def standarize_mol_by_inchi(mol, neutralize=True):
newmol = AddHs(mol)
sinchi, code, msg = generate_inchi(newmol, FixedH=False, RecMet=False)
if neutralize:
nsinchi = neutralize_inchi(sinchi)
else:
nsinchi = sinchi
newmol = MolFromInchi(nsinchi, removeHs=False)
newmol = AddHs(newmol, explicitOnly=True)
return newmol
def xyz_to_rdmol(nxyz, smiles):
mol = get_mol(smiles)
mol = AddHs(mol)
num_atoms = len(nxyz)
conformer = Conformer(num_atoms)
for i, quad in enumerate(nxyz):
conformer.SetAtomPosition(i, quad[1:])
mol.AddConformer(conformer)
return mol
def generate_conformers(mol, add_hydrogens=True,
rmsd_threshold=2.0, # Arbitrarily selected
num_conformers=None, # None means best guess
parallelism=None,
forcefield=DEFAULT_FORCEFIELD,
log=logging):
if add_hydrogens:
log.info("Adding implicit hydrogens")
mol = AddHs(mol)
if num_conformers is None:
num_conformers = get_num_confs_for_mol(mol)
log.info("Attempting to generate {0} conformations with min RMSD of {1:.4f}".format(num_conformers, rmsd_threshold))
orig_conf_ids = EmbedMultipleConfs(mol, numConfs=num_conformers,
pruneRmsThresh=rmsd_threshold,
ignoreSmoothingFailures=True) # Prevents crashes in some situations
log.info("Generated {0} initial conformations".format(len(orig_conf_ids)))
log.info("Optimizing and calculating energies using {0}".format(forcefield))
conf_energy = optimize_conformers(mol, interfragment=True,
parallelism=parallelism,
forcefield=forcefield)
sorted_by_energy = sorted(conf_energy.iteritems(), key=operator.itemgetter(1))
log.info("Filtering similar conformers")
selected = []
min_rmsd, max_rmsd = float('inf'), float('-inf')
for idx, id_energy in enumerate(sorted_by_energy):
conf_id, energy = id_energy
keep = True
for comp_id, other_energy in sorted_by_energy[idx+1:]:
rmsd = AlignMol(mol, mol, prbCid=comp_id, refCid=conf_id)
if rmsd <= rmsd_threshold:
mol.RemoveConformer(conf_id)
keep = False
break
else:
if rmsd < min_rmsd:
min_rmsd = rmsd
if rmsd > max_rmsd:
max_rmsd = rmsd
if keep:
selected.append(id_energy)
log.debug("Removed {0} after post-optimization RMSD filtering".format(len(orig_conf_ids) - len(selected)))
log.info("RMSD: min={0:.4f} max={1:.4f}".format(min_rmsd, max_rmsd))
return mol, selected
def embed_etkdg(mol, seed):
from rdkit.Chem import AllChem, AddHs
mol = AddHs(mol)
params = AllChem.ETKDG()
params.randomSeed = seed
AllChem.EmbedMolecule(mol, params)
return mol
def rdkit_3d_descirptors(mols, regex="(NPR1)|(NPR2)|(PMI1)|(PMI2)|(PMI3)|(SpherocityIndex)|(InertialShapeFactor)|"
"(Eccentricity)|(Asphericity)"):
""" embeds molecules in 3D and calculates a set of RDKit descriptors for given molecules (RDKit) ``mols``
:param mols: {str} RDKit molecules
:param regex: {str} regular expression to match RDKit functions
:return: {pandas DataFrame} descriptor names and values
"""
# embed molecules in 3D
mols = [AddHs(m) for m in mols]
for i, m in enumerate(mols):
AllChem.EmbedMolecule(m, AllChem.ETKDG())
# AllChem.MMFFOptimizeMolecule(m)
# create results dictionary with descriptors as keys and append list of values for all mols
rslt = dict()
desc_regex = re.compile(regex)
for descriptor in Descriptors3D.__dict__.keys():
if desc_regex.match(descriptor):
print("\t%s..." % descriptor)
func = getattr(Descriptors3D, descriptor)
pbar = ProgressBar()
rslt[descriptor] = list()
for mol in pbar(mols):
rslt[descriptor].append(func(mol))
return pd.DataFrame(rslt)
def calculate(self):
"""
Check if the SMILES is valid then update the info.
:return: RDKit Mol object
"""
try:
m = MolFromSmiles("".join(p.config['long_prefix']) +
"".join(self.smiles.element))
self.smiles.properties[p.s_valid] = False
if m is not None:
m = AddHs(m)
AllChem.EmbedMolecule(m)
AllChem.UFFOptimizeMolecule(m)
self.smiles.properties["InChI"] = MolToInchi(m)
except Exception as e:
print("Error rdkit : " + repr(e))
m = None
if m is not None:
self.smiles.properties[p.s_valid] = True
with p.lock_update_data:
p.tree_info[p.info_good] += 1
self.smiles.properties[p.s_id] = p.tree_info[p.info_good]
else:
with p.lock_update_data:
p.tree_info[p.info_bad] += 1
self.smiles.properties[p.s_id] = p.tree_info[p.info_bad]
return m
def add_hs(self,
inplace=False,
add_coords=True,
explicit_only=False,
only_on_atoms=False):
""" Add hydrogens to self.
Args:
inplace (bool):
Whether to add Hs to `Mol`, or return a new `Mol`.
add_coords (bool):
Whether to set 3D coordinate for added Hs.
explicit_only (bool):
Whether to add only explicit Hs, or also implicit ones.
only_on_atoms (iterable<bool>):
An iterable specifying the atoms to add Hs.
Returns:
skchem.Mol:
`Mol` with Hs added.
"""
if inplace:
msg = 'Inplace addition of Hs is not yet supported.'
raise NotImplementedError(msg)
raw = AddHs(self,
addCoords=add_coords,
onlyOnAtoms=only_on_atoms,
explicitOnly=explicit_only)
return self.__class__.from_super(raw)
def add_hydrogen(cls, mol_in, addCoords=True):
"""Explicit all hydrogens.
:param mol_in: RDKit Mol
:param addCoords: Add coordinate to added Hs, bool
:return mol_out: RDKit Mol
"""
return AddHs(mol_in, explicitOnly=False, addCoords=addCoords)
def generate(self,
max_generated_conformers=50,
prune_thresh=0.01,
maxattempts_per_conformer=5,
output=None,
threads=1):
'''
Generates conformers
Note the number max_generated _conformers required is related to the
number of rotatable bonds
'''
self.mol = AddHs(self.mol, addCoords=True)
self.initial_confs = EmbedMultipleConfs(
self.mol,
numConfs=max_generated_conformers,
pruneRmsThresh=prune_thresh,
maxAttempts=maxattempts_per_conformer,
useRandomCoords=False,
# Despite what the documentation says -1 is a seed!!
# It doesn't mean random generation
numThreads=threads,
randomSeed=random.randint(1, 10000000))
if len(self.initial_confs) == 0:
output.write((f"Generated {len(self.initial_confs)} "
"initial confs\n"))
output.write((f"Trying again with {max_generated_conformers * 10} "
"attempts and random coords\n"))
self.initial_confs = EmbedMultipleConfs(
self.mol,
numConfs=max_generated_conformers,
pruneRmsThresh=prune_thresh,
useRandomCoords=True,
maxAttempts=10 * maxattempts_per_conformer,
# Despite what the documentation says -1 is a seed!!
# It doesn't mean random
# generatrion
numThreads=threads,
randomSeed=random.randint(1, 10000000))
output.write("Generated " + str(len(self.initial_confs)) +
" initial confs\n")
return self.initial_confs
def docksmile(smile, filename):
'''
coverts a smile string to pdbqt and runs autodock vina,
returns the binding energy of its top pose
Vina configuration details in config.txt
'''
#print(smile, filename)
if not isinstance(smile, str):
raise TypeError('Input is not a class of string')
m = MolFromSmiles(smile)
# assert valid smiles
if m is None:
raise ValueError(smile, 'is not a valid smile string')
mh = AddHs(m)
embed = AllChem.EmbedMolecule(mh, useRandomCoords=False)
#check if rdkit successfully generates structure
if embed!=0:
print('RDkit fails to embed molecule', smile, '; file:%s.pdb'%filename)
return smile, np.nan
# generate pdb file
#pdb = MolToPDBFile(mh, 'input/'+filename+'.pdb', flavor=4)
pdb = MolToPDBBlock(mh, flavor=4)
open('/tmp/'+filename+'.pdb', 'w').write(pdb)
# convert pdb to pdbqt
try:
out = subprocess.run([py_path, lig_path, '-l', '/tmp/'+filename+'.pdb', '-o','/tmp/'+filename+'.pdbqt'])
except subprocess.CalledProcessError as e:
print(e.output)
if not os.path.exists('/tmp/'+filename+'.pdbqt'):
print("%s does't exist" % (filename+'.pdbqt'))
return smile, np.nan
try:
result = subprocess.run(['sh', './run_spike_open_docking.sh', filename], stdout=subprocess.PIPE)
result = result.stdout.decode('utf-8')
except subprocess.CalledProcessError as er:
print(er.output)
print(smile, '; file:%s.pdbqt'%filename)
return smile, np.nan
#print(filename+'.pdbqt','docking success')
# read energy from output
energy = np.nan
strings = re.split('\n', result)
for line in strings:
if line[0:4] == ' 1':
energy = float(re.split(' +', line)[2])
#print(energy )
return smile, energy
def __init__(self, smiles, forcefield="mmff"):
'''
Initialises the class
'''
self.mol = MolFromSmiles(smiles)
self.full_clusters = []
self.forcefield = forcefield
self.conf_energies = []
self.initial_confs = None
self.smiles = smiles
def processline(t, step, line):
global lensum
if t.incr():
return 1
if step == 0:
lensum += len(line)
else:
m = MolFromSmiles(line)
if step == 100:
lensum += len(line)
elif step == 105:
lensum += len(sha256(line).hexdigest())
elif step in (110, 120):
with open(tmpname, 'wb+') as f:
print(line, file=f)
if step == 120:
os.fsync(f.fileno())
lensum += os.stat(tmpname).st_size
elif step == 210:
lensum += m.GetNumAtoms()
elif step == 220:
lensum += m.GetNumBonds()
elif step == 300:
lensum += len(MolToSmiles(m))
elif step == 400:
lensum += len(MolToMolBlock(m))
elif step == 420:
m2 = AddHs(m)
EmbedMolecule(m2, randomSeed=2020)
m2 = RemoveHs(m2)
m2.SetProp("_Name", "test")
lensum += len(MolToMolBlock(m2))
elif step == 600:
lensum += mol2file(m, 'svg')
elif step == 610:
lensum += mol2file(m, 'png')
else:
raise ValueError("Not implemented step " + str(step))
return 0
def get_max_atom_bond_size(smiles_iterator, explicit_hs=True):
""" Convienence function to get max_atoms, max_bonds for a set of input
SMILES """
max_atoms = 0
max_bonds = 0
for smiles in tqdm(smiles_iterator):
mol = MolFromSmiles(smiles)
if explicit_hs:
mol = AddHs(mol)
max_atoms = max([max_atoms, len(mol.GetAtoms())])
max_bonds = max([max_bonds, len(mol.GetBonds())])
return dict(max_atoms=max_atoms, max_bonds=max_bonds * 2)
def rdmols_from_document(document, build_from="inchi", add_hs=True):
"""
Convert back a document to a set of rdmols. This method is a companion of "as_document".
:param document: a document produced by the "as_mongo_document" method, dict
:param build_from: the type of depiction to be used to build back the rdmols, str in ["inchi", "smiles"]
:param add_hs: add Hs to RDKit mol object, default is True
:returns list_list_rdmols: list of list of rdmols
"""
assert build_from in ["inchi", "smiles"]
assert add_hs in [True, False]
list_list_rdmols = list()
list_stoechiometry = document['list_stoechiometry']
if build_from == 'inchi':
for list_inchis in document['list_list_inchis']:
list_rdmols = list()
for inchi in list_inchis:
rd_mol = MolFromInchi(inchi, sanitize=True)
if add_hs:
rd_mol = AddHs(rd_mol)
list_rdmols.append(rd_mol)
list_list_rdmols.append(list_rdmols)
elif build_from == 'smiles':
for list_smiles in document['list_list_smiles']:
list_rdmols = list()
for smiles in list_smiles:
rd_mol = MolFromSmiles(smiles, sanitize=True)
if add_hs:
rd_mol = AddHs(rd_mol)
list_rdmols.append(rd_mol)
list_list_rdmols.append(list_rdmols)
else:
raise NotImplementedError()
return list_list_rdmols, list_stoechiometry
def _predict_rt(self, smiles: str) -> Optional[float]:
"""Predict Retention Time from SMILES string using provided predictor.
Parameters
----------
smiles : str
SMILES string of input compound.
Returns
-------
predicted_rt : Optional[float]
Predicted retention time, None if errors occur during prediction,
for example if certain features of the input compound that are
required for the prediction cannot be calculated.
"""
mol = MolFromSmiles(smiles)
mol = AddHs(mol)
fp = self.fp_calculator(mol)
# Transform dict into array of values (fingerprint)
if self.rt_important_features:
fp = np.array(
[fp[feature] for feature in self.rt_important_features]
).reshape(1, -1)
def validate_np_val(val: float) -> bool:
"""Make sure value is numeric, not NaN, and not infinity.
Parameters
----------
val : float
Value to check.
Returns
-------
bool
True if input value is numeric, False otherwise.
"""
if isinstance(val, float) and not np.isnan(val) and not np.isinf(val):
return True
return False
if all([validate_np_val(val) for val in fp[0]]):
predicted_rt = self.rt_predictor.predict(fp)[0]
else:
return None
return predicted_rt
def smiles_reaction_matrix(smarts, *sources, **kwargs):
sep = kwargs.setdefault('sep', ' ')
molValue = int(kwargs.get('molValue', 400))
logValue = float(kwargs.get('logValue', 4.0))
reaction = ReactionFromSmarts(smarts)
smilesLists = [load_smiles_file(source) for source in sources]
products = reaction_matrix(reaction, *smilesLists)
for reactants, product in products:
cids = [r.GetProp("_Name") for r in reactants]
product_id = '.'.join(cids)
for mol in product:
smiles = MolToSmiles(mol, isomericSmiles=True)
mol.UpdatePropertyCache(strict=False)
mh = AddHs(mol, addCoords=True)
mwt = MolWt(mol)
if mwt <= molValue:
logp = MolLogP(mol)
if logp < logValue:
yield sep.join((smiles, product_id, str(mwt), str(logp)))+"\n"
def construct_feature_matrices(self, smiles, train=True):
""" construct a molecule from the given smiles string and return atom
and bond classes.
Returns
dict with entries
'n_atom' : number of atoms in the molecule
'n_bond' : number of bonds in the molecule
'atom' : (n_atom,) length list of atom classes
'bond' : (n_bond,) list of bond classes
'connectivity' : (n_bond, 2) array of source atom, target atom pairs.
"""
self.atom_tokenizer.train = train
self.bond_tokenizer.train = train
logger = logging.getLogger(__name__)
mol = MolFromSmiles(smiles)
if self.explicit_hs:
mol = AddHs(mol)
n_atom = mol.GetNumAtoms()
n_bond = 2 * mol.GetNumBonds()
# If its an isolated atom, add a self-link
if n_bond == 0:
n_bond = 1
logger.warning(f'Found molecule {smiles} with zero bonds')
atom_feature_matrix = np.zeros(n_atom, dtype='int')
bond_feature_matrix = np.zeros(n_bond, dtype='int')
bond_indices = np.zeros(n_bond, dtype='int')
connectivity = np.zeros((n_bond, 2), dtype='int')
bond_index = 0
for n, atom in enumerate(mol.GetAtoms()):
# Atom Classes
atom_feature_matrix[n] = self.atom_tokenizer(
self.atom_features(atom))
start_index = atom.GetIdx()
for bond in atom.GetBonds():
# Is the bond pointing at the target atom
rev = bond.GetBeginAtomIdx() != start_index
# Bond Classes
bond_feature_matrix[bond_index] = self.bond_tokenizer(
self.bond_features(bond, flipped=rev))
# Connect edges to original bonds
bond_indices[bond_index] = bond.GetIdx()
# Connectivity
if not rev: # Original direction
connectivity[bond_index, 0] = bond.GetBeginAtomIdx()
connectivity[bond_index, 1] = bond.GetEndAtomIdx()
else: # Reversed
connectivity[bond_index, 0] = bond.GetEndAtomIdx()
connectivity[bond_index, 1] = bond.GetBeginAtomIdx()
bond_index += 1
# Track the largest atom and bonds seen
if train:
if n_atom > self.max_atoms:
self.max_atoms = n_atom
if mol.GetNumBonds() > self.max_bonds:
self.max_bonds = mol.GetNumBonds()
return {
'n_atom': n_atom,
'n_bond': mol.GetNumBonds(), # the real number of bonds
'bond_indices': bond_indices,
'atom': atom_feature_matrix,
'bond': bond_feature_matrix,
'connectivity': connectivity,
}
#!/usr/bin/python2
# Little harness for timing how long it takes to embed a molecule
# which seems extremely variable on one machine,
from __future__ import print_function, division
import sys, time, os
from rdkit.Chem import MolFromSmiles, AddHs, RemoveHs
from rdkit.Chem.AllChem import EmbedMolecule
if __name__ == "__main__":
dotimestamp = int(os.getenv('MOLEMBED_TIME', '0'))
doaddh = int(os.getenv('MOLEMBED_ADDH', '0'))
rseed = int(os.getenv('MOLEMBED_SEED', '0'))
t0 = time.time()
for line in sys.stdin.readlines():
s = line.strip()
if dotimestamp:
t1 = time.time()
dt = (t1 - t0) * 1e3
print('%.3f' % dt, s)
t0 = t1
else:
print(s)
m = MolFromSmiles(s)
if doaddh:
m2 = AddHs(m)
else:
m2 = m
EmbedMolecule(m2, randomSeed=rseed)
def addhs(self, mol):
from rdkit.Chem import AddHs
return AddHs(mol)
def construct_feature_matrices(self, smiles):
""" construct a molecule from the given smiles string and return atom
and bond classes.
Returns
dict with entries
'n_atom' : number of atoms in the molecule
'n_bond' : number of bonds in the molecule
'atom' : (n_atom,) length list of atom classes
'bond' : (n_bond,) list of bond classes
'connectivity' : (n_bond, 2) array of source atom, target atom pairs.
"""
mol = MolFromSmiles(smiles)
if self.explicit_hs:
mol = AddHs(mol)
n_atom = len(mol.GetAtoms())
n_bond = 2 * len(mol.GetBonds())
# If its an isolated atom, add a self-link
if n_bond == 0:
n_bond = 1
atom_feature_matrix = np.zeros(n_atom, dtype='int')
bond_feature_matrix = np.zeros(n_bond, dtype='int')
connectivity = np.zeros((n_bond, 2), dtype='int')
bond_index = 0
atom_seq = mol.GetAtoms()
atoms = [atom_seq[i] for i in range(n_atom)]
for n, atom in enumerate(atoms):
# Atom Classes
atom_feature_matrix[n] = self.atom_tokenizer(
self.atom_features(atom))
start_index = atom.GetIdx()
for bond in atom.GetBonds():
# Is the bond pointing at the target atom
rev = bond.GetBeginAtomIdx() != start_index
# Bond Classes
bond_feature_matrix[bond_index] = self.bond_tokenizer(
self.bond_features(bond, flipped=rev))
# Connectivity
if not rev: # Original direction
connectivity[bond_index, 0] = bond.GetBeginAtomIdx()
connectivity[bond_index, 1] = bond.GetEndAtomIdx()
else: # Reversed
connectivity[bond_index, 0] = bond.GetEndAtomIdx()
connectivity[bond_index, 1] = bond.GetBeginAtomIdx()
bond_index += 1
return {
'n_atom': n_atom,
'n_bond': n_bond,
'atom': atom_feature_matrix,
'bond': bond_feature_matrix,
'connectivity': connectivity,
}
#! /usr/bin/env python
import sys
from rdkit.Chem import SDMolSupplier, MolToPDBFile, AllChem, AddHs, RemoveHs
from rdkit.Chem.Draw import MolsToGridImage
spl = SDMolSupplier(sys.argv[1])
mols = [m for m in spl]
for i, m in enumerate(mols):
m = AddHs(m)
AllChem.EmbedMolecule(m, useBasicKnowledge=True, maxAttempts=100)
AllChem.MMFFOptimizeMolecule(m)
RemoveHs(m)
MolToPDBFile(m, 'ligand_%d.pdb' % i)
img = MolsToGridImage(mols,
legends=["ligand_%d" % i for i in range(len(mols))])
img.save('ligands.png')
class ConformerGenerator(object):
'''
Generates conformations of molecules from 2D representation.
'''
def __init__(self, smiles, forcefield="mmff"):
'''
Initialises the class
'''
self.mol = MolFromSmiles(smiles)
self.full_clusters = []
self.forcefield = forcefield
self.conf_energies = []
self.initial_confs = None
self.smiles = smiles
def generate(self,
max_generated_conformers=50,
prune_thresh=0.01,
maxattempts_per_conformer=5,
output=None,
threads=1):
'''
Generates conformers
Note the number max_generated _conformers required is related to the
number of rotatable bonds
'''
self.mol = AddHs(self.mol, addCoords=True)
self.initial_confs = EmbedMultipleConfs(
self.mol,
numConfs=max_generated_conformers,
pruneRmsThresh=prune_thresh,
maxAttempts=maxattempts_per_conformer,
useRandomCoords=False,
# Despite what the documentation says -1 is a seed!!
# It doesn't mean random generation
numThreads=threads,
randomSeed=random.randint(1, 10000000))
if len(self.initial_confs) == 0:
output.write((f"Generated {len(self.initial_confs)} "
"initial confs\n"))
output.write((f"Trying again with {max_generated_conformers * 10} "
"attempts and random coords\n"))
self.initial_confs = EmbedMultipleConfs(
self.mol,
numConfs=max_generated_conformers,
pruneRmsThresh=prune_thresh,
useRandomCoords=True,
maxAttempts=10 * maxattempts_per_conformer,
# Despite what the documentation says -1 is a seed!!
# It doesn't mean random
# generatrion
numThreads=threads,
randomSeed=random.randint(1, 10000000))
output.write("Generated " + str(len(self.initial_confs)) +
" initial confs\n")
return self.initial_confs
def minimise(self, output=None):
'''
Minimises conformers using a force field
'''
if "\\" in self.smiles or "/" in self.smiles:
output.write(("WARNING: Smiles string contains slashes, "
"which specify cis/trans stereochemistry.\n"))
output.write(("Force-field minimization may change the "
"stereochemistry.\n"))
if self.forcefield != "mmff" and self.forcefield != "uff":
raise ValueError("Unrecognised force field")
if self.forcefield == "mmff":
props = MMFFGetMoleculeProperties(self.mol)
for i in range(0, len(self.initial_confs)):
potential = MMFFGetMoleculeForceField(self.mol,
props,
confId=i)
if potential is None:
output.write("MMFF not available, using UFF\n")
potential = UFFGetMoleculeForceField(self.mol, confId=i)
assert potential is not None
output.write(f"Minimising conformer number {i}\n")
potential.Minimize()
mmff_energy = potential.CalcEnergy()
self.conf_energies.append((i, mmff_energy))
elif self.forcefield == "uff":
for i in range(0, len(self.initial_confs)):
potential = UFFGetMoleculeForceField(self.mol, confId=i)
assert potential is not None
potential.Minimize()
uff_energy = potential.CalcEnergy()
self.conf_energies.append((i, uff_energy))
self.conf_energies = sorted(self.conf_energies, key=lambda tup: tup[1])
return self.mol
def cluster(self,
rms_tolerance=0.1,
max_ranked_conformers=10,
energy_window=5,
Report_e_tol=10,
output=None):
'''
Removes duplicates after minimization
'''
self.counter = 0
self.factormax = 3
self.mol_no_h = RemoveHs(self.mol)
calcs_performed = 0
self.full_clusters = []
confs = self.conf_energies[:]
ignore = []
ignored = 0
for i, pair_1 in enumerate(confs):
if i == 0:
index_0, energy_0 = pair_1
output.write((f"clustering cluster {i} of "
f"{len(self.conf_energies)}\n"))
index_1, energy_1 = pair_1
if abs(energy_1 - energy_0) > Report_e_tol:
output.write(("Breaking because hit Report Energy Window, "
f"E was {energy_1} kcal/mol "
f"and minimum was {energy_0} \n"))
break
if i in ignore:
ignored += i
continue
self.counter += 1
if self.counter == self.factormax * max_ranked_conformers:
output.write('Breaking because hit MaxNConfs \n')
break
clustered = [[self.mol.GetConformer(id=index_1), energy_1, 0.00]]
ignore.append(i)
for j, pair_2 in enumerate(confs):
if j > 1:
index_2, energy_2 = pair_2
if j in ignore:
ignored += 1
continue
if abs(energy_1 - energy_2) > energy_window:
break
if abs(energy_1 - energy_2) <= 1e-3:
clustered.append([
self.mol.GetConformer(id=index_2), energy_2, 0.00
])
ignore.append(j)
rms = GetConformerRMS(self.mol_no_h, index_1, index_2)
calcs_performed += 1
if rms <= rms_tolerance:
clustered.append([
self.mol.GetConformer(id=index_2), energy_2,
rms
])
ignore.append(j)
self.full_clusters.append(clustered)
output.write(f"{ignored} ignore passes made\n")
output.write((f"{calcs_performed} overlays needed out "
f"of a possible {len(self.conf_energies) ** 2}\n"))
ranked_clusters = []
for i, cluster in enumerate(self.full_clusters):
if i < self.factormax * max_ranked_conformers:
ranked_clusters.append(cluster[0])
return ranked_clusters
def recluster(self,
path,
rms_tolerance=0.1,
max_ranked_conformers=10,
energy_window=5,
output=None,
clustered_confs=[],
molecule=None,
key=None,
fallback_to_align=False):
self.removed = []
self.counter = 0
i = -1
for conf_a in clustered_confs:
i += 1
j = i
if self.counter == max_ranked_conformers:
for k in range(i, len(clustered_confs)):
if os.path.isfile(key + "_Conf_" + str(k + 1) + ".xyz"):
os.remove(key + "_Conf_" + str(k + 1) + ".xyz")
output.write("Removed " + key + "_Conf_" + str(k + 1) +
".xyz\n")
break
if i in self.removed:
continue
self.counter += 1
for conf_b in clustered_confs[i + 1:]:
j += 1
if conf_b[1] - conf_a[1] > energy_window:
break
if j in self.removed:
continue
try:
rms = obfit_rmsd(key + "_Conf_" + str(i + 1),
key + "_Conf_" + str(j + 1),
str(molecule),
path=path)
except (subprocess.CalledProcessError, ValueError) as e:
if fallback_to_align:
output.write(
'obfit failed, falling back to obabel --align')
output.write(f'Exception {e}\n')
try:
rms = align_rmsd(f"{key}_Conf_{str(i + 1)}",
f"{key}_Conf_{str(j + 1)}", path)
except ValueError:
continue
else:
continue
output.write("Comparing " + str(i + 1) + " " + str(j + 1) +
' RMSD ' + str(rms) + "\n")
if rms > rms_tolerance:
pos = _atomic_pos_from_conformer(conf_b[0])
elements = _extract_atomic_type(conf_b[0])
pos = [[-float(coor[k]) for k in range(3)] for coor in pos]
coords = list(zip(elements, pos))
filename = os.path.join(
path, key + "_Conf_" + str(j + 1) + "_inv.xyz")
write_xyz(coords=coords,
filename=filename,
comment=conf_b[1])
try:
file1 = key + "_Conf_" + str(i + 1)
file2 = key + "_Conf_" + str(j + 1) + "_inv"
rmsinv = obfit_rmsd(file1, file2, str(molecule))
except (subprocess.CalledProcessError, ValueError) as e:
if fallback_to_align:
output.write(
'obfit failed, falling back to obabel --align')
output.write(f'Exception {e}\n')
try:
i_key = f"{key}_Conf_{str(i + 1)}"
inv_key = f"{key}_Conf_{str(j + 1)}_inv"
rmsinv = align_rmsd(i_key, inv_key)
except ValueError:
continue
else:
continue
rms = min([rms, rmsinv])
os.remove(key + "_Conf_" + str(j + 1) + "_inv.xyz")
output.write((f"Comparing {i + 1} {j + 1} "
f"RMSD after checking inversion {rms}\n"))
if rms <= rms_tolerance:
self.removed.append(j)
output.write("Removed Conf_" + str(j + 1) + "\n")
os.remove(key + "_Conf_" + str(j + 1) + ".xyz")
def process(self):
data1 = np.load(self.raw_paths[0])
data2 = np.load(self.raw_paths[1])
data1_feed_dict = {
'E': torch.as_tensor(data1['E']),
'N': torch.as_tensor(data1['N']),
'R': torch.as_tensor(data1['R_qm'] if self.qm else data1['R_mmff']),
'D': torch.as_tensor(data1['D_qm'] if self.qm else data1['D_mmff']),
'Q': torch.as_tensor(data1['Q']),
'Z': torch.as_tensor(data1['Z'])
}
data2_feed_dict = {
'E': torch.as_tensor(data2['E']),
'N': torch.as_tensor(data2['N']),
'R': torch.as_tensor(data2['R_qm'] if self.qm else data2['R_mmff']),
'D': torch.as_tensor(data2['D_qm'] if self.qm else data2['D_mmff']),
'Q': torch.as_tensor(data2['Q']),
'Z': torch.as_tensor(data2['Z'])
}
data1_size = data1['E'].shape[0]
data2_size = data2['E'].shape[0]
if not self.sep_heavy_atom:
data_size = data1_size + data2_size
else:
in_part1 = (self.num_heavy_atom < 14)
heavy_atom_data = pd.read_csv(self.raw_paths[2] if in_part1 else self.raw_paths[3])
num_heavy_atom = torch.as_tensor(heavy_atom_data['numberHA']).long()
atom_mask = (num_heavy_atom == self.num_heavy_atom)
atom_mask = atom_mask.view(-1)
data_dict_used = data1_feed_dict if in_part1 else data2_feed_dict
for key in data_dict_used.keys():
data_dict_used[key] = data_dict_used[key][atom_mask]
'''
Here is a trick to make sure later part only calculate data_dict_used
'''
data_size = data_dict_used['E'].shape[0]
data1_feed_dict = data_dict_used
data_array = np.empty(data_size, dtype=Data)
for i in tqdm(range(data_size)):
data_index = i if i < data1_size else i - data1_size
if i < data1_size:
tmp_data = _get_ith_data(data_index, **data1_feed_dict)
else:
tmp_data = _get_ith_data(data_index, **data2_feed_dict)
tmp_data = self.pre_transform(tmp_data, edge_version='cutoff', do_sort_edge=True, cal_efg=False,
cutoff=self.cutoff, boundary_factor=None, use_center=None,
mol=AddHs(MolFromSmiles('C')),
cal_3body_term=self.cal_3body_term, bond_atom_sep=self.bond_atom_sep,
record_long_range=self.record_long_range)
data_array[i] = tmp_data
data_list = [data_array[i] for i in range(data_size)]
print('collating...')
data1, slices = self.collate(data_list)
print('saving...')
torch.save((data1, slices), self.processed_paths[0])
def _addHs(mol, explicitOnly=False, addCoords=False):
return AddHs(mol, explicitOnly=explicitOnly, addCoords=addCoords)
def MolToMol2Block(mol, confId=-1, addHs=False, addCharges=False):
"""Returns a Mol2 string block for a molecule
ARGUMENTS:
- mol: the molecule
- confId: (optional) selects which conformation to output (-1 = default)
if set to None will return all conformers
RETURNS:
a string
"""
#
# References
# - Format specs http://www.tripos.com/data/support/mol2.pdf
# - Atom typing http://www.sdsc.edu/CCMS/Packages/cambridge/pluto/atom_types.html
#
confIds = (confId, )
if confId == None:
confIds = Chem.Mol.GetNumConformers()
blocks = []
# add explicit hydrogens (since mol2 reader requires them)
if addHs:
h_coords = mol.GetNumConformers() > 0 and mol.GetConformer(-1).Is3D()
try:
mol = AddHs(mol, addCoords=h_coords)
except RuntimeError:
mol = AddHs(mol, addCoords=False)
# compute charges
if addCharges:
ComputeGasteigerCharges(mol)
for confId in confIds:
molecule = """@<TRIPOS>MOLECULE
{}
{} {} 0 0 0
SMALL
GASTEIGER\n\n""".format(
mol.GetProp("_Name") if mol.HasProp("_Name") else "UNK",
mol.GetNumAtoms(), mol.GetNumBonds())
# FIXME "USER_CHARGES" could become 'Gasteiger charges'
# FIXME "SMALL" means small molecule but could become "PROTEIN"
pos = _get_positions(mol, confId)
atom_lines = [
"{:>4} {:>4} {:>13.4f} {:>9.4f} {:>9.4f} {:<5} {} {} {:>7.4f}".
format(a.GetIdx() + 1,
a.GetSymbol(),
float(pos[a.GetIdx()][0]),
float(pos[a.GetIdx()][1]),
float(pos[a.GetIdx()][2]),
_sybyl_atom_type(a), 1, "UNL",
float(a.GetProp('_GasteigerCharge').replace(',', '.'))
if a.HasProp('_GasteigerCharge') else 0.0)
for a in mol.GetAtoms()
]
atom_lines = ["@<TRIPOS>ATOM"] + atom_lines
atom_lines = "\n".join(atom_lines) + "\n"
bond_lines = [
"{:>5} {:>5} {:>5} {:>2}".format(
bid + 1,
b.GetBeginAtomIdx() + 1,
b.GetEndAtomIdx() + 1, "ar"
if b.GetBondTypeAsDouble() == 1.5 else "am"
if _amide_bond(b) else str(int(b.GetBondTypeAsDouble())))
for bid, (b) in enumerate(mol.GetBonds())
]
bond_lines = ["@<TRIPOS>BOND"] + bond_lines + ["\n"]
bond_lines = "\n".join(bond_lines)
block = molecule + atom_lines + bond_lines
blocks.append(block)
return "".join(blocks)