def testCountBounds(self):
m = Chem.MolFromSmiles(
'COc1ccc(CCNC(=O)c2ccccc2C(=O)NCCc2ccc(OC)cc2)cc1')
fp1 = rdFingerprintGenerator.GetRDKitFPGenerator(
fpSize=2048, countSimulation=True).GetFingerprint(m)
fp2 = rdFingerprintGenerator.GetRDKitFPGenerator(
fpSize=2048, countSimulation=True,
countBounds=(1, 8, 16, 32)).GetFingerprint(m)
self.assertNotEqual(fp1.GetNumOnBits(), fp2.GetNumOnBits())
fp1 = rdFingerprintGenerator.GetTopologicalTorsionGenerator(
fpSize=2048, countSimulation=True).GetFingerprint(m)
fp2 = rdFingerprintGenerator.GetTopologicalTorsionGenerator(
fpSize=2048, countSimulation=True,
countBounds=(1, 8, 16, 32)).GetFingerprint(m)
self.assertNotEqual(fp1.GetNumOnBits(), fp2.GetNumOnBits())
fp1 = rdFingerprintGenerator.GetMorganGenerator(
fpSize=2048, countSimulation=True).GetFingerprint(m)
fp2 = rdFingerprintGenerator.GetMorganGenerator(
fpSize=2048, countSimulation=True,
countBounds=(1, 8, 16, 32)).GetFingerprint(m)
self.assertNotEqual(fp1.GetNumOnBits(), fp2.GetNumOnBits())
fp1 = rdFingerprintGenerator.GetAtomPairGenerator(
fpSize=2048, countSimulation=True).GetFingerprint(m)
fp2 = rdFingerprintGenerator.GetAtomPairGenerator(
fpSize=2048, countSimulation=True,
countBounds=(1, 8, 16, 32)).GetFingerprint(m)
self.assertNotEqual(fp1.GetNumOnBits(), fp2.GetNumOnBits())
def make_fp_generator(fp_type, settings):
if fp_type == 'morgan':
arguments = {'includeChirality': True}
for arg in settings:
arguments[arg] = settings[arg]
fp_gen = rdFingerprintGenerator.GetMorganGenerator(**arguments)
elif fp_type == 'atom_pair':
arguments = {'includeChirality': True}
for arg in settings:
arguments[arg] = settings[arg]
fp_gen = rdFingerprintGenerator.GetAtomPairGenerator(**arguments)
elif fp_type == 'rdkit':
arguments = {}
for arg in settings:
arguments[arg] = settings[arg]
fp_gen = rdFingerprintGenerator.GetRDKitFPGenerator(**arguments)
elif fp_type == 'toplogical':
arguments = {'includeChirality': True}
for arg in settings:
arguments[arg] = settings[arg]
fp_gen = rdFingerprintGenerator.GetTopologicalTorsionGenerator(
**arguments)
else:
fp_gen = False
return fp_gen
def _get_morgan(self):
gen_mo = rdFingerprintGenerator.GetMorganGenerator()
fps = list()
for mol in self.df['mols']:
fp = np.array(gen_mo.GetFingerprint(mol))
fps.append(fp)
fps = np.array(fps)
return sparse.csr_matrix(fps).astype('int')
def get_morgan_features(mols):
invGen =rdFingerprintGenerator.GetMorganFeatureAtomInvGen()
gen_mo = rdFingerprintGenerator.GetMorganGenerator(atomInvariantsGenerator=invGen)
fps = list()
for mol in mols:
fp = np.array(gen_mo.GetFingerprint(mol))
fps.append(fp)
fps = np.array(fps)
return sparse.csr_matrix(fps).astype('int')
def generate_fingeprints(smiles):
# Load these here so they're only needed on the worker machines.
from rdkit import Chem
from rdkit.Chem import rdFingerprintGenerator
morgan_fp = ''
rdkit_fp = ''
atompair_fp = ''
tt_fp = ''
try:
mol = Chem.MolFromSmiles(smiles)
# Morgan
morgan_fp = rdFingerprintGenerator.GetMorganGenerator().GetFingerprint(
mol).ToBase64()
# Feature Morgan
# TODO
# RDKit
rdkit_fp = rdFingerprintGenerator.GetRDKitFPGenerator().GetFingerprint(
mol).ToBase64()
# Layered
# TODO
# Atom pairs
atompair_fp = rdFingerprintGenerator.GetAtomPairGenerator(
).GetFingerprint(mol).ToBase64()
# MACCS
# TODO
# Topological Torsion
tt_fp = rdFingerprintGenerator.GetTopologicalTorsionGenerator(
).GetFingerprint(mol).ToBase64()
# Pattern
# TODO
# E-state
# TODO
except Exception as e:
print(f'Exception {e} processing {smiles}')
return {}
# NOTE: add any new fingerprints to fingerprint_columns.
return {
'morgan_fp': morgan_fp,
'rdkit_fp': rdkit_fp,
'atompair_fp': atompair_fp,
'tt_fp': tt_fp
}
def testMorganGenerator(self):
m = Chem.MolFromSmiles('CCCC(=O)O')
g = rdFingerprintGenerator.GetMorganGenerator(3)
fp = g.GetSparseCountFingerprint(m)
nz = fp.GetNonzeroElements()
self.assertEqual(len(nz), 14)
invgen = rdFingerprintGenerator.GetMorganAtomInvGen()
g = rdFingerprintGenerator.GetMorganGenerator(radius=3, atomInvariantsGenerator=invgen)
fp = g.GetSparseCountFingerprint(m)
nz = fp.GetNonzeroElements()
self.assertEqual(len(nz), 14)
invgen = rdFingerprintGenerator.GetMorganFeatureAtomInvGen()
g = rdFingerprintGenerator.GetMorganGenerator(radius=3, atomInvariantsGenerator=invgen)
fp = g.GetSparseCountFingerprint(m)
nz = fp.GetNonzeroElements()
self.assertEqual(len(nz), 13)
ms = [Chem.MolFromSmiles(x, sanitize=False) for x in ('C1=CC=CN=N1', 'C1C=CC=NN=1')]
for m in ms:
m.UpdatePropertyCache()
Chem.GetSymmSSSR(m)
g = rdFingerprintGenerator.GetMorganGenerator(radius=2, useBondTypes=True)
self.assertNotEqual(g.GetSparseCountFingerprint(ms[0]), g.GetSparseCountFingerprint(ms[1]))
g = rdFingerprintGenerator.GetMorganGenerator(radius=2, useBondTypes=False)
self.assertEqual(g.GetSparseCountFingerprint(ms[0]), g.GetSparseCountFingerprint(ms[1]))
binvgen = rdFingerprintGenerator.GetMorganBondInvGen(useBondTypes=False)
g2 = rdFingerprintGenerator.GetMorganGenerator(radius=2, bondInvariantsGenerator=binvgen)
self.assertEqual(g.GetSparseCountFingerprint(ms[0]), g2.GetSparseCountFingerprint(ms[0]))
self.assertEqual(g.GetSparseCountFingerprint(ms[1]), g2.GetSparseCountFingerprint(ms[1]))
def write_fingerprints(sesh):
option = st.selectbox('What fingerprint?', (
'choose one',
'Morgan',
))
if option in ['Morgan', 'MACCS']:
if st.button('Generate fingerprints'):
pbar = st.progress(0)
gen_mo = rdFingerprintGenerator.GetMorganGenerator(512)
for count, mol in enumerate(sesh.df['mols']):
fp = rdFingerprintGenerator
sesh.fp.append(gen_mo.GetFingerprint(mol))
pbar.progress(int((count + 1) / len(sesh.df) * 100))
sesh.fp = np.array(sesh.fp)
def calculate_fingerprint(mol, method='morgan'):
rdmol = to_rdkit_Mol(mol)
rdmol.UpdatePropertyCache(strict = False)
Chem.GetSymmSSSR(rdmol)
Dict = {
'rdkit': rdFingerprintGenerator.GetRDKitFPGenerator(fpSize=64),
'morgan': rdFingerprintGenerator.GetMorganGenerator(fpSize=64),
'topological-torsion': rdFingerprintGenerator.GetTopologicalTorsionGenerator(fpSize=64),
'atom-pairs': rdFingerprintGenerator.GetAtomPairGenerator(fpSize=64),
}
rep = []
fp = [Dict[method].GetFingerprint(rdmol, fromAtoms=[i]) for i in range(len(mol.atoms))]
for atomic_fp in fp:
arr = np.zeros((1, ))
DataStructs.ConvertToNumpyArray(atomic_fp, arr)
rep.append(arr)
return np.array(rep)
def generate_fingerprints_and_create_list(self):
#generate fingerprints of predicted ligands and known ligands:
gen_mo = rdFingerprintGenerator.GetMorganGenerator(fpSize=2048,
radius=2)
predicted_fps = [
gen_mo.GetFingerprint(mol) for mol in self.predicted['molecules']
]
true_fps = [
gen_mo.GetFingerprint(mol) for mol in self.true_pos['molecules']
]
similarities = list()
for count, mol in enumerate(predicted_fps):
tanimoto_values = ([
DataStructs.TanimotoSimilarity(mol, i) for i in true_fps
])
index_of_highest = np.argmax(tanimoto_values)
similarities.append(tanimoto_values[index_of_highest])
#module code is in: https://github.com/rdkit/rdkit/tree/master/Contrib/SA_Score
sa_score = [
sascorer.calculateScore(i)
for i in list(self.predicted['molecules'])
]
#create a list holding the QED drug-likeness score
#reference: https://doi.org/10.1038/nchem.1243
qeds = [qed(mol) for mol in self.predicted['molecules']]
#create a list holding logp:
logp = [Descriptors.MolLogP(m) for m in self.predicted['molecules']]
#filter catalog usage instructions are here: https://github.com/rdkit/rdkit/pull/536
params = FilterCatalogParams()
params.AddCatalog(FilterCatalogParams.FilterCatalogs.BRENK)
catalog = FilterCatalog(params)
self.brenk = np.array(
[catalog.HasMatch(m) for m in self.predicted['molecules']])
#add these lists as columns to the 'predicted' pd.DataFrame
self.predicted['similarities'] = similarities
self.predicted['sa_score'] = sa_score
self.predicted['qeds'] = qeds
self.predicted['logp'] = logp
print(self.predicted['logp'] < 6)
shortlist_mask = ((self.predicted['similarities'] < 0.2) &
(self.predicted['sa_score'] < 4) &
(self.predicted['qeds'] > 0.25) &
(self.predicted['logp'] < 6) & (~self.brenk))
def fingerprint_morgan(radius, fpSize=2048, count=False):
"""Morgan fingerprint of the specified size (list of int).
Args:
radius: The number of iterations to grow the fingerprint.
fpSize: Size of the generated fingerprint (defaults to 2048).
count: The default value of False will generate fingerprint bits
(0 or 1) whereas a value of True will generate the count of each
fingerprint value.
"""
generator = rdFingerprintGenerator.GetMorganGenerator(radius=radius,
fpSize=fpSize)
if count:
fingerprint_fn = _fingerprint_fn_count(generator)
else:
fingerprint_fn = _fingerprint_fn_bits(generator)
fingerprint_fn.__name__ = f'fingerprint_morgan(radius={radius},' + \
f'fpSize={fpSize},count={count})'
return fingerprint_fn
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 testBulk(self):
m1 = Chem.MolFromSmiles('CCC')
m2 = Chem.MolFromSmiles('OCCCCC')
m3 = Chem.MolFromSmiles('CCCCC')
g = rdFingerprintGenerator.GetAtomPairGenerator()
results = rdFingerprintGenerator.GetSparseCountFPs(
[m1, m2, m3], rdFingerprintGenerator.AtomPairFP)
self.assertEqual(results[0], g.GetSparseCountFingerprint(m1))
self.assertEqual(results[1], g.GetSparseCountFingerprint(m2))
self.assertEqual(results[2], g.GetSparseCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetMorganGenerator(2)
results = rdFingerprintGenerator.GetSparseCountFPs(
[m1, m2, m3], rdFingerprintGenerator.MorganFP)
self.assertEqual(results[0], g.GetSparseCountFingerprint(m1))
self.assertEqual(results[1], g.GetSparseCountFingerprint(m2))
self.assertEqual(results[2], g.GetSparseCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetRDKitFPGenerator()
results = rdFingerprintGenerator.GetSparseCountFPs(
[m1, m2, m3], rdFingerprintGenerator.RDKitFP)
self.assertEqual(results[0], g.GetSparseCountFingerprint(m1))
self.assertEqual(results[1], g.GetSparseCountFingerprint(m2))
self.assertEqual(results[2], g.GetSparseCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetTopologicalTorsionGenerator()
results = rdFingerprintGenerator.GetSparseCountFPs(
[m1, m2, m3], rdFingerprintGenerator.TopologicalTorsionFP)
self.assertEqual(results[0], g.GetSparseCountFingerprint(m1))
self.assertEqual(results[1], g.GetSparseCountFingerprint(m2))
self.assertEqual(results[2], g.GetSparseCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetAtomPairGenerator()
results = rdFingerprintGenerator.GetSparseFPs(
[m1, m2, m3], rdFingerprintGenerator.AtomPairFP)
self.assertEqual(results[0], g.GetSparseFingerprint(m1))
self.assertEqual(results[1], g.GetSparseFingerprint(m2))
self.assertEqual(results[2], g.GetSparseFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetMorganGenerator(2)
results = rdFingerprintGenerator.GetSparseFPs(
[m1, m2, m3], rdFingerprintGenerator.MorganFP)
self.assertEqual(results[0], g.GetSparseFingerprint(m1))
self.assertEqual(results[1], g.GetSparseFingerprint(m2))
self.assertEqual(results[2], g.GetSparseFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetRDKitFPGenerator()
results = rdFingerprintGenerator.GetSparseFPs(
[m1, m2, m3], rdFingerprintGenerator.RDKitFP)
self.assertEqual(results[0], g.GetSparseFingerprint(m1))
self.assertEqual(results[1], g.GetSparseFingerprint(m2))
self.assertEqual(results[2], g.GetSparseFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetTopologicalTorsionGenerator()
results = rdFingerprintGenerator.GetSparseFPs(
[m1, m2, m3], rdFingerprintGenerator.TopologicalTorsionFP)
self.assertEqual(results[0], g.GetSparseFingerprint(m1))
self.assertEqual(results[1], g.GetSparseFingerprint(m2))
self.assertEqual(results[2], g.GetSparseFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetAtomPairGenerator()
results = rdFingerprintGenerator.GetCountFPs(
[m1, m2, m3], rdFingerprintGenerator.AtomPairFP)
self.assertEqual(results[0], g.GetCountFingerprint(m1))
self.assertEqual(results[1], g.GetCountFingerprint(m2))
self.assertEqual(results[2], g.GetCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetMorganGenerator(2)
results = rdFingerprintGenerator.GetCountFPs(
[m1, m2, m3], rdFingerprintGenerator.MorganFP)
self.assertEqual(results[0], g.GetCountFingerprint(m1))
self.assertEqual(results[1], g.GetCountFingerprint(m2))
self.assertEqual(results[2], g.GetCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetRDKitFPGenerator()
results = rdFingerprintGenerator.GetCountFPs(
[m1, m2, m3], rdFingerprintGenerator.RDKitFP)
self.assertEqual(results[0], g.GetCountFingerprint(m1))
self.assertEqual(results[1], g.GetCountFingerprint(m2))
self.assertEqual(results[2], g.GetCountFingerprint(m3))
self.assertEqual(len(results), 3)
g = rdFingerprintGenerator.GetTopologicalTorsionGenerator()
results = rdFingerprintGenerator.GetCountFPs(
[m1, m2, m3], rdFingerprintGenerator.TopologicalTorsionFP)
self.assertEqual(results[0], g.GetCountFingerprint(m1))
self.assertEqual(results[1], g.GetCountFingerprint(m2))
self.assertEqual(results[2], g.GetCountFingerprint(m3))
self.assertEqual(len(results), 3)
def testMorganGenerator(self):
m = Chem.MolFromSmiles('CCCCC')
g = rdFingerprintGenerator.GetMorganGenerator(3)
fp = g.GetSparseCountFingerprint(m)
nz = fp.GetNonzeroElements()
self.assertEqual(len(nz), 7)
def fingerprint_molecules(mols):
fps = [
rdFingerprintGenerator.GetMorganGenerator().GetFingerprint(mol)
for mol in mols
]
return fps
