def test_compute_charge_dictionary(self):
from rdkit.Chem.AllChem import ComputeGasteigerCharges
for fname in (self.ligand_file, self.protein_file):
_, mol = rgf.load_molecule(fname)
ComputeGasteigerCharges(mol)
charge_dict = rgf.compute_charge_dictionary(mol)
self.assertEqual(len(charge_dict), mol.GetNumAtoms())
for i in range(mol.GetNumAtoms()):
self.assertIn(i, charge_dict)
self.assertIsInstance(charge_dict[i], (float, int))
def test_load_molecule(self):
# adding hydrogens and charges is tested in dc.utils
from rdkit.Chem.AllChem import Mol
for add_hydrogens in (True, False):
for calc_charges in (True, False):
mol_xyz, mol_rdk = rgf.load_molecule(self.ligand_file, add_hydrogens,
calc_charges)
num_atoms = mol_rdk.GetNumAtoms()
self.assertIsInstance(mol_xyz, np.ndarray)
self.assertIsInstance(mol_rdk, Mol)
self.assertEqual(mol_xyz.shape, (num_atoms, 3))
def setUp(self):
current_dir = os.path.dirname(os.path.realpath(__file__))
# simple flat ring
from rdkit.Chem import MolFromSmiles
self.cycle4 = MolFromSmiles('C1CCC1')
self.cycle4.Compute2DCoords()
# load and sanitize two real molecules
_, self.prot = rgf.load_molecule(
os.path.join(current_dir, 'data', '3ws9_protein_fixer_rdkit.pdb'),
add_hydrogens=False,
calc_charges=False,
sanitize=True)
_, self.lig = rgf.load_molecule(
os.path.join(current_dir, 'data', '3ws9_ligand.sdf'),
add_hydrogens=False,
calc_charges=False,
sanitize=True)
def test_featurize_splif(self):
prot_xyz, prot_rdk = rgf.load_molecule(self.protein_file)
lig_xyz, lig_rdk = rgf.load_molecule(self.ligand_file)
distance = rgf.compute_pairwise_distances(
protein_xyz=prot_xyz, ligand_xyz=lig_xyz)
bins = [(1, 2), (2, 3)]
dicts = rgf.featurize_splif(
prot_xyz,
prot_rdk,
lig_xyz,
lig_rdk,
contact_bins=bins,
pairwise_distances=distance,
ecfp_degree=2)
expected_dicts = [
rgf.compute_splif_features_in_range(
prot_rdk, lig_rdk, distance, c_bin, ecfp_degree=2) for c_bin in bins
]
self.assertIsInstance(dicts, list)
self.assertEqual(dicts, expected_dicts)
def test_compute_all_ecfp(self):
_, mol = rgf.load_molecule(self.ligand_file)
num_atoms = mol.GetNumAtoms()
for degree in range(1, 4):
# TODO test if dict contains smiles
ecfp_all = rgf.compute_all_ecfp(mol, degree=degree)
self.assertIsInstance(ecfp_all, dict)
self.assertEqual(len(ecfp_all), num_atoms)
self.assertEqual(list(ecfp_all.keys()), list(range(num_atoms)))
num_ind = np.random.choice(range(1, num_atoms))
indices = list(np.random.choice(num_atoms, num_ind, replace=False))
ecfp_selected = rgf.compute_all_ecfp(mol, indices=indices, degree=degree)
self.assertIsInstance(ecfp_selected, dict)
self.assertEqual(len(ecfp_selected), num_ind)
self.assertEqual(sorted(ecfp_selected.keys()), sorted(indices))