def get_filter_values(mol):
"""
calculate the values, for a given molecule, that are used to filter
return as a dictionary
"""
assert isinstance(mol, Chem.Mol)
values = {}
values["MW"] = desc.CalcExactMolWt(mol)
values["logP"] = crip.MolLogP(mol)
values["HBA"] = lip.NumHAcceptors(mol)
values["HBD"] = lip.NumHDonors(mol)
values["tPSA"] = desc.CalcTPSA(mol)
values["rot_bonds"] = lip.NumRotatableBonds(mol)
values["rigid_bonds"] = mol.GetNumBonds() - values["rot_bonds"] # assume mutual exclusion
values["num_rings"] = lip.RingCount(mol)
values["num_hetero_atoms"] = lip.NumHeteroatoms(mol)
values["charge"] = rdmolops.GetFormalCharge(mol) # trusting this charge calculation method
values["num_carbons"], values["num_charges"], values["max_ring_size"] = get_atom_props(mol)
try:
values["hc_ratio"] = float(values["num_hetero_atoms"]) / float(values["num_carbons"])
except ZeroDivisionError:
values["hc_ratio"] = 100000000 # if there are zero carbons
values["fc"] = len(list(Brics.FindBRICSBonds(mol))) # how many BRICS bonds, related to complexity
values["is_good"] = True # default to true, but not yet observed
atoms = [atom.GetSymbol() for atom in mol.GetAtoms()] # get all the atoms, and make the list unique (only types)
atoms = set(atoms)
atoms = list(atoms)
values["atoms"] = atoms
values["num_chiral_centers"] = len(Chem.FindMolChiralCenters(mol, includeUnassigned=True))
values["rejections"] = [] # empty list to store the reasons for rejection
return values
def _generateFPs(mol, fragmentMethod='Morgan'):
aBits = {}
fp = None
# circular Morgan fingerprint fragmentation, we use a simple invariant than ususal here
if fragmentMethod == 'Morgan':
tmp = {}
fp = AllChem.GetMorganFingerprint(
mol,
radius=2,
invariants=utilsFP.generateAtomInvariant(mol),
bitInfo=tmp)
aBits = utilsFP.getMorganEnvironment(mol, tmp, fp=fp, minRad=2)
fp = fp.GetNonzeroElements()
# path-based RDKit fingerprint fragmentation
elif fragmentMethod == 'RDK':
fp = AllChem.UnfoldedRDKFingerprintCountBased(mol,
maxPath=5,
minPath=3,
bitInfo=aBits)
fp = fp.GetNonzeroElements()
# get the final BRICS fragmentation (= smallest possible BRICS fragments of a molecule)
elif fragmentMethod == 'Brics':
fragMol = BRICS.BreakBRICSBonds(mol)
propSmi = _prepBRICSSmiles(fragMol)
fp = Counter(propSmi.split('.'))
else:
print("Unknown fragment method")
return fp, aBits
def get_ddi_mask(atc42SMLES, med_voc):
# ATC3_List[22] = {0}
# ATC3_List[25] = {0}
# ATC3_List[27] = {0}
fraction = []
for k, v in med_voc.idx2word.items():
tempF = set()
for SMILES in atc42SMLES[v]:
try:
m = BRICS.BRICSDecompose(Chem.MolFromSmiles(SMILES))
for frac in m:
tempF.add(frac)
except:
pass
fraction.append(tempF)
fracSet = []
for i in fraction:
fracSet += i
fracSet = list(set(fracSet)) # set of all segments
ddi_matrix = np.zeros((len(med_voc.idx2word), len(fracSet)))
for i, fracList in enumerate(fraction):
for frac in fracList:
ddi_matrix[i, fracSet.index(frac)] = 1
return ddi_matrix
def pair_frags(fname, out, method='Recap', is_mf=True):
smiles = pd.read_table(fname).Smiles.dropna()
pairs = []
for i, smile in enumerate(tqdm(smiles)):
smile = utils.clean_mol(smile)
mol = Chem.MolFromSmiles(smile)
if method == 'recap':
frags = np.array(sorted(Recap.RecapDecompose(mol).GetLeaves().keys()))
else:
frags = BRICS.BRICSDecompose(mol)
frags = np.array(sorted({re.sub(r'\[\d+\*\]', '*', f) for f in frags}))
if len(frags) == 1: continue
du, hy = Chem.MolFromSmiles('*'), Chem.MolFromSmiles('[H]')
subs = np.array([Chem.MolFromSmiles(f) for f in frags])
subs = np.array([Chem.RemoveHs(Chem.ReplaceSubstructs(f, du, hy, replaceAll=True)[0]) for f in subs])
subs = np.array([m for m in subs if m.GetNumAtoms() > 1])
match = np.array([[m.HasSubstructMatch(f) for f in subs] for m in subs])
frags = subs[match.sum(axis=0) == 1]
frags = sorted(frags, key=lambda x:-x.GetNumAtoms())[:voc.n_frags]
frags = [Chem.MolToSmiles(Chem.RemoveHs(f)) for f in frags]
max_comb = len(frags) if is_mf else 1
for ix in range(1, max_comb+1):
combs = combinations(frags, ix)
for comb in combs:
input = '.'.join(comb)
if len(input) > len(smile): continue
if mol.HasSubstructMatch(Chem.MolFromSmarts(input)):
pairs.append([input, smile])
df = pd.DataFrame(pairs, columns=['Frags', 'Smiles'])
df.to_csv(out, sep='\t', index=False)
def get_sgs(record_dict, n_min, n_max, method="exhaustive"):
if method == "exhaustive":
return Chem.rdmolops.FindAllSubgraphsOfLengthMToN(
record_dict["mol"], n_min, n_max)
elif method == "RECAP":
hierarchy = Recap.RecapDecompose(record_dict["mol"])
sgs = []
for substructure in hierarchy.GetAllChildren().values():
substructure = Chem.DeleteSubstructs(substructure.mol,
Chem.MolFromSmarts('[#0]'))
edge_idxs = get_substructure_bond_idx(substructure,
record_dict["mol"])
if edge_idxs is not None:
sgs.append(edge_idxs)
return subset_sgs_sizes([sgs], n_min, n_max)
elif method == "BRICS":
substructures = BRICS.BRICSDecompose(record_dict["mol"])
sgs = []
for substructure in substructures:
substructure = Chem.DeleteSubstructs(
Chem.MolFromSmiles(substructure), Chem.MolFromSmarts('[#0]'))
edge_idxs = get_substructure_bond_idx(substructure,
record_dict["mol"])
if edge_idxs is not None:
sgs.append(edge_idxs)
return subset_sgs_sizes([sgs], n_min, n_max)
def brics(
mol: Chem.Mol,
singlepass: bool = True,
remove_parent: bool = False,
sanitize: bool = True,
fix: bool = True,
):
"""Run BRICS on the molecules and potentially fix dummy atoms.
Args:
mol: a molecule.
singlepass: Single pass for `BRICSDecompose`.
remove_parent: Remove parent from the fragments.
sanitize: Wether to sanitize the fragments.
fix: Wether to fix the fragments.
"""
frags = BRICS.BRICSDecompose(mol, returnMols=True, singlePass=singlepass)
frags = list(frags)
if fix:
frags = [dm.fix_mol(x) for x in frags]
if sanitize:
frags = [dm.sanitize_mol(x) for x in frags]
if remove_parent:
frags.pop(0)
frags = [x for x in frags if x is not None]
return frags
def generate_chemicals_from_fragments(smiles_list, n=10):
"""
reconstruct chemicals from fragments
Paramters
-----------------
smiles_list: list of string
list of smiles of fragments
n: int
number of chemicals to be generated
Returns
---------------
smiles_list: list of string
list of newly generated smiles
"""
# convert smiles to mol objects
all_components = [Chem.MolFromSmiles(f) for f in smiles_list]
builder = BRICS.BRICSBuild(all_components)
generated_mol_list = []
for i in (range(n)):
m = next(builder)
m.UpdatePropertyCache(strict=True)
generated_mol_list.append(m)
smiles_list = [Chem.MolToSmiles(m) for m in generated_mol_list]
return smiles_list
def fragmentate_chemicals(SMILES_list, return_only_fragments=True):
"""
fragmentate chemicals by BRICs algorithm
Parameters
---------------------
SMILES_list: list of string
list of smiles
return_only_fragments: bool
if true, return only fragment parts
Returns
----------------------
fragmentated_smiles: list of string
list of fragmentated chemicals
"""
mols = [Chem.MolFromSmiles(SMILES) for SMILES in SMILES_list]
fragmentated_smiles = [BRICS.BRICSDecompose(mol) for mol in tqdm(mols)]
# nested list to normal list
fragmentated_smiles = (list(
itertools.chain.from_iterable(fragmentated_smiles)))
fragmentated_smiles = list(set(fragmentated_smiles))
if return_only_fragments:
fragmentated_smiles = [
i for i in fragmentated_smiles if i.find("*") > 0
]
return fragmentated_smiles
def fragment_iterative(mol, min_length=3):
bond_data = list(BRICS.FindBRICSBonds(mol))
try:
idxs, labs = zip(*bond_data)
except Exception:
return []
bonds = []
for a1, a2 in idxs:
bond = mol.GetBondBetweenAtoms(a1, a2)
bonds.append(bond.GetIdx())
order = np.argsort(bonds).tolist()
bonds = [bonds[i] for i in order]
frags, temp = [], deepcopy(mol)
for bond in bonds:
res = break_on_bond(temp, bond)
if len(res) == 1:
frags.append(temp)
break
head, tail = res
if get_size(head) < min_length or get_size(tail) < min_length:
continue
frags.append(head)
temp = deepcopy(tail)
return frags
def fragment_recursive(mol, frags):
try:
bonds = list(BRICS.FindBRICSBonds(mol))
if bonds == []:
frags.append(mol)
return frags
idxs, labs = list(zip(*bonds))
bond_idxs = []
for a1, a2 in idxs:
bond = mol.GetBondBetweenAtoms(a1, a2)
bond_idxs.append(bond.GetIdx())
order = np.argsort(bond_idxs).tolist()
bond_idxs = [bond_idxs[i] for i in order]
broken = Chem.FragmentOnBonds(mol,
bondIndices=[bond_idxs[0]],
dummyLabels=[(0, 0)])
head, tail = Chem.GetMolFrags(broken, asMols=True)
print(mol_to_smiles(head), mol_to_smiles(tail))
frags.append(head)
fragment_recursive(tail, frags)
except Exception:
pass
def fragmenter(thefile):
os.remove('output.txt')
id = []
for line in open(thefile):
line = line.strip()
id.append(line)
df = pd.DataFrame()
df = id
count = 0
mylist = []
for y in df:
base = Chem.MolFromSmiles(df[count])
catalog = BRICS.BRICSDecompose(base)
mcat = [Chem.MolFromSmiles(x) for x in catalog]
ms = BRICS.BRICSBuild(mcat)
for m in ms:
a = Chem.MolToSmiles(m)
mylist.append(a)
count = count + 1
df2 = pd.DataFrame({'smiles': mylist})
f3 = open('output.txt', 'w+')
for j in mylist:
print(j, file=f3)
f3.close()
return mylist
def break_on_rotatable_bonds_to_mol(inmol):
"""Takes a mol and breaks it on all of the rotatable bonds (well, most of them) - returns a single mol, or None if there are no rotatable bonds"""
# Get the indices of the atoms around the bonds to be broken
atom_pairs = [p[0] for p in BRICS.FindBRICSBonds(inmol) if p]
# Return if no bonds found... as it was given
if not atom_pairs:
return inmol
# Get the bond indices
bonds = [
inmol.GetBondBetweenAtoms(at1, at2).GetIdx()
for (at1, at2) in atom_pairs
]
# Fragment the molecule
fragged_mol = Chem.rdmolops.FragmentOnBonds(inmol, bonds, addDummies=False)
return fragged_mol
def identify_rotatable_bond_atom_pairs(mol):
"""find the atom quadruplets around rotatable bonds of a molecule"""
# List of tuples of 4 atoms
atom_sets = []
# Get the atoms on the ends of rotatable bonds
atom_pairs = [p[0] for p in BRICS.FindBRICSBonds(mol) if p]
# Go through and get one of the neighbours for each
for a1,a2 in atom_pairs:
# Get the neighbours for a1 (removing a2)
a1_neighbours = [n.GetIdx() for n in mol.GetAtomWithIdx(a1).GetNeighbors()]
a1_neighbours.remove(a2)
# Get the neighbours for a2 (removing a1)
a2_neighbours = [n.GetIdx() for n in mol.GetAtomWithIdx(a2).GetNeighbors()]
a2_neighbours.remove(a1)
# Add one from either side of the double bond
atom_sets.append((a1_neighbours[0], a1, a2, a2_neighbours[0]))
# Now have 4 atoms from which we can calculate a dihedral angle
return atom_sets
def splitMol(self, mol, bondsToKeep):
''' fragments a molecule on a particular set of BRICS bonds.
Partially sanitizes the results
'''
bbnds = BRICS.FindBRICSBonds(mol)
bndsToTry = []
lbls = []
for aids, lbl in bbnds:
if lbl in bondsToKeep:
bndsToTry.append(mol.GetBondBetweenAtoms(
aids[0], aids[1]).GetIdx())
lbls.append([int(x) for x in lbl])
if not bndsToTry:
return []
res = Chem.FragmentOnSomeBonds(mol, bndsToTry, dummyLabels=lbls)
# We need at least a partial sanitization for the rest of what we will be doing:
for entry in res:
entry.UpdatePropertyCache(False)
Chem.FastFindRings(entry)
return res
def combine_frag(self):
self.generate_frag_templates()
print('Merging fragments together to generate compounds...')
for current_template in self.potential_cpd_templates:
fragms = [Chem.MolFromSmiles(x) for x in sorted(current_template)]
ms = BRICS.BRICSBuild(fragms)
prods = [next(ms) for x in range(1)]
# mini_frags = self.collect_mini_frags_from_each_template(current_template)
# percent = len(mini_frags)
# counter = 0
for i in range(1):
# for j in range(len(mini_frags)):
sampler = Chem.MolToSmiles(prods[i], True)
# if mini_frags[j] in sampler:
# counter+=1
# if counter == percent:
if sampler not in self.templates:
print(sampler)
self.templates.append(sampler)
self.chembank.write(sampler + '\n')
def fragment_database(self):
fName = 'C:/RDKit_2017_03_2/Data/FunctionalGroups.txt'
fparams = FragmentCatalog.FragCatParams(1, 6, fName)
self.fcat = FragmentCatalog.FragCatalog(fparams)
## macrocycle_file = 'macrocycles_IDs.csv'
## suppl = [i.split(',')[0] for i in open(self.directory+name,'r').read().splitlines()][1:] # read all the macrocycle smiles from file
## ms = [Chem.MolFromSmiles(i) for i in suppl] # mols of macrocycles
zinc_file = 'data/smiles_database.csv'
zinc_suppl = [
i.split(',')[1]
for i in open(self.directory + zinc_file, 'r').read().splitlines()
][1:]
zinc_ms = [Chem.MolFromSmiles(i) for i in zinc_suppl]
pre_synthetic_frag_database = [
BRICS.BRICSDecompose(i) for i in zinc_ms
]
self.synthetic_frag_database = list(
set(chain.from_iterable(pre_synthetic_frag_database)))
def getBits(mol):
'''
Parameters
----------
mol : rdkit mol object to be broken up into fragments by breaking
rotable bonds
Returns
-------
mols : A list of rdkit mol objects
'''
# find the rotatable bonds
bonds = mol.GetSubstructMatches(RotatableBondSmarts)
bonds = [((x, y), (0, 0)) for x, y in bonds]
p = BRICS.BreakBRICSBonds(mol, bonds=bonds)
mols = [mol for mol in Chem.GetMolFrags(p, asMols=True)]
return mols
def __init__(self, radius, fpSize, IC50function, molFile):
self.fpgen = rdFingerprintGenerator.GetMorganGenerator(
radius=radius, fpSize=fpSize)
self.getIC50 = IC50function
self.molFile = molFile
# Open SMILES file and convert each sequence to rdkit molecule
with open(self.molFile) as f:
raw_text = f.read()
raw_data = raw_text.split("\n")
mol_list = [Chem.MolFromSmiles(x) for x in raw_data[:1000]]
self.ms = [rdMolStandardize.FragmentParent(x) for x in mol_list]
# Get a count of the BRICS bonds within the molecules
cntr = Counter()
for m in self.ms:
bbnds = BRICS.FindBRICSBonds(m)
for aids, lbls in bbnds:
cntr[lbls] += 1
freqs = sorted([(y, x) for x, y in cntr.items()], reverse=True)
# Keep the top 10 bonds
self.bondsToKeep = [y for x, y in freqs]
def get_scaffolds(self,
scaffolding_method=ScaffoldingMethod.MurckoScaffold):
"""Compute deemed scaffolds for a given compound.
Args:
scaffolding_method (ScaffoldingMethod, optional):
Defaults to MurckoScaffold. Scaffolding method to use
Returns:
list[rdkit.Chem.rdchem.Mol]: Scaffolds found in the component.
"""
try:
scaffolds = []
if scaffolding_method == ScaffoldingMethod.MurckoScaffold:
scaffolds = [(MurckoScaffold.GetScaffoldForMol(self.mol_no_h))]
elif scaffolding_method == ScaffoldingMethod.MurckoGeneric:
scaffolds = [
(MurckoScaffold.MakeScaffoldGeneric(self.mol_no_h))
]
elif scaffolding_method == ScaffoldingMethod.Brics:
scaffolds = BRICS.BRICSDecompose(self.mol_no_h)
brics_smiles = [
re.sub(r"(\[[0-9]*\*\])", "[H]", i) for i in scaffolds
] # replace dummy atoms with H's to get matches https://sourceforge.net/p/rdkit/mailman/message/35261974/
brics_mols = [
rdkit.Chem.MolFromSmiles(x) for x in brics_smiles
]
for mol in brics_mols:
rdkit.Chem.RemoveHs(mol)
brics_hits = [
self.mol_no_h.GetSubstructMatches(i) for i in brics_mols
]
for index, brics_hit in enumerate(brics_hits):
smiles = rdkit.Chem.MolToSmiles(brics_mols[index])
name = scaffolding_method.name
source = 'RDKit scaffolds'
key = f'{name}_{smiles}'
brics_hit = conversions.listit(brics_hit)
if not smiles:
continue
if key not in self._scaffolds:
self._scaffolds[key] = SubstructureMapping(
name, smiles, source, brics_hit)
return brics_mols
for s in scaffolds:
scaffold_atom_names = [
atom.GetProp('name') for atom in s.GetAtoms()
]
mapping = []
for at_name in scaffold_atom_names:
idx = [
atom.GetIdx() for atom in self.mol.GetAtoms()
if atom.GetProp('name') == at_name
][0]
mapping.append(idx)
smiles = rdkit.Chem.MolToSmiles(s)
name = scaffolding_method.name
source = 'RDKit scaffolds'
if not smiles:
continue
if name in self._scaffolds:
self._scaffolds[name].mappings.append(mapping)
else:
self._scaffolds[name] = SubstructureMapping(
name, smiles, source, [mapping])
return scaffolds
except (RuntimeError, ValueError):
raise CCDUtilsError(
f'Computing scaffolds using method {scaffolding_method.name} failed.'
)
print 'Argument List:', str(sys.argv)
if len(sys.argv ) == 3 :
inp_sdf_file = sys.argv[1]
out_sdf_file = sys.argv[2]
else:
sys.exit ("Usage: fragmenter.py infile outfile")
try:
suppl = Chem.SDMolSupplier(inp_sdf_file)
catalog=set()
for mol in suppl:
if mol is None: continue
print mol.GetNumAtoms()
#AllChem.Compute2DCoords(mol)
pieces = BRICS.BRICSDecompose(mol)
catalog.update(pieces)
print('Generated: ', len(catalog), ' fragments.')
ofile = Chem.SDWriter(out_sdf_file)
for frg in catalog:
cmol = Chem.MolFromSmiles(frg)
AllChem.Compute2DCoords(cmol)
ofile.write(cmol)
except IOError:
print >> sys.stderr, "Input file could not be opened"
sys.exit(1)
fgData = """AcidChloride C(=O)Cl Acid Chloride
CarboxylicAcid C(=O)[O;H,-] Carboxylic acid
SulfonylChloride [$(S-!@[#6])](=O)(=O)(Cl) Sulfonyl Chloride
Amine [N;!H0;$(N-[#6]);!$(N-[!#6]);!$(N-C=[O,N,S])] Amine
BoronicAcid [$(B-!@[#6])](O)(O) Boronic Acid
Isocyanate [$(N-!@[#6])](=!@C=!@O) Isocyanate
Alcohol [O;H1;$(O-!@[#6;!$(C=!@[O,N,S])])] Alcohol
Aldehyde [CH;D2;!$(C-[!#6])]=O Aldehyde
Halogen [$([Cl,Br,I]-!@[#6]);!$([Cl,Br,I]-!@C-!@[F,Cl,Br,I]);!$([Cl,Br,I]-[C,S](=[O,S,N]))] Halogen"""
fglines = [re.split(r'\t+', x.strip()) for x in fgData.split('\n')]
hLabels = [x[0] for x in fglines]
patts = [Chem.MolFromSmarts(x[1]) for x in fglines]
labels = inLs[0].strip().split(delim) + hLabels + ['HasBRICSBond?']
print(delim.join(labels))
for line in inLs[1:]:
splitL = line.strip().split(delim)
mol = Chem.MolFromSmiles(splitL[1])
for fg in patts:
if mol.HasSubstructMatch(fg):
splitL.append('True')
else:
splitL.append('False')
bricsRes = BRICS.BRICSDecompose(mol)
if len(bricsRes) > 1:
splitL.append('True')
else:
splitL.append('False')
print(delim.join(splitL))
random.seed(1)
max_number_of_generated_structures = 100
molecules = [
molecule for molecule in Chem.SDMolSupplier('logSdataset1290_2d.sdf')
if molecule is not None
]
# molecules = [molecule for molecule in Chem.SmilesMolSupplier('logSdataset1290_2d.smi',
# delimiter='\t', titleLine=False)
# if molecule is not None]
print(len(molecules))
fragments = set()
for molecule in molecules:
fragment = BRICS.BRICSDecompose(molecule, minFragmentSize=2)
# print(fragment)
# print(list(BRICS.FindBRICSBonds(molecule)))
fragments.update(fragment)
print(len(fragments))
# print (fragments)
generated_structures = BRICS.BRICSBuild(
[Chem.MolFromSmiles(smiles) for smiles in fragments])
writer = Chem.SDWriter('generated_structures.sdf')
# writer = Chem.SmilesWriter('generated_structures.smi')
number_of_generated_structures = 0
for generated_structure in generated_structures:
generated_structure.UpdatePropertyCache(True)
AllChem.Compute2DCoords(generated_structure)
writer.write(generated_structure)
def get_fc0_frags(self):
"""
This function breaks every bond and enumerates all the FC0 fragments. This is a needed first step, since these
are used to build up all the more complex fragments.
:return:
"""
fc0_frags = []
brics = BRICS.FindBRICSBonds(self.init_mol)
brics_counter = Counter(
counter=100
) # counter starts at 100 since it's not going to be in use as a p.atom
edges = []
for bond, brics_type in brics:
i, j = bond
x, y = brics_type
pseudo_i = brics_counter.increment()
pseudo_j = brics_counter.increment(
) # this is always done in pairs so that eg. 100 and 101 go together
self.brics_legend[pseudo_i] = int(
x) # remember to store the BRICS typing, we need it later
self.brics_legend[pseudo_j] = int(y)
edges.append((pseudo_i, pseudo_j)) # keep a record of this bond
self.breakAndReplace(i, j, pseudo_i, pseudo_j)
self.edges = edges
# this while loop finds parts of a graph that are no longer connected, meaning FC0 fragments in this case
untouched = set(self.atoms.keys()
) # keep a list of atoms we haven't investigated yet
while untouched: # if there are still atoms we haven't processed
currSet = set()
currSet.add(untouched.pop())
newMol = RecomposerMol()
newAtoms = {}
newPseudoAtoms = {}
new_brics_legend = {}
new_brics_legend.update(self.brics_legend)
while currSet:
# perform graph search
currAtom = self.atoms[currSet.pop()]
for bond in currAtom.bonds:
idx = None
if currAtom.index == bond.endAtomIdx:
idx = bond.beginAtomIdx
elif currAtom.index == bond.beginAtomIdx:
idx = bond.endAtomIdx
# add connected atoms to curr mol atom set
if idx in untouched:
currSet.add(idx)
untouched.remove(idx)
else:
continue
newAtoms[currAtom.index] = self._copyAtom(
currAtom) # watch out for reference errors!
if currAtom.atom_num == 0:
newPseudoAtoms[currAtom.isotope] = currAtom.index
newMol.atoms = newAtoms
newMol.pseudoIndex = newPseudoAtoms
newMol.brics_legend = new_brics_legend
fc0_frags.append(newMol)
self.frag_cache.append(newMol)
# now need to index the fragments to make it easier to bond them
fc0_index = {}
for i, frag in enumerate(fc0_frags):
for pseudo_atom in frag.pseudoIndex.keys():
fc0_index[
pseudo_atom] = i # each FC0 fragment should be accessible by pseudoatom
self.fc0_index = fc0_index
self.fragments[0] = fc0_frags
def fragment_target(self):
self.target_fragments = list(
BRICS.BRICSDecompose(Chem.MolFromSmiles(self.target)))
if (len(Chem.MolToSmiles(rm3).split(".")) > 1):
count1 = count1 + 1
[functional_list.append(i) for i in Chem.MolToSmiles(rm3).split(".")]
else:
if(len(Chem.MolToSmiles(rm2).split(".")) > 1):
count2 = count2 + 1
[functional_list.append(i) for i in Chem.MolToSmiles(rm2).split(".")]
else:
if(len(Chem.MolToSmiles(rm).split(".")) > 1):
count3 = count3 + 1
[functional_list.append(i) for i in Chem.MolToSmiles(rm).split(".")]
else:
pieces_smi = Chem.BRICS.BRICSDecompose(temp)
pieces = [Chem.MolFromSmiles(x) for x in BRICS.BRICSDecompose(temp)]
count_fail_no_match += 1
print(can_smi)
except:
count_fail += 1
print(list(set(functional_list))) #retrieve only the found functional groups
print(len(list(set(functional_list))))
print(count1, count2, count3)
print("total processed: "+ str(count1+count2+count3))
print("no substructured: "+ str(count_fail_no_match))
print("fail processed: "+ str(count_fail))
def update_atom_position(mol1, mol2):
mol_copy = Chem.Mol(mol2)
# This is a work-around to get a seedSmarts for the FMCS algorithm
# and prevent the occassional hanging of FMCS
# Might be unnecessary with future versions of rdkit
core_frags = BRICS.BRICSDecompose(Chem.RemoveHs(mol1))
frag_smarts = []
for frag in enumerate(core_frags):
smi_str = (re.sub('[[1-9][0-9]{0,2}\*]', '[*]', frag[1]))
frag_smarts.append(
Chem.MolToSmarts(Chem.MolFromSmiles(smi_str)).replace(
':', '~').replace('-', '~').replace('=',
'~').replace('#0', '*'))
seed = None
for query in frag_smarts:
if mol_copy.HasSubstructMatch(Chem.MolFromSmarts(query)):
seed = query
break
# Now get MCSS
res = rdFMCS.FindMCS([mol1, mol_copy], seedSmarts=seed)
mcs_q = Chem.MolFromSmarts(res.smartsString)
# Get atom IDs
template = mol1.GetSubstructMatches(mcs_q)[0]
hit_atom = mol_copy.GetSubstructMatches(mcs_q)[0]
# Update XYZ coords of MCSS
running_distance = 0
for i in range(0, len(template)):
origin = mol1.GetConformer().GetAtomPosition(template[i])
pos = mol_copy.GetConformer().GetAtomPosition(hit_atom[i])
p1 = np.array([origin.x, origin.y, origin.z])
p2 = np.array([pos.x, pos.y, pos.z])
sq_dist = np.sum((p1 - p2)**2, axis=0)
dist = np.sqrt(sq_dist)
running_distance += dist
mol_copy.GetConformer().SetAtomPosition(hit_atom[i],
(origin.x, origin.y, origin.z))
if running_distance > 0.1:
# relax atoms outside MCSS
res_atom = []
for atom in mol_copy.GetAtoms():
if atom.GetIdx() not in hit_atom:
res_atom.append(atom.GetIdx())
# do minimization
mp = ChemicalForceFields.MMFFGetMoleculeProperties(mol_copy)
ff = ChemicalForceFields.MMFFGetMoleculeForceField(mol_copy, mp)
for val in hit_atom:
ff.AddFixedPoint(val)
for val in res_atom:
ff.MMFFAddPositionConstraint(val, 1, 5)
ff.Minimize()
return mol_copy
number_of_free_bonds = 0 # The number of free bond(s) restricted. If 0, all fragments are saved
# load molecules
molecules = [
molecule for molecule in Chem.SmilesMolSupplier(
'logS_molecules_1290.smi', delimiter='\t', titleLine=False)
if molecule is not None
]
# molecules = [molecule for molecule in Chem.SDMolSupplier('logSdataset1290_2d.sdf') if molecule is not None]
print('number of molecules :', len(molecules))
# generate fragments
fragments = set()
for molecule in molecules:
fragment = BRICS.BRICSDecompose(molecule, minFragmentSize=2)
fragments.update(fragment)
# select and arange fragments
new_fragments = []
number_of_generated_structures = 0
for fragment in fragments:
free_bond = []
free_bond = [index for index, atom in enumerate(fragment) if atom == '*']
flag = False
if number_of_free_bonds == 0:
if len(free_bond):
flag = True
else:
if len(free_bond) == number_of_free_bonds:
flag = True
# In[ ]: # In[8]: from rdkit.Chem import AllChem from rdkit.Chem import BRICS from rdkit.Chem import rdMolDescriptors # In[21]: smiles = Chem.MolToSmiles(mol) frag = BRICS.BRICSDecompose(mol) print(smiles) # In[20]: print(len(frag)) # In[11]: print(frag) # In[12]:
if tests[9]:
logger.info('Writing: Mol blocks')
t1 = time.time()
for mol in mols:
mb = Chem.MolToMolBlock(mol)
t2 = time.time()
logger.info('Results10: %.2f seconds' % (t2 - t1))
ts.append(t2 - t1)
if tests[10]:
from rdkit.Chem import BRICS
logger.info('BRICS decomposition')
t1 = time.time()
for mol in mols:
d = BRICS.BreakBRICSBonds(mol)
t2 = time.time()
logger.info('Results11: %.2f seconds' % (t2 - t1))
ts.append(t2 - t1)
if tests[11]:
logger.info('Generate 2D coords')
t1 = time.time()
for mol in mols:
AllChem.Compute2DCoords(mol)
t2 = time.time()
logger.info('Results12: %.2f seconds' % (t2 - t1))
ts.append(t2 - t1)
if tests[12]:
logger.info('Generate topological fingerprints')
# ['*C(=O)CC', '*CCOC(=O)CC', '*CCOc1ccccc1', '*OCCOc1ccccc1', '*c1ccccc1']
# # 3.2 BRICS方法
# RDKit 还提供了另一种把分子切成片段的方法——BRICS方法。 BRICS方法主要是根据可合成的的键对分子进行切断，因此其返回的数据结构是来自于该分子的不同分子片段， 虚拟原子（*）是告诉我们是如何切断的。
# 对下图中的分子进行BRICS分解
smi = 'C=CC(=O)N1CCC(CC1)C2CCNC3=C(C(=NN23)C4=CC=C(C=C4)OC5=CC=CC=C5)C(=O)N'
m = Chem.MolFromSmiles(smi)
Draw.MolToImageFile(
m,
"/drug_development/studyRdkit/st_rdcit/img/mol34.jpg",
size=(600, 400),
legend=
'zanubrutinib(C=CC(=O)N1CCC(CC1)C2CCNC3=C(C(=NN23)C4=CC=C(C=C4)OC5=CC=CC=C5)C(=O)N)'
)
frags = (BRICS.BRICSDecompose(m))
print(frags)
mols = []
for fsmi in frags:
mols.append(Chem.MolFromSmiles(fsmi))
img = Draw.MolsToGridImage(mols,
molsPerRow=3,
subImgSize=(200, 200),
legends=['' for x in mols])
img.save('/drug_development/studyRdkit/st_rdcit/img/mol35.jpg')
# 四、组合分子片段--BRICS方法
# 以上述片段进行BRICS组合产生分子
newms = BRICS.BRICSBuild(mols)
newms = list(newms)