def test_ensemble_model():
X = np.vstack(
(np.arange(30, 10,
-2, dtype='float64'), np.arange(100,
90,
-1,
dtype='float64'))).T
Y = np.arange(10, dtype='float64')
rf = regressors.randomforest(random_state=42)
nn = regressors.neuralnetwork(solver='lbfgs', random_state=42)
ensemble = ensemble_model((rf, nn))
# we do not need to fit underlying models, they change when we fit enseble
ensemble.fit(X, Y)
pred = ensemble.predict(X)
mean_pred = np.vstack((rf.predict(X), nn.predict(X))).mean(axis=0)
assert_array_almost_equal(pred, mean_pred)
assert_almost_equal(ensemble.score(X, Y), r2_score(Y, pred))
# ensemble of a single model should behave exactly like this model
nn = neuralnetwork(solver='lbfgs', random_state=42)
ensemble = ensemble_model((nn, ))
ensemble.fit(X, Y)
assert_array_almost_equal(ensemble.predict(X), nn.predict(X))
assert_almost_equal(ensemble.score(X, Y), nn.score(X, Y))
def __init__(self, protein=None, n_jobs=-1, version=1, spr=0, **kwargs):
self.protein = protein
self.n_jobs = n_jobs
self.version = version
self.spr = spr
model = randomforest(n_estimators=500,
oob_score=True,
n_jobs=n_jobs,
**kwargs)
if version == 1:
cutoff = 12
descriptors = close_contacts(protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 2:
cutoff = np.array([0, 2, 4, 6, 8, 10, 12])
descriptors = close_contacts(protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 3:
cutoff = 12
cc = close_contacts(protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
vina = autodock_vina_descriptor(protein)
descriptors = ensemble_descriptor((vina, cc))
super(rfscore, self).__init__(model,
descriptors,
score_title='rfscore_v%i' % self.version)
def test_regressors():
X = np.vstack((np.arange(30, 10, -2, dtype='float64'),
np.arange(100, 90, -1, dtype='float64'))).T
Y = np.arange(10, dtype='float64')
np.random.seed(42)
for regressor in (regressors.svm(C=10),
regressors.randomforest(random_state=42),
regressors.neuralnetwork(solver='lbfgs',
random_state=42,
hidden_layer_sizes=(20, 20)),
regressors.mlr()):
regressor.fit(X, Y)
pred = regressor.predict(X)
assert_true((np.abs(pred.flatten() - Y) < 1).all())
assert_greater(regressor.score(X, Y), 0.9)
pickled = pickle.dumps(regressor)
reloaded = pickle.loads(pickled)
pred_reloaded = reloaded.predict(X)
assert_array_almost_equal(pred, pred_reloaded)
def __init__(self, protein = None, n_jobs = -1, version = 1, spr = 0, **kwargs):
self.protein = protein
self.n_jobs = n_jobs
self.version = version
self.spr = spr
model = randomforest(n_estimators = 500, oob_score = True, n_jobs = n_jobs, **kwargs)
if version == 1:
cutoff = 12
descriptors = close_contacts(protein, cutoff = cutoff, protein_types = protein_atomic_nums, ligand_types = ligand_atomic_nums)
elif version == 2:
cutoff = np.array([ 0, 2, 4, 6, 8, 10, 12])
descriptors = close_contacts(protein, cutoff = cutoff, protein_types = protein_atomic_nums, ligand_types = ligand_atomic_nums)
elif version == 3:
cutoff = 12
cc = close_contacts(protein, cutoff = cutoff, protein_types = protein_atomic_nums, ligand_types = ligand_atomic_nums)
vina = autodock_vina_descriptor(protein)
descriptors = ensemble_descriptor((vina, cc))
super(rfscore,self).__init__(model, descriptors, score_title = 'rfscore')
def __init__(self, protein=None, n_jobs=-1, version=1, spr=0, **kwargs):
self.protein = protein
self.n_jobs = n_jobs
self.version = version
self.spr = spr
if version == 1:
cutoff = 12
mtry = 6
descriptors = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 2:
cutoff = np.array([0, 2, 4, 6, 8, 10, 12])
mtry = 14
descriptors = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 3:
cutoff = 12
mtry = 6
cc = close_contacts_descriptor(protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
vina_scores = [
'vina_gauss1', 'vina_gauss2', 'vina_repulsion',
'vina_hydrophobic', 'vina_hydrogen', 'vina_num_rotors'
]
vina = oddt_vina_descriptor(protein, vina_scores=vina_scores)
descriptors = ensemble_descriptor((vina, cc))
model = randomforest(n_estimators=500,
oob_score=True,
n_jobs=n_jobs,
max_features=mtry,
bootstrap=True,
min_samples_split=6,
**kwargs)
super(rfscore, self).__init__(model,
descriptors,
score_title='rfscore_v%i' % self.version)
def __init__(self, protein=None, n_jobs=-1, version=1, spr=0, **kwargs):
"""Scoring function implementing RF-Score variants. It predicts the
binding affinity (pKi/d) of ligand in a complex utilizng simple
descriptors (close contacts of atoms <12A) with sophisticated
machine-learning model (random forest). The third variand supplements
those contacts with Vina partial scores. For futher details see RF-Score
publications v1[1]_, v2[2]_, v3[3]_.
Parameters
----------
protein : oddt.toolkit.Molecule object
Receptor for the scored ligands
n_jobs: int (default=-1)
Number of cores to use for scoring and training. By default (-1)
all cores are allocated.
version: int (default=1)
Scoring function variant. The deault is the simplest one (v1).
spr: int (default=0)
The minimum number of contacts in each pair of atom types in
the training set for the column to be included in training.
This is a way of removal of not frequent and empty contacts.
References
----------
.. [1] Ballester PJ, Mitchell JBO. A machine learning approach to
predicting protein-ligand binding affinity with applications to
molecular docking. Bioinformatics. 2010;26: 1169-1175.
doi:10.1093/bioinformatics/btq112
.. [2] Ballester PJ, Schreyer A, Blundell TL. Does a more precise
chemical description of protein-ligand complexes lead to more
accurate prediction of binding affinity? J Chem Inf Model. 2014;54:
944-955. doi:10.1021/ci500091r
.. [3] Li H, Leung K-S, Wong M-H, Ballester PJ. Improving AutoDock Vina
Using Random Forest: The Growing Accuracy of Binding Affinity
Prediction by the Effective Exploitation of Larger Data Sets.
Mol Inform. WILEY-VCH Verlag; 2015;34: 115-126.
doi:10.1002/minf.201400132
"""
self.protein = protein
self.n_jobs = n_jobs
self.version = version
self.spr = spr
if version == 1:
cutoff = 12
mtry = 6
descriptors = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 2:
cutoff = np.array([0, 2, 4, 6, 8, 10, 12])
mtry = 14
descriptors = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 3:
cutoff = 12
mtry = 6
cc = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
vina_scores = ['vina_gauss1',
'vina_gauss2',
'vina_repulsion',
'vina_hydrophobic',
'vina_hydrogen',
'vina_num_rotors']
vina = oddt_vina_descriptor(protein, vina_scores=vina_scores)
descriptors = ensemble_descriptor((vina, cc))
model = randomforest(n_estimators=500,
oob_score=True,
n_jobs=n_jobs,
max_features=mtry,
bootstrap=True,
min_samples_split=6,
**kwargs)
super(rfscore, self).__init__(model, descriptors,
score_title='rfscore_v%i' % self.version)
def __init__(self, protein = None, n_jobs = -1, **kwargs):
self.protein = protein
self.n_jobs = n_jobs
model = randomforest(n_estimators = 500, oob_score = True, n_jobs = n_jobs, **kwargs)
descriptors = close_contacts(protein, cutoff = cutoff, protein_types = protein_atomic_nums, ligand_types = ligand_atomic_nums)
super(rfscore,self).__init__(model, descriptors, score_title = 'rfscore')
assert cls.score(X, Y) == 1.0
prob = cls.predict_proba(X)
assert_array_almost_equal(prob, [[0, 1]] * 5 + [[1, 0]] * 5, decimal=1)
log_prob = cls.predict_log_proba(X)
assert_array_almost_equal(np.log(prob), log_prob)
pickled = pickle.dumps(cls)
reloaded = pickle.loads(pickled)
prob_reloaded = reloaded.predict_proba(X)
assert_array_almost_equal(prob, prob_reloaded)
@pytest.mark.parametrize('reg', [
regressors.svm(C=10),
regressors.randomforest(random_state=42),
regressors.neuralnetwork(
solver='lbfgs', random_state=42, hidden_layer_sizes=(20, 20)),
regressors.mlr()
])
def test_regressors(reg):
X = np.vstack(
(np.arange(30, 10,
-2, dtype='float64'), np.arange(100,
90,
-1,
dtype='float64'))).T
Y = np.arange(10, dtype='float64')
np.random.seed(42)
def __init__(self, protein=None, n_jobs=-1, version=1, spr=0, **kwargs):
"""Scoring function implementing RF-Score variants. It predicts the
binding affinity (pKi/d) of ligand in a complex utilizng simple
descriptors (close contacts of atoms <12A) with sophisticated
machine-learning model (random forest). The third variand supplements
those contacts with Vina partial scores. For futher details see RF-Score
publications v1[1]_, v2[2]_, v3[3]_.
Parameters
----------
protein : oddt.toolkit.Molecule object
Receptor for the scored ligands
n_jobs: int (default=-1)
Number of cores to use for scoring and training. By default (-1)
all cores are allocated.
version: int (default=1)
Scoring function variant. The deault is the simplest one (v1).
spr: int (default=0)
The minimum number of contacts in each pair of atom types in
the training set for the column to be included in training.
This is a way of removal of not frequent and empty contacts.
References
----------
.. [1] Ballester PJ, Mitchell JBO. A machine learning approach to
predicting protein-ligand binding affinity with applications to
molecular docking. Bioinformatics. 2010;26: 1169-1175.
doi:10.1093/bioinformatics/btq112
.. [2] Ballester PJ, Schreyer A, Blundell TL. Does a more precise
chemical description of protein-ligand complexes lead to more
accurate prediction of binding affinity? J Chem Inf Model. 2014;54:
944-955. doi:10.1021/ci500091r
.. [3] Li H, Leung K-S, Wong M-H, Ballester PJ. Improving AutoDock Vina
Using Random Forest: The Growing Accuracy of Binding Affinity
Prediction by the Effective Exploitation of Larger Data Sets.
Mol Inform. WILEY-VCH Verlag; 2015;34: 115-126.
doi:10.1002/minf.201400132
"""
self.protein = protein
self.n_jobs = n_jobs
self.version = version
self.spr = spr
if version == 1:
cutoff = 12
mtry = 6
descriptors = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 2:
cutoff = np.array([0, 2, 4, 6, 8, 10, 12])
mtry = 14
descriptors = close_contacts_descriptor(
protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
elif version == 3:
cutoff = 12
mtry = 6
cc = close_contacts_descriptor(protein,
cutoff=cutoff,
protein_types=protein_atomic_nums,
ligand_types=ligand_atomic_nums)
vina_scores = [
'vina_gauss1', 'vina_gauss2', 'vina_repulsion',
'vina_hydrophobic', 'vina_hydrogen', 'vina_num_rotors'
]
vina = oddt_vina_descriptor(protein, vina_scores=vina_scores)
descriptors = ensemble_descriptor((vina, cc))
# elif version == 5:
# cutoff = np.array([0, 2, 4, 6, 8, 10, 12])
# mtry = 14
# descriptors = close_contacts_descriptor(
# protein,
# cutoff=cutoff,
# protein_types=protein_atomic_nums,
# ligand_types=ligand_atomic_nums)
model = randomforest(n_estimators=500,
oob_score=True,
n_jobs=n_jobs,
max_features=mtry,
bootstrap=True,
min_samples_split=6,
**kwargs)
super(rfscore, self).__init__(model,
descriptors,
score_title='rfscore_v%i' % self.version)