def testOrderDepend(self):
""" test order dependence of some descriptors:
"""
data = [('C=CC=C',21.01955,2.73205),
('O=CC=O',25.01955,2.73205),
('FCC(=O)CF',46.7548875,2.98816),
('O=C1C=CC(=O)C=C1',148.705216,2.8265),
('C12C(F)=C(O)C(F)C1C(F)=C(O)C(F)2',315.250442,2.4509),
('C12CC=CCC1C(=O)C3CC=CCC3C(=O)2',321.539522,1.95986)]
for smi,CT,bal in data:
m = Chem.MolFromSmiles(smi)
newBal = GraphDescriptors.BalabanJ(m, forceDMat = 1)
assert feq(newBal,bal,1e-4),'mol %s %f!=%f'%(smi,newBal,bal)
m = Chem.MolFromSmiles(smi)
newCT = GraphDescriptors.BertzCT(m, forceDMat = 1)
assert feq(newCT,CT,1e-4),'mol %s (CT calc = %f) should have CT = %f'%(smi,newCT,CT)
m = Chem.MolFromSmiles(smi)
newCT = GraphDescriptors.BertzCT(m, forceDMat = 1)
assert feq(newCT,CT,1e-4),'mol %s (CT calc = %f) should have CT = %f'%(smi,newCT,CT)
newBal = GraphDescriptors.BalabanJ(m, forceDMat = 1)
assert feq(newBal,bal,1e-4),'mol %s %f!=%f'%(smi,newBal,bal)
m = Chem.MolFromSmiles(smi)
newBal = GraphDescriptors.BalabanJ(m, forceDMat = 1)
assert feq(newBal,bal,1e-4),'mol %s %f!=%f'%(smi,newBal,bal)
newCT = GraphDescriptors.BertzCT(m, forceDMat = 1)
assert feq(newCT,CT,1e-4),'mol %s (CT calc = %f) should have CT = %f'%(smi,newCT,CT)
def testIssue125(self):
# test an issue with calculating BalabanJ
smi = 'O=C(OC)C1=C(C)NC(C)=C(C(OC)=O)C1C2=CC=CC=C2[N+]([O-])=O'
m1 = Chem.MolFromSmiles(smi)
m2 = Chem.MolFromSmiles(smi)
Chem.MolToSmiles(m1)
j1 = GraphDescriptors.BalabanJ(m1)
j2 = GraphDescriptors.BalabanJ(m2)
assert feq(j1, j2)
def testBalabanJ(self):
""" test calculation of the Balaban J value
J values are from Balaban's paper and have had roundoff
errors and typos corrected.
"""
data = [# alkanes
('CC',1.0),('CCC',1.6330),
('CCCC',1.9747),('CC(C)C',2.3238),
('CCCCC',2.1906),('CC(C)CC',2.5396),('CC(C)(C)C',3.0237),
('CCCCCC',2.3391),('CC(C)CCC',2.6272),('CCC(C)CC',2.7542),('CC(C)(C)CC',3.1685),
('CC(C)C(C)C',2.9935),
# cycloalkanes
('C1CCCCC1',2.0000),
('C1C(C)CCCC1',2.1229),
('C1C(CC)CCCC1',2.1250),
('C1C(C)C(C)CCC1',2.2794),
('C1C(C)CC(C)CC1',2.2307),
('C1C(C)CCC(C)C1',2.1924),
('C1C(CCC)CCCC1',2.0779),
('C1C(C(C)C)CCCC1',2.2284),
('C1C(CC)C(C)CCC1',2.2973),
('C1C(CC)CC(C)CC1',2.2317),
('C1C(CC)CCC(C)C1',2.1804),
('C1C(C)C(C)C(C)CC1',2.4133),
('C1C(C)C(C)CC(C)C1',2.3462),
('C1C(C)CC(C)CC1(C)',2.3409),
# aromatics
('c1ccccc1',3.0000),
('c1c(C)cccc1',3.0215),
('c1c(CC)cccc1',2.8321),
('c1c(C)c(C)ccc1',3.1349),
('c1c(C)cc(C)cc1',3.0777),
('c1c(C)ccc(C)c1',3.0325),
('c1c(CCC)cccc1',2.6149),
('c1c(C(C)C)cccc1',2.8483),
('c1c(CC)c(C)ccc1',3.0065),
('c1c(CC)cc(C)cc1',2.9369),
('c1c(CC)ccc(C)c1',2.8816),
('c1c(C)c(C)c(C)cc1',3.2478),
('c1c(C)c(C)cc(C)c1',3.1717),
('c1c(C)cc(C)cc1(C)',3.1657)
]
for smi,res in data:
m = Chem.MolFromSmiles(smi)
j = GraphDescriptors.BalabanJ(m,forceDMat=1)
assert feq(j,res),'mol %s (J=%f) should have J=%f'%(smi,j,res)
j = GraphDescriptors.BalabanJ(m)
assert feq(j,res),'second pass: mol %s (J=%f) should have J=%f'%(smi,j,res)
'HBD', 'jIndex'
]
for name in prop_names:
d[f'{name}'] = []
for i, s in enumerate(smiles):
if (i % 10000 == 0):
print(i)
m = Chem.MolFromSmiles(s)
if (m == None or 'i' in s or '.' in s):
DUD = DUD.drop(i)
print(s, i)
else:
d['QED'].append(QED.default(m))
d['logP'].append(Crippen.MolLogP(m))
d['molWt'].append(Descriptors.MolWt(m))
d['maxCharge'].append(Descriptors.MaxPartialCharge(m))
d['minCharge'].append(Descriptors.MinPartialCharge(m))
d['valence'].append(Descriptors.NumValenceElectrons(m))
d['TPSA'].append(rdMolDescriptors.CalcTPSA(m))
d['HBA'].append(rdMolDescriptors.CalcNumHBA(m))
d['HBD'].append(rdMolDescriptors.CalcNumHBD(m))
d['jIndex'].append(GraphDescriptors.BalabanJ(m))
df = pd.DataFrame.from_dict(d)
df_merge = pd.merge(df, DUD, on=df.index)
#df_merge.to_csv('/home/mcb/jboitr/data/DUD_full.csv')
df_merge.to_csv('C:/Users/jacqu/Documents/data/DUD_full.csv')
def calc_rdkit(mol):
descriptors = pd.Series(
np.array([
Crippen.MolLogP(mol),
Crippen.MolMR(mol),
Descriptors.FpDensityMorgan1(mol),
Descriptors.FpDensityMorgan2(mol),
Descriptors.FpDensityMorgan3(mol),
Descriptors.FractionCSP3(mol),
Descriptors.HeavyAtomMolWt(mol),
Descriptors.MaxAbsPartialCharge(mol),
Descriptors.MaxPartialCharge(mol),
Descriptors.MinAbsPartialCharge(mol),
Descriptors.MinPartialCharge(mol),
Descriptors.MolWt(mol),
Descriptors.NumRadicalElectrons(mol),
Descriptors.NumValenceElectrons(mol),
EState.EState.MaxAbsEStateIndex(mol),
EState.EState.MaxEStateIndex(mol),
EState.EState.MinAbsEStateIndex(mol),
EState.EState.MinEStateIndex(mol),
EState.EState_VSA.EState_VSA1(mol),
EState.EState_VSA.EState_VSA10(mol),
EState.EState_VSA.EState_VSA11(mol),
EState.EState_VSA.EState_VSA2(mol),
EState.EState_VSA.EState_VSA3(mol),
EState.EState_VSA.EState_VSA4(mol),
EState.EState_VSA.EState_VSA5(mol),
EState.EState_VSA.EState_VSA6(mol),
EState.EState_VSA.EState_VSA7(mol),
EState.EState_VSA.EState_VSA8(mol),
EState.EState_VSA.EState_VSA9(mol),
Fragments.fr_Al_COO(mol),
Fragments.fr_Al_OH(mol),
Fragments.fr_Al_OH_noTert(mol),
Fragments.fr_aldehyde(mol),
Fragments.fr_alkyl_carbamate(mol),
Fragments.fr_alkyl_halide(mol),
Fragments.fr_allylic_oxid(mol),
Fragments.fr_amide(mol),
Fragments.fr_amidine(mol),
Fragments.fr_aniline(mol),
Fragments.fr_Ar_COO(mol),
Fragments.fr_Ar_N(mol),
Fragments.fr_Ar_NH(mol),
Fragments.fr_Ar_OH(mol),
Fragments.fr_ArN(mol),
Fragments.fr_aryl_methyl(mol),
Fragments.fr_azide(mol),
Fragments.fr_azo(mol),
Fragments.fr_barbitur(mol),
Fragments.fr_benzene(mol),
Fragments.fr_benzodiazepine(mol),
Fragments.fr_bicyclic(mol),
Fragments.fr_C_O(mol),
Fragments.fr_C_O_noCOO(mol),
Fragments.fr_C_S(mol),
Fragments.fr_COO(mol),
Fragments.fr_COO2(mol),
Fragments.fr_diazo(mol),
Fragments.fr_dihydropyridine(mol),
Fragments.fr_epoxide(mol),
Fragments.fr_ester(mol),
Fragments.fr_ether(mol),
Fragments.fr_furan(mol),
Fragments.fr_guanido(mol),
Fragments.fr_halogen(mol),
Fragments.fr_hdrzine(mol),
Fragments.fr_hdrzone(mol),
Fragments.fr_HOCCN(mol),
Fragments.fr_imidazole(mol),
Fragments.fr_imide(mol),
Fragments.fr_Imine(mol),
Fragments.fr_isocyan(mol),
Fragments.fr_isothiocyan(mol),
Fragments.fr_ketone(mol),
Fragments.fr_ketone_Topliss(mol),
Fragments.fr_lactam(mol),
Fragments.fr_lactone(mol),
Fragments.fr_methoxy(mol),
Fragments.fr_morpholine(mol),
Fragments.fr_N_O(mol),
Fragments.fr_Ndealkylation1(mol),
Fragments.fr_Ndealkylation2(mol),
Fragments.fr_NH0(mol),
Fragments.fr_NH1(mol),
Fragments.fr_NH2(mol),
Fragments.fr_Nhpyrrole(mol),
Fragments.fr_nitrile(mol),
Fragments.fr_nitro(mol),
Fragments.fr_nitro_arom(mol),
Fragments.fr_nitro_arom_nonortho(mol),
Fragments.fr_nitroso(mol),
Fragments.fr_oxazole(mol),
Fragments.fr_oxime(mol),
Fragments.fr_para_hydroxylation(mol),
Fragments.fr_phenol(mol),
Fragments.fr_phenol_noOrthoHbond(mol),
Fragments.fr_phos_acid(mol),
Fragments.fr_phos_ester(mol),
Fragments.fr_piperdine(mol),
Fragments.fr_piperzine(mol),
Fragments.fr_priamide(mol),
Fragments.fr_prisulfonamd(mol),
Fragments.fr_pyridine(mol),
Fragments.fr_quatN(mol),
Fragments.fr_SH(mol),
Fragments.fr_sulfide(mol),
Fragments.fr_sulfonamd(mol),
Fragments.fr_sulfone(mol),
Fragments.fr_term_acetylene(mol),
Fragments.fr_tetrazole(mol),
Fragments.fr_thiazole(mol),
Fragments.fr_thiocyan(mol),
Fragments.fr_thiophene(mol),
Fragments.fr_unbrch_alkane(mol),
Fragments.fr_urea(mol),
GraphDescriptors.BalabanJ(mol),
GraphDescriptors.BertzCT(mol),
GraphDescriptors.Chi0(mol),
GraphDescriptors.Chi0n(mol),
GraphDescriptors.Chi0v(mol),
GraphDescriptors.Chi1(mol),
GraphDescriptors.Chi1n(mol),
GraphDescriptors.Chi1v(mol),
GraphDescriptors.Chi2n(mol),
GraphDescriptors.Chi2v(mol),
GraphDescriptors.Chi3n(mol),
GraphDescriptors.Chi3v(mol),
GraphDescriptors.Chi4n(mol),
GraphDescriptors.Chi4v(mol),
GraphDescriptors.HallKierAlpha(mol),
GraphDescriptors.Ipc(mol),
GraphDescriptors.Kappa1(mol),
GraphDescriptors.Kappa2(mol),
GraphDescriptors.Kappa3(mol),
Lipinski.HeavyAtomCount(mol),
Lipinski.NHOHCount(mol),
Lipinski.NOCount(mol),
Lipinski.NumAliphaticCarbocycles(mol),
Lipinski.NumAliphaticHeterocycles(mol),
Lipinski.NumAliphaticRings(mol),
Lipinski.NumAromaticCarbocycles(mol),
Lipinski.NumAromaticHeterocycles(mol),
Lipinski.NumAromaticRings(mol),
Lipinski.NumHAcceptors(mol),
Lipinski.NumHDonors(mol),
Lipinski.NumHeteroatoms(mol),
Lipinski.NumRotatableBonds(mol),
Lipinski.NumSaturatedCarbocycles(mol),
Lipinski.NumSaturatedHeterocycles(mol),
Lipinski.NumSaturatedRings(mol),
Lipinski.RingCount(mol),
MolSurf.LabuteASA(mol),
MolSurf.PEOE_VSA1(mol),
MolSurf.PEOE_VSA10(mol),
MolSurf.PEOE_VSA11(mol),
MolSurf.PEOE_VSA12(mol),
MolSurf.PEOE_VSA13(mol),
MolSurf.PEOE_VSA14(mol),
MolSurf.PEOE_VSA2(mol),
MolSurf.PEOE_VSA3(mol),
MolSurf.PEOE_VSA4(mol),
MolSurf.PEOE_VSA5(mol),
MolSurf.PEOE_VSA6(mol),
MolSurf.PEOE_VSA7(mol),
MolSurf.PEOE_VSA8(mol),
MolSurf.PEOE_VSA9(mol),
MolSurf.SlogP_VSA1(mol),
MolSurf.SlogP_VSA10(mol),
MolSurf.SlogP_VSA11(mol),
MolSurf.SlogP_VSA12(mol),
MolSurf.SlogP_VSA2(mol),
MolSurf.SlogP_VSA3(mol),
MolSurf.SlogP_VSA4(mol),
MolSurf.SlogP_VSA5(mol),
MolSurf.SlogP_VSA6(mol),
MolSurf.SlogP_VSA7(mol),
MolSurf.SlogP_VSA8(mol),
MolSurf.SlogP_VSA9(mol),
MolSurf.SMR_VSA1(mol),
MolSurf.SMR_VSA10(mol),
MolSurf.SMR_VSA2(mol),
MolSurf.SMR_VSA3(mol),
MolSurf.SMR_VSA4(mol),
MolSurf.SMR_VSA5(mol),
MolSurf.SMR_VSA6(mol),
MolSurf.SMR_VSA7(mol),
MolSurf.SMR_VSA8(mol),
MolSurf.SMR_VSA9(mol),
MolSurf.TPSA(mol)
]))
return descriptors
def calculate(self, D):
return float(RDKit.BalabanJ(self.mol, dMat=D))
def get_descriptors(smiles):
"""
Get a dictionary of RDKit descriptors from a SMILES string.
Parameters
----------
smiles : str
The SMILES string of the chemical of interest
Returns
-------
descriptors : dict
A collection of molecular descriptors
Notes: Developed with RDKit 2019.03.4, although doc pages listed 2019.03.1
"""
mol = Chem.MolFromSmiles(smiles)
mol = Chem.AddHs(mol)
Chem.EmbedMolecule(mol, Chem.ETKDG())
descriptors = {}
# Starting with simple descriptors:
# https://www.rdkit.org/docs/source/rdkit.Chem.Descriptors.html
# Molecular weight
descriptors['molwt'] = Descriptors.ExactMolWt(mol)
# Partial charge metrics
descriptors['max_abs_partial_charge'] = Descriptors.MaxAbsPartialCharge(mol)
descriptors['max_partial_charge'] = Descriptors.MaxPartialCharge(mol)
descriptors['min_abs_partial_charge'] = Descriptors.MinAbsPartialCharge(mol)
descriptors['min_partial_charge'] = Descriptors.MinPartialCharge(mol)
# Basic electron counts
descriptors['num_radical_electrons'] = Descriptors.NumRadicalElectrons(mol)
descriptors['num_valence_electrons'] = Descriptors.NumValenceElectrons(mol)
# 3-D descriptors
# https://www.rdkit.org/docs/source/rdkit.Chem.Descriptors3D.html
# Calculating these should produce the same result, according to some basic tests
# descriptors['asphericity'] = rdMolDescriptors.CalcAsphericity(mol)
# descriptors['eccentricity'] = rdMolDescriptors.CalcEccentricity(mol)
descriptors['asphericity'] = Descriptors3D.Asphericity(mol)
descriptors['eccentricity'] = Descriptors3D.Eccentricity(mol)
descriptors['inertial_shape_factor'] = Descriptors3D.InertialShapeFactor(mol)
descriptors['radius_of_gyration'] = Descriptors3D.RadiusOfGyration(mol)
descriptors['spherocity_index'] = Descriptors3D.SpherocityIndex(mol)
# Graph descriptors
# https://www.rdkit.org/docs/source/rdkit.Chem.GraphDescriptors.html
descriptors['balaban_j'] = GraphDescriptors.BalabanJ(mol)
descriptors['bertz_ct'] = GraphDescriptors.BertzCT(mol)
descriptors['chi0'] = GraphDescriptors.Chi0(mol)
descriptors['chi0n'] = GraphDescriptors.Chi0n(mol)
descriptors['chi0v'] = GraphDescriptors.Chi0v(mol)
descriptors['chi1'] = GraphDescriptors.Chi1(mol)
descriptors['chi1n'] = GraphDescriptors.Chi1n(mol)
descriptors['chi1v'] = GraphDescriptors.Chi1v(mol)
descriptors['chi2n'] = GraphDescriptors.Chi2n(mol)
descriptors['chi2v'] = GraphDescriptors.Chi2v(mol)
descriptors['chi3n'] = GraphDescriptors.Chi3n(mol)
descriptors['chi3v'] = GraphDescriptors.Chi3v(mol)
descriptors['chi4n'] = GraphDescriptors.Chi4n(mol)
descriptors['chi4v'] = GraphDescriptors.Chi4v(mol)
descriptors['hall_kier_alpha'] = GraphDescriptors.HallKierAlpha(mol)
descriptors['kappa1'] = GraphDescriptors.Kappa1(mol)
descriptors['kappa2'] = GraphDescriptors.Kappa2(mol)
descriptors['kappa3'] = GraphDescriptors.Kappa3(mol)
# Predicted properties from Wildman and Crippen
descriptors['log_p'] = Descriptors.MolLogP(mol)
descriptors['refractivity'] = Descriptors.MolMR(mol)
return descriptors