def testIssue80(self):
from rdkit.Chem import Lipinski
m = Chem.MolFromSmiles('CCOC')
ref = Crippen.MolLogP(m)
Lipinski.NHOHCount(m)
probe = Crippen.MolLogP(m)
self.assertTrue(probe == ref)
def lipinski_rule(mol):
fingerprint = rdMolDescriptors.GetHashedAtomPairFingerprintAsBitVect(mol)
return [
Lipinski.NHOHCount(mol) <= 5,
Lipinski.NOCount(mol) <= 10,
Descriptors.ExactMolWt(mol) <= 500,
LogP('logP').run(fingerprint) <= 5]
def get_descriptors(mol, write=False):
# Make a copy of the molecule dataframe
desc = [
Lipinski.NumAromaticHeterocycles(mol),
Lipinski.NumAromaticRings(mol),
Lipinski.NumHDonors(mol),
Lipinski.RingCount(mol),
Lipinski.NHOHCount(mol),
Lipinski.NumHeteroatoms(mol),
Lipinski.NumAliphaticCarbocycles(mol),
Lipinski.NumSaturatedCarbocycles(mol),
Lipinski.NumAliphaticHeterocycles(mol),
Lipinski.NumHAcceptors(mol),
Lipinski.NumSaturatedHeterocycles(mol),
Lipinski.NumAliphaticRings(mol),
Descriptors.NumRadicalElectrons(mol),
Descriptors.MaxPartialCharge(mol),
Descriptors.NumValenceElectrons(mol),
Lipinski.FractionCSP3(mol),
Descriptors.MaxAbsPartialCharge(mol),
Lipinski.NumAromaticCarbocycles(mol),
Lipinski.NumSaturatedRings(mol),
Lipinski.NumRotatableBonds(mol)
]
desc = [0 if i != i else i for i in desc]
return desc
def test1(self):
" testing first 200 mols from NCI "
# figure out which rotor version we are using
m = Chem.MolFromSmiles("CC(C)(C)c1cc(O)c(cc1O)C(C)(C)C")
if Lipinski.NumRotatableBonds(m) == 2:
rot_prop = NonStrict
else:
rot_prop = Strict
suppl = Chem.SDMolSupplier(self.inFileName)
idx = 1
for m in suppl:
if m:
calc = Lipinski.NHOHCount(m)
orig = int(m.GetProp('NUM_LIPINSKIHDONORS'))
assert calc == orig, 'bad num h donors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NOCount(m)
orig = int(m.GetProp('NUM_LIPINSKIHACCEPTORS'))
assert calc == orig, 'bad num h acceptors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumHDonors(m)
orig = int(m.GetProp('NUM_HDONORS'))
assert calc == orig, 'bad num h donors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumHAcceptors(m)
orig = int(m.GetProp('NUM_HACCEPTORS'))
assert calc == orig, 'bad num h acceptors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumHeteroatoms(m)
orig = int(m.GetProp('NUM_HETEROATOMS'))
assert calc == orig, 'bad num heteroatoms for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumRotatableBonds(m)
orig = int(m.GetProp(rot_prop))
assert calc == orig, 'bad num rotors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
# test the underlying numrotatable bonds
calc = rdMolDescriptors.CalcNumRotatableBonds(
m, rdMolDescriptors.NumRotatableBondsOptions.NonStrict)
orig = int(m.GetProp(NonStrict))
assert calc == orig, 'bad num rotors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = rdMolDescriptors.CalcNumRotatableBonds(
m, rdMolDescriptors.NumRotatableBondsOptions.Strict)
orig = int(m.GetProp(Strict))
assert calc == orig, 'bad num rotors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
idx += 1
def auto_sampling(mult_factor, mol, log):
auto_samples = 0
auto_samples += 3 * (Lipinski.NumRotatableBonds(mol)
) # x3, for C3 rotations
auto_samples += 3 * (Lipinski.NHOHCount(mol)) # x3, for OH/NH rotations
auto_samples += 3 * (Lipinski.NumSaturatedRings(mol)
) # x3, for boat/chair/envelope confs
if auto_samples == 0:
auto_samples = mult_factor
else:
auto_samples = mult_factor * auto_samples
return auto_samples
def auto_sampling(mult_factor,mol,args,log): if args.metal_complex: if len(args.metal_idx) > 0: mult_factor = mult_factor*3*len(args.metal_idx) # this accounts for possible trans/cis isomers in metal complexes auto_samples = 0 auto_samples += 3*(Lipinski.NumRotatableBonds(mol)) # x3, for C3 rotations auto_samples += 3*(Lipinski.NHOHCount(mol)) # x3, for OH/NH rotations auto_samples += 3*(Lipinski.NumSaturatedRings(mol)) # x3, for boat/chair/envelope confs if auto_samples == 0: auto_samples = mult_factor else: auto_samples = mult_factor*auto_samples return auto_samples
def descriptors(self, mol):
aromatic_frac = self.arofrac(mol)
mw = Descriptors.ExactMolWt(mol, False)
valence_e = Descriptors.NumValenceElectrons(mol)
h_acceptors = Lipinski.NumHAcceptors(mol)
h_donors = Lipinski.NumHDonors(mol)
NO_counts = Lipinski.NOCount(mol)
NHOH_count = Lipinski.NHOHCount(mol)
rotors = Lipinski.NumRotatableBonds(mol)
SP3_frac = Lipinski.FractionCSP3(mol)
logP = Crippen.MolLogP(mol)
SP_bonds = len(mol.GetSubstructMatches(Chem.MolFromSmarts('[^1]')))
return([aromatic_frac,mw,valence_e,h_acceptors,h_donors,NO_counts,NHOH_count, rotors,SP3_frac,logP,SP_bonds])
def CalculateNumLipinskiHDonors(mol):
"""
Caculation of the number of Hydrogen Bond Donors
(nitrogen–hydrogen and oxygen–hydrogen bonds)
--->nLipinskiHD
:param mol: molecular
:type mol: rdkit.Chem.rdchem.Mol
:return: the number of Hydrogen Bond Donors
:rtype: int
"""
nLipinskiHD = Lipinski.NHOHCount(mol)
return nLipinskiHD
def test1(self):
" testing first 200 mols from NCI "
suppl = Chem.SDMolSupplier(self.inFileName)
idx = 1
oldDonorSmarts = Chem.MolFromSmarts('[NH1,NH2,OH1]')
OldDonorCount = lambda x, y=oldDonorSmarts: Lipinski._NumMatches(x, y)
oldAcceptorSmarts = Chem.MolFromSmarts('[N,O]')
OldAcceptorCount = lambda x, y=oldAcceptorSmarts: Lipinski._NumMatches(
x, y)
for m in suppl:
if m:
calc = Lipinski.NHOHCount(m)
orig = int(m.GetProp('NUM_LIPINSKIHDONORS'))
assert calc == orig, 'bad num h donors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NOCount(m)
orig = int(m.GetProp('NUM_LIPINSKIHACCEPTORS'))
assert calc == orig, 'bad num h acceptors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumHDonors(m)
orig = int(m.GetProp('NUM_HDONORS'))
assert calc == orig, 'bad num h donors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumHAcceptors(m)
orig = int(m.GetProp('NUM_HACCEPTORS'))
assert calc == orig, 'bad num h acceptors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumHeteroatoms(m)
orig = int(m.GetProp('NUM_HETEROATOMS'))
assert calc == orig, 'bad num heteroatoms for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
calc = Lipinski.NumRotatableBonds(m)
orig = int(m.GetProp('NUM_ROTATABLEBONDS'))
assert calc == orig, 'bad num rotors for mol %d (%s): %d != %d' % (
idx, m.GetProp('SMILES'), calc, orig)
idx += 1
def extract(x, from_smiles):
if from_smiles:
mol = Chem.MolFromSmiles(x)
else:
mol = x
if (mol is None) or (len(mol.GetAtoms()) == 0):
if include_3D:
return [0] * 29
else:
return [0] * 24
else:
logP = Crippen.MolLogP(mol)
refractivity = Crippen.MolMR(mol)
weight = Descriptors.MolWt(mol)
exact_weight = Descriptors.ExactMolWt(mol)
heavy_weight = Descriptors.HeavyAtomMolWt(mol)
heavy_count = Lipinski.HeavyAtomCount(mol)
nhoh_count = Lipinski.NHOHCount(mol)
no_count = Lipinski.NOCount(mol)
hacceptor_count = Lipinski.NumHAcceptors(mol)
hdonor_count = Lipinski.NumHDonors(mol)
hetero_count = Lipinski.NumHeteroatoms(mol)
rotatable_bond_count = Lipinski.NumRotatableBonds(mol)
valance_electron_count = Descriptors.NumValenceElectrons(mol)
amide_bond_count = rdMolDescriptors.CalcNumAmideBonds(mol)
aliphatic_ring_count = Lipinski.NumAliphaticRings(mol)
aromatic_ring_count = Lipinski.NumAromaticRings(mol)
saturated_ring_count = Lipinski.NumSaturatedRings(mol)
aliphatic_cycle_count = Lipinski.NumAliphaticCarbocycles(mol)
aliphaticHetero_cycle_count = Lipinski.NumAliphaticHeterocycles(
mol)
aromatic_cycle_count = Lipinski.NumAromaticCarbocycles(mol)
aromaticHetero_cycle_count = Lipinski.NumAromaticHeterocycles(mol)
saturated_cycle_count = Lipinski.NumSaturatedCarbocycles(mol)
saturatedHetero_cycle_count = Lipinski.NumSaturatedHeterocycles(
mol)
tpsa = rdMolDescriptors.CalcTPSA(mol)
if include_3D:
mol_3D = Chem.AddHs(mol)
AllChem.EmbedMolecule(mol_3D)
AllChem.MMFFOptimizeMolecule(mol_3D)
eccentricity = rdMolDescriptors.CalcEccentricity(mol_3D)
asphericity = rdMolDescriptors.CalcAsphericity(mol_3D)
spherocity = rdMolDescriptors.CalcSpherocityIndex(mol_3D)
inertial = rdMolDescriptors.CalcInertialShapeFactor(mol_3D)
gyration = rdMolDescriptors.CalcRadiusOfGyration(mol_3D)
return [
logP, refractivity, weight, exact_weight, heavy_weight,
heavy_count, nhoh_count, no_count, hacceptor_count,
hdonor_count, hetero_count, rotatable_bond_count,
valance_electron_count, amide_bond_count,
aliphatic_ring_count, aromatic_ring_count,
saturated_ring_count, aliphatic_cycle_count,
aliphaticHetero_cycle_count, aromatic_cycle_count,
aromaticHetero_cycle_count, saturated_cycle_count,
saturatedHetero_cycle_count, tpsa, eccentricity,
asphericity, spherocity, inertial, gyration
]
else:
return [
logP, refractivity, weight, exact_weight, heavy_weight,
heavy_count, nhoh_count, no_count, hacceptor_count,
hdonor_count, hetero_count, rotatable_bond_count,
valance_electron_count, amide_bond_count,
aliphatic_ring_count, aromatic_ring_count,
saturated_ring_count, aliphatic_cycle_count,
aliphaticHetero_cycle_count, aromatic_cycle_count,
aromaticHetero_cycle_count, saturated_cycle_count,
saturatedHetero_cycle_count, tpsa
]
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
# Sets up a list to hold the index of compounds that violate two of Lipinski's
# rules of five. The counter is to track that index number.
lipinski_violators = []
counter = 0
print("Scanning molecules for Lipinski violations.")
for mol in tqdm(mols):
# Assume no violations
dono_viol = False
acceptor_viol = False
mw_viol = False
logp_viol = False
# Use RDKit functions to get hdonors, acceptors, molecular weight and
# logP.
hdonors = Lipinski.NHOHCount(mol)
hacceptors = Lipinski.NOCount(mol)
mw = rdMolDescriptors.CalcExactMolWt(mol)
logp = Crippen.MolLogP(mol)
# Make the checks if the current mol actually violates a role.
if hdonors > 5:
dono_viol = True
if hacceptors > 10:
acceptor_viol = True
if mw > 500:
mw_viol = True
if logp > 5:
logp_viol = True
# Check if the violation sum is greater than one and assign the molecule
for line in data.readlines():
split = line.split(' ')
mol = Chem.MolFromSmiles(split[0])
fingerprint = rdMolDescriptors.GetHashedAtomPairFingerprintAsBitVect(mol)
pKa = acid_model.run(GetAvalonFP(mol) +
MACCSkeys.GenMACCSKeys(mol) +
fingerprint)
sim_pKa = sim_model.run(split[0], acids)
X.append([pKa,
sim_pKa,
Lipinski.NumHDonors(mol),
Lipinski.NumHAcceptors(mol),
Lipinski.NHOHCount(mol)])
Y.append(float(split[1][:-1]))
scaler = preprocessing.StandardScaler()
X = scaler.fit_transform(np.asarray(X))
Y = np.asarray(Y)
X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.1,
random_state=1)
model = MLPRegressor(solver='adam', alpha=1e-5, hidden_layer_sizes=(1048, 128),
random_state=1, verbose=1, max_iter=1000, batch_size=500)
model.fit(X_train, y_train)
print(model.score(X_test, y_test))