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 testNumBitsPerFeature(self):
m1 = Chem.MolFromSmiles('CCCO')
g = rdFingerprintGenerator.GetRDKitFPGenerator(minPath=1, maxPath=2)
fp = g.GetFingerprint(m1)
self.assertEqual(fp.GetNumOnBits(), 8)
g = rdFingerprintGenerator.GetRDKitFPGenerator(minPath=1, maxPath=2, numBitsPerFeature=1)
fp = g.GetFingerprint(m1)
self.assertEqual(fp.GetNumOnBits(), 4)
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 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 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(mols):
"""
Generate RDKit fingerprint from list of molecules.
Parameters
----------
mols : list of rdkit.Chem.rdchem.Mol
List of molecules.
Returns
-------
list of rdkit.DataStructs.cDataStructs.ExplicitBitVect
List of fingerprints.
"""
rdkit_gen = rdFingerprintGenerator.GetRDKitFPGenerator(maxPath=5)
fingerprints = [rdkit_gen.GetFingerprint(mol) for mol in mols]
return fingerprints
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 testRDKitFPGenerator(self):
m = Chem.MolFromSmiles('CCCCC')
g = rdFingerprintGenerator.GetRDKitFPGenerator()
fp = g.GetSparseCountFingerprint(m)
nz = fp.GetNonzeroElements()
self.assertEqual(len(nz), 8)
