oe_molecule, residue_name)
# Run epik on mol2 file
mae_file_path = output_basepath + '-epik.mae'
schrodinger.run_epik(mol2_file_path, mae_file_path, tautomerize=False,
max_structures=100, min_probability=np.exp(-MAX_ENERGY_PENALTY), ph=7.4)
# Convert maestro file to sdf and mol2
output_sdf_filename = output_basepath + '-epik.sdf'
output_mol2_filename = output_basepath + '-epik.mol2'
schrodinger.run_structconvert(mae_file_path, output_sdf_filename)
schrodinger.run_structconvert(mae_file_path, output_mol2_filename)
# Read SDF file.
ifs_sdf = oechem.oemolistream()
ifs_sdf.SetFormat(oechem.OEFormat_SDF)
ifs_sdf.open(output_sdf_filename)
sdf_molecule = oechem.OEGraphMol()
# Read MOL2 file.
ifs_mol2 = oechem.oemolistream()
ifs_mol2.open(output_mol2_filename)
mol2_molecule = oechem.OEMol()
# Assign charges.
charged_molecules = list()
index = 0
while oechem.OEReadMolecule(ifs_sdf, sdf_molecule):
oechem.OEReadMolecule(ifs_mol2, mol2_molecule)
index += 1
print("Charging molecule %d" % (index))
try:
# Charge molecule.
charged_molecule = openeye.get_charges(mol2_molecule, max_confs=800, strictStereo=False, normalize=True, keep_confs=None)
# Assign Tripos types
oechem.OETriposAtomTypeNames(charged_molecule)
oechem.OETriposBondTypeNames(charged_molecule)
# Store tags.
oechem.OECopySDData(charged_molecule, sdf_molecule)
# Store molecule
charged_molecules.append(charged_molecule)
except Exception as e:
print(e)
print("Skipping protomer/tautomer because of failed charging.")
# Clean up
ifs_sdf.close()
ifs_mol2.close()
# Write state penalites.
outfile = open(output_basepath + '-state-penalties.out', 'w')
for (index, charged_molecule) in enumerate(charged_molecules):
# Get Epik data.
epik_Ionization_Penalty = float(oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty"))
epik_Ionization_Penalty_Charging = float(oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty_Charging"))
epik_Ionization_Penalty_Neutral = float(oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty_Neutral"))
epik_State_Penalty = float(oechem.OEGetSDData(charged_molecule, "r_epik_State_Penalty"))
epik_Tot_Q = int(oechem.OEGetSDData(charged_molecule, "i_epik_Tot_Q"))
outfile.write('%16.8f\n' % epik_State_Penalty)
outfile.close()
# Write as PDB
charged_pdb_filename = output_basepath + '-epik-charged.pdb'
ofs = oechem.oemolostream(charged_pdb_filename)
flavor = oechem.OEOFlavor_PDB_CurrentResidues | oechem.OEOFlavor_PDB_ELEMENT | oechem.OEOFlavor_PDB_BONDS | oechem.OEOFlavor_PDB_HETBONDS | oechem.OEOFlavor_PDB_BOTH
ofs.SetFlavor(oechem.OEFormat_PDB, flavor)
for (index, charged_molecule) in enumerate(charged_molecules):
# Fix residue names
for atom in charged_molecule.GetAtoms():
residue = oechem.OEAtomGetResidue(atom)
residue.SetName(residue_name)
oechem.OEAtomSetResidue(atom, residue)
#oechem.OEWritePDBFile(ofs, charged_molecule, flavor)
oechem.OEWriteMolecule(ofs, charged_molecule)
ofs.close()
# Write molecules as mol2.
charged_mol2_filename = output_basepath + '-epik-charged.mol2'
write_mol2_preserving_atomnames(charged_mol2_filename, charged_molecules, residue_name)
mae_directory_path = "./mae_files"
if not os.path.exists(mae_directory_path):
os.makedirs(mae_directory_path)
print("{} directory created.".format(mae_directory_path))
# Sequencial pKa calculation method is used starting form pH 7.0.
for key in eMolID_oemol_dict.keys():
print("Running Epik for molecule {} ...".format(key))
mol2_file_path = mol2_directory_path + "/" + str(key) + ".mol2"
mae_file_path = mae_directory_path + "/" + str(key) + ".mae"
schrodinger.run_epik(mol2_file_path,
mae_file_path,
max_structures=100,
ph=7.0,
ph_tolerance=None,
tautomerize=True,
extract_range=None,
max_atoms=150,
scan=True)
print("\n")
##### CONVERT EPIK OUTPUT (.MAE FILE) TO SDF #####
sdf_directory_path = "./sdf_files"
if not os.path.exists(sdf_directory_path):
os.makedirs(sdf_directory_path)
print("{} directory created.".format(sdf_directory_path))
for key in eMolID_oemol_dict.keys():
def enumerate_conformations(name,
pdbfile=None,
smiles=None,
pdbname=None,
pH=7.4):
"""Run Epik to get protonation states using PDB residue templates for naming.
Parameters
----------
name : str
Common name of molecule (used to create subdirectory)
smiles : str
Isomeric SMILES string
pdbname : str
Three-letter PDB code (e.g. 'DB8')
"""
# Create output subfolder
# output_basepath = os.path.join(output_dir, name)
# if not os.path.isdir(output_basepath):
# os.mkdir(output_basepath)
# output_basepath = os.path.join(output_basepath, name)
oehandler = openeye.oechem.OEThrow
# String stream output
oss = oechem.oeosstream()
oehandler.SetOutputStream(oss)
log = "New run:\nPDB code: {pdbname}; Molecule: {name}; pH {pH}\n".format(
**locals())
success_status = True
if pdbname:
# Make sure to only use one entry if there are multiple
if ' ' in pdbname:
pdbnames = pdbname.split(' ')
log += "Splitting '%s' into first entry only: '%s'" % (pdbname,
pdbnames[0])
pdbname = pdbnames[0]
# Retrieve PDB (for atom names)
url = 'http://ligand-expo.rcsb.org/reports/%s/%s/%s_model.pdb' % (
pdbname[0], pdbname, pdbname)
pdb_filename = name + '-rcsb_download.pdb'
log += "Retrieving PDB structure from RCSB ligand expo: {}.\n".format(
pdb_filename)
retrieve_url(url, pdb_filename)
log += "Parsing PDB file.\n"
pdb_molecule = read_molecule(pdb_filename)
# Retrieve SDF (for everything else)
url = 'http://ligand-expo.rcsb.org/reports/%s/%s/%s_model.sdf' % (
pdbname[0], pdbname, pdbname)
sdf_filename = name + '-rcsb_download.sdf'
log += "Retrieving SDF structure from RCSB ligand expo: {}.\n".format(
sdf_filename)
retrieve_url(url, sdf_filename)
log += "Parsing SDF file.\n"
sdf_molecule = read_molecule(sdf_filename)
# Replace atom names in SDF
log += "Canonicalizing atom names.\n"
for (sdf_atom, pdb_atom) in zip(sdf_molecule.GetAtoms(),
pdb_molecule.GetAtoms()):
sdf_atom.SetName(pdb_atom.GetName())
# Assign Tripos atom types
log += "Assign atom type names.\n"
oechem.OETriposAtomTypeNames(sdf_molecule)
oechem.OETriposBondTypeNames(sdf_molecule)
oe_molecule = sdf_molecule
# We already know the residue name
residue_name = pdbname
# For the moment, disabling these two types of input
# elif smiles:
# # Generate molecule geometry with OpenEye
# logging.info(("Generating molecule {}".format(name)))
# oe_molecule = openeye.smiles_to_oemol(smiles)
# # Assign Tripos atom types
# oechem.OETriposAtomTypeNames(oe_molecule)
# oechem.OETriposBondTypeNames(oe_molecule)
# try:
# logging.info("Charging initial")
# write_mol2_preserving_atomnames(name + '-debug.mol2', oe_molecule, 'debug')
# oe_molecule = openeye.get_charges(oe_molecule, keep_confs=1)
# except RuntimeError as e:
# traceback.print_exc()
# logging.info(("Skipping molecule " + name))
# return
# residue_name = re.sub('[^A-Za-z]+', '', name.upper())[:3]
# logging.info("resname = %s", residue_name)
# oe_molecule.SetTitle(residue_name) # fix iupac name issue with mol2convert
# elif pdbfile:
# residue_name = re.sub('[^A-Za-z]+', '', name.upper())[:3]
# logging.info("Loading molecule molecule {0} from {1}".format(name, pdbfile))
# oe_molecule = read_molecule(pdbfile)
# # Assign Tripos atom types
# oechem.OETriposAtomTypeNames(oe_molecule)
# oechem.OETriposBondTypeNames(oe_molecule)
# try:
# logging.info("Charging initial")
# write_mol2_preserving_atomnames(name + '-debug.mol2', oe_molecule, 'debug')
# oe_molecule = openeye.get_charges(oe_molecule, keep_confs=1)
# except RuntimeError as e:
# traceback.print_exc()
# logging.info(("Skipping molecule " + name))
# return
else:
raise Exception('Must provide SMILES string or pdbname, or pdbfile')
# Save mol2 file, preserving atom names
log += "Running Epik.\n"
mol2_file_path = name + '-before_epik.mol2'
write_mol2_preserving_atomnames(mol2_file_path, oe_molecule, residue_name)
# Run epik on mol2 file
mae_file_path = name + '-epik.mae'
schrodinger.run_epik(mol2_file_path,
mae_file_path,
tautomerize=False,
max_structures=50,
min_probability=np.exp(-MAX_ENERGY_PENALTY),
ph=pH)
log += "Epik run completed.\n"
# Convert maestro file to sdf and mol2
output_sdf_filename = name + '-after_epik.sdf'
output_mol2_filename = name + '-after_epik.mol2'
# logging.info("Creating sdf")
schrodinger.run_structconvert(mae_file_path, output_sdf_filename)
# logging.info("Creating mol2")
schrodinger.run_structconvert(mae_file_path, output_mol2_filename)
# Read SDF file.
ifs_sdf = oechem.oemolistream()
ifs_sdf.SetFormat(oechem.OEFormat_SDF)
ifs_sdf.open(output_sdf_filename)
sdf_molecule = oechem.OEGraphMol()
# Read MOL2 file.
ifs_mol2 = oechem.oemolistream()
ifs_mol2.open(output_mol2_filename)
mol2_molecule = oechem.OEMol()
# Assign charges.
# reset count of error handler
oehandler.Clear()
log += "Assigning charges to protonation states.\n"
charged_molecules = list()
index = 0
failed_states = set()
while oechem.OEReadMolecule(ifs_sdf, sdf_molecule):
oechem.OEReadMolecule(ifs_mol2, mol2_molecule)
index += 1
log += "State {0:d}\n".format(index)
try:
# Charge molecule.
charged_molecule_conformers = omtoe.get_charges(mol2_molecule,
max_confs=800,
strictStereo=False,
normalize=True,
keep_confs=-1)
log += "Charging stage output:\n"
OEOutput = str(oss)
log += OEOutput
log += "\nCharging state completed.\n"
# Restore coordinates to original
charged_molecule = select_conformers(charged_molecule_conformers,
mol2_molecule,
keep_confs=None)
# Assign Tripos types
oechem.OETriposAtomTypeNames(charged_molecule)
oechem.OETriposBondTypeNames(charged_molecule)
# Store tags.
oechem.OECopySDData(charged_molecule, sdf_molecule)
# Store molecule
charged_molecules.append(charged_molecule)
# Check for failure in the log
openeye_charge_log_parser(OEOutput, True)
oehandler.Clear()
except Exception as e:
failed_states.add(index)
logging.info(e)
log += "State failed charging.\n"
log += str(e)
log += "\n"
filename_failure = name + '-conformers-failed-state-{}-.mol2'.format(
index)
try:
write_mol2_preserving_atomnames(filename_failure,
charged_molecule_conformers,
residue_name)
except:
log += "Could not store result, most likely failed during Omega step!\n"
success_status = False
oehandler.Clear()
# Clean up
ifs_sdf.close()
ifs_mol2.close()
# Write state penalties.
outfile = open(name + '-state-penalties.out', 'w')
for (index, charged_molecule) in enumerate(charged_molecules):
# Get Epik data.
log += "Writing Epik data for state {:d}\n".format(index + 1)
epik_Ionization_Penalty = float(
oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty"))
epik_Ionization_Penalty_Charging = float(
oechem.OEGetSDData(charged_molecule,
"r_epik_Ionization_Penalty_Charging"))
epik_Ionization_Penalty_Neutral = float(
oechem.OEGetSDData(charged_molecule,
"r_epik_Ionization_Penalty_Neutral"))
epik_State_Penalty = float(
oechem.OEGetSDData(charged_molecule, "r_epik_State_Penalty"))
epik_Tot_Q = int(oechem.OEGetSDData(charged_molecule, "i_epik_Tot_Q"))
outfile.write('%16.8f\n' % epik_State_Penalty)
outfile.close()
# Write as PDB
charged_pdb_filename = name + '-charged_output.pdb'
ofs = oechem.oemolostream(charged_pdb_filename)
flavor = oechem.OEOFlavor_PDB_CurrentResidues | oechem.OEOFlavor_PDB_ELEMENT | oechem.OEOFlavor_PDB_BONDS | oechem.OEOFlavor_PDB_HETBONDS | oechem.OEOFlavor_PDB_BOTH
ofs.SetFlavor(oechem.OEFormat_PDB, flavor)
for (index, charged_molecule) in enumerate(charged_molecules):
# Fix residue names
for atom in charged_molecule.GetAtoms():
residue = oechem.OEAtomGetResidue(atom)
residue.SetName(residue_name)
oechem.OEAtomSetResidue(atom, residue)
oechem.OEWriteMolecule(ofs, charged_molecule)
ofs.close()
# Write molecules as mol2.
charged_mol2_filename = name + '-charged_output.mol2'
write_mol2_preserving_atomnames(charged_mol2_filename, charged_molecules,
residue_name)
log += "Run completed.\n"
if success_status:
log += "Status: Success\n"
else:
log += "Status: Failure\n"
log += "Failed states: {}\n".format(" ".join(
[str(state) for state in sorted(list(failed_states))]))
with open("log.txt", 'w') as logfile:
logfile.write(log)
return log, success_status
def run_epik(name, filename, residue_name, perceive_bonds=False):
"""Generate conformer with OpenEye omega, protonation states with Schrodinger Epik, and charges with OpenEye AM1-BCC.
Parameters
----------
name : str
The name of the output directory to generate.
filename : str
The mol2, PDB, or SDF file to read in.
residue_name : str
Three uppercase letters to name residue.
perceive_bonds : bool, optional, default=False
If True, will use geometry to perceive connectivity.
This is necessary for PDB files.
"""
# Generate molecule geometry with OpenEye
print("Generating molecule %s from %s" % (name, filename))
oe_molecule = read_molecules(filename)
if perceive_bonds:
oechem.OEDetermineConnectivity(oe_molecule)
# Assign geometry and charges with Omega
oe_molecule = openeye.get_charges(oe_molecule, max_confs=1, strictStereo=False, normalize=True, keep_confs=1)
# Create output subfolder
output_basepath = os.path.join(output_dir, name)
if not os.path.isdir(output_basepath):
os.mkdir(output_basepath)
output_basepath = os.path.join(output_basepath, name)
# Save mol2 file with residue name = first three uppercase letters
print "Running epik on molecule {}".format(name)
mol2_file_path = output_basepath + '-input.mol2'
residue_name = re.sub('[^A-Za-z]+', '', name.upper())[:3]
#openeye.molecule_to_mol2(oe_molecule, mol2_file_path, residue_name=residue_name)
from openeye import oechem
ofs = oechem.oemolostream(mol2_file_path)
oechem.OEWriteMol2File(ofs, oe_molecule, True, False)
ofs.close()
# Run epik on mol2 file
mae_file_path = output_basepath + '-epik.mae'
schrodinger.run_epik(mol2_file_path, mae_file_path, tautomerize=False,
max_structures=100, min_probability=np.exp(-6), ph=7.4)
# Convert maestro file to sdf and mol2
output_sdf_filename = output_basepath + '-epik.sdf'
output_mol2_filename = output_basepath + '-epik.mol2'
schrodinger.run_structconvert(mae_file_path, output_sdf_filename)
schrodinger.run_structconvert(mae_file_path, output_mol2_filename)
# Read SDF file.
ifs_sdf = oechem.oemolistream()
ifs_sdf.SetFormat(oechem.OEFormat_SDF)
ifs_sdf.open(output_sdf_filename)
sdf_molecule = oechem.OEMol()
uncharged_molecules = read_molecules(output_sdf_filename)
# Read MOL2 file.
ifs_mol2 = oechem.oemolistream()
ifs_mol2.open(output_mol2_filename)
mol2_molecule = oechem.OEMol()
uncharged_molecules = read_molecules(output_sdf_filename)
# Assign charges.
charged_molecules = list()
index = 0
while oechem.OEReadMolecule(ifs_sdf, sdf_molecule):
molecule = oechem.OEReadMolecule(ifs_mol2, mol2_molecule)
index += 1
print "Charging molecule %d / %d" % (index, len(uncharged_molecules))
try:
# Charge molecule.
charged_molecule = openeye.get_charges(sdf_molecule, max_confs=800, strictStereo=False, normalize=True, keep_confs=None)
# Store tags.
oechem.OECopySDData(charged_molecule, sdf_molecule)
charged_molecules.append(charged_molecule)
except Exception as e:
print(e)
print("Skipping protomer/tautomer because of failed charging.")
# Clean up
ifs_sdf.close()
ifs_mol2.close()
# Write molecules.
charged_mol2_filename = output_basepath + '-epik-charged.mol2'
ofs = oechem.oemolostream(charged_mol2_filename)
for (index, charged_molecule) in enumerate(charged_molecules):
oechem.OEWriteMolecule(ofs, charged_molecule)
ofs.close()
# Write state penalites.
outfile = open(output_basepath + '-state-penalties.out', 'w')
for (index, charged_molecule) in enumerate(charged_molecules):
# Get Epik data.
epik_Ionization_Penalty = float(oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty"))
epik_Ionization_Penalty_Charging = float(oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty_Charging"))
epik_Ionization_Penalty_Neutral = float(oechem.OEGetSDData(charged_molecule, "r_epik_Ionization_Penalty_Neutral"))
epik_State_Penalty = float(oechem.OEGetSDData(charged_molecule, "r_epik_State_Penalty"))
epik_Tot_Q = int(oechem.OEGetSDData(charged_molecule, "i_epik_Tot_Q"))
outfile.write('%16.8f\n' % epik_State_Penalty)
outfile.close()