def hard_osimertinib() -> GoalDirectedBenchmark:
smiles = 'COc1cc(N(C)CCN(C)C)c(NC(=O)C=C)cc1Nc2nccc(n2)c3cn(C)c4ccccc34'
modifier = ClippedScoreModifier(upper_x=0.8)
similar_to_osimertinib = TanimotoScoringFunction(smiles,
fp_type='FCFP4',
score_modifier=modifier)
but_not_too_similar = TanimotoScoringFunction(
smiles,
fp_type='ECFP6',
score_modifier=MinGaussianModifier(mu=0.85, sigma=0.1))
tpsa_over_100 = RdkitScoringFunction(descriptor=tpsa,
score_modifier=MaxGaussianModifier(
mu=100, sigma=10))
logP_scoring = RdkitScoringFunction(descriptor=logP,
score_modifier=MinGaussianModifier(
mu=1, sigma=1))
make_osimertinib_great_again = ArithmeticMeanScoringFunction([
similar_to_osimertinib, but_not_too_similar, tpsa_over_100,
logP_scoring
])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name='Osimertinib MPO',
objective=make_osimertinib_great_again,
contribution_specification=specification)
def start_pop_ranolazine() -> GoalDirectedBenchmark:
ranolazine = 'COc1ccccc1OCC(O)CN2CCN(CC(=O)Nc3c(C)cccc3C)CC2'
modifier = ClippedScoreModifier(upper_x=0.7)
similar_to_ranolazine = TanimotoScoringFunction(ranolazine,
fp_type='AP',
score_modifier=modifier)
logP_under_4 = RdkitScoringFunction(descriptor=logP,
score_modifier=MaxGaussianModifier(
mu=7, sigma=1))
aroma = RdkitScoringFunction(descriptor=num_aromatic_rings,
score_modifier=MinGaussianModifier(mu=1,
sigma=1))
fluorine = RdkitScoringFunction(descriptor=AtomCounter('F'),
score_modifier=GaussianModifier(mu=1,
sigma=1.0))
optimize_ranolazine = ArithmeticMeanScoringFunction(
[similar_to_ranolazine, logP_under_4, fluorine, aroma])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name='Ranolazine MPO',
objective=optimize_ranolazine,
contribution_specification=specification,
starting_population=[ranolazine])
def hard_cobimetinib() -> GoalDirectedBenchmark:
smiles = 'OC1(CN(C1)C(=O)C1=C(NC2=C(F)C=C(I)C=C2)C(F)=C(F)C=C1)C1CCCCN1'
modifier = ClippedScoreModifier(upper_x=0.7)
os_tf = TanimotoScoringFunction(smiles,
fp_type='FCFP4',
score_modifier=modifier)
os_ap = TanimotoScoringFunction(smiles,
fp_type='ECFP6',
score_modifier=MinGaussianModifier(
mu=0.75, sigma=0.1))
rot_b = RdkitScoringFunction(descriptor=num_rotatable_bonds,
score_modifier=MinGaussianModifier(mu=3,
sigma=1))
rings = RdkitScoringFunction(descriptor=num_aromatic_rings,
score_modifier=MaxGaussianModifier(mu=3,
sigma=1))
t_cns = ArithmeticMeanScoringFunction(
[os_tf, os_ap, rot_b, rings,
CNS_MPO_ScoringFunction()])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name='Cobimetinib MPO',
objective=t_cns,
contribution_specification=specification)
def hard_fexofenadine() -> GoalDirectedBenchmark:
"""
make fexofenadine less greasy
:return:
"""
smiles = 'CC(C)(C(=O)O)c1ccc(cc1)C(O)CCCN2CCC(CC2)C(O)(c3ccccc3)c4ccccc4'
modifier = ClippedScoreModifier(upper_x=0.8)
similar_to_fexofenadine = TanimotoScoringFunction(smiles,
fp_type='AP',
score_modifier=modifier)
tpsa_over_90 = RdkitScoringFunction(descriptor=tpsa,
score_modifier=MaxGaussianModifier(
mu=90, sigma=10))
logP_under_4 = RdkitScoringFunction(descriptor=logP,
score_modifier=MinGaussianModifier(
mu=4, sigma=1))
optimize_fexofenadine = ArithmeticMeanScoringFunction(
[similar_to_fexofenadine, tpsa_over_90, logP_under_4])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name='Fexofenadine MPO',
objective=optimize_fexofenadine,
contribution_specification=specification)
def ranolazine_mpo() -> GoalDirectedBenchmark:
"""
Make start_pop_ranolazine more polar and add a fluorine
"""
ranolazine = "COc1ccccc1OCC(O)CN2CCN(CC(=O)Nc3c(C)cccc3C)CC2"
modifier = ClippedScoreModifier(upper_x=0.7)
similar_to_ranolazine = TanimotoScoringFunction(ranolazine,
fp_type="AP",
score_modifier=modifier)
logP_under_4 = RdkitScoringFunction(descriptor=logP,
score_modifier=MaxGaussianModifier(
mu=7, sigma=1))
tpsa_f = RdkitScoringFunction(descriptor=tpsa,
score_modifier=MaxGaussianModifier(mu=95,
sigma=20))
fluorine = RdkitScoringFunction(descriptor=AtomCounter("F"),
score_modifier=GaussianModifier(mu=1,
sigma=1.0))
optimize_ranolazine = GeometricMeanScoringFunction(
[similar_to_ranolazine, logP_under_4, fluorine, tpsa_f])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(
name="Ranolazine MPO",
objective=optimize_ranolazine,
contribution_specification=specification,
starting_population=[ranolazine],
)
def scaffold_hop() -> GoalDirectedBenchmark:
"""
Keep the decoration, and similarity to start point, but change the scaffold.
"""
smiles = "CCCOc1cc2ncnc(Nc3ccc4ncsc4c3)c2cc1S(=O)(=O)C(C)(C)C"
pharmacophor_sim = TanimotoScoringFunction(
smiles,
fp_type="PHCO",
score_modifier=ClippedScoreModifier(upper_x=0.75))
deco = SMARTSScoringFunction(
"[#6]-[#6]-[#6]-[#8]-[#6]~[#6]~[#6]~[#6]~[#6]-[#7]-c1ccc2ncsc2c1",
inverse=False)
# anti scaffold
scaffold = SMARTSScoringFunction(
"[#7]-c1n[c;h1]nc2[c;h1]c(-[#8])[c;h0][c;h1]c12", inverse=True)
scaffold_hop_obj = ArithmeticMeanScoringFunction(
[pharmacophor_sim, deco, scaffold])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name="Scaffold Hop",
objective=scaffold_hop_obj,
contribution_specification=specification)
def test_thresholded_is_special_case_of_clipped_for_positive_input():
threshold = 4.584
thresholded_modifier = ThresholdedLinearModifier(threshold=threshold)
clipped_modifier = ClippedScoreModifier(upper_x=threshold)
values = np.array([0, 2.3, 8.545, 3.23, 0.12, 55.555])
assert np.allclose(thresholded_modifier(values), clipped_modifier(values))
def similarity(smiles: str, name: str, fp_type: str = 'ECFP4', threshold: float = 0.7,
rediscovery: bool = False) -> GoalDirectedBenchmark:
category = 'rediscovery' if rediscovery else 'similarity'
benchmark_name = f'{name} {category}'
modifier = ClippedScoreModifier(upper_x=threshold)
scoring_function = TanimotoScoringFunction(target=smiles, fp_type=fp_type, score_modifier=modifier)
if rediscovery:
specification = uniform_specification(1)
else:
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name=benchmark_name,
objective=scoring_function,
contribution_specification=specification)
def decoration_hop() -> GoalDirectedBenchmark:
smiles = 'CCCOc1cc2ncnc(Nc3ccc4ncsc4c3)c2cc1S(=O)(=O)C(C)(C)C'
pharmacophor_sim = TanimotoScoringFunction(smiles, fp_type='PHCO',
score_modifier=ClippedScoreModifier(upper_x=0.85))
# change deco
deco1 = SMARTSScoringFunction('CS([#6])(=O)=O', inverse=True)
deco2 = SMARTSScoringFunction('[#7]-c1ccc2ncsc2c1', inverse=True)
# keep scaffold
scaffold = SMARTSScoringFunction('[#7]-c1n[c;h1]nc2[c;h1]c(-[#8])[c;h0][c;h1]c12', inverse=False)
deco_hop1_fn = ArithmeticMeanScoringFunction([pharmacophor_sim, deco1, deco2, scaffold])
specification = uniform_specification(1, 10, 100)
return GoalDirectedBenchmark(name='Deco Hop',
objective=deco_hop1_fn,
contribution_specification=specification)
def test_clipped_function():
min_x = 4.4
max_x = 8.8
min_score = -3.3
max_score = 9.2
modifier = ClippedScoreModifier(upper_x=max_x, lower_x=min_x, high_score=max_score, low_score=min_score)
# values smaller than min_x should be assigned min_score
for x in [-2, 0, 4, 4.4]:
assert modifier(x) == min_score
# values larger than max_x should be assigned min_score
for x in [8.8, 9.0, 1000]:
assert modifier(x) == max_score
# values in between are interpolated
slope = (max_score - min_score) / (max_x - min_x)
for x in [4.4, 4.8, 5.353, 8.034, 8.8]:
dx = x - min_x
dy = dx * slope
assert modifier(x) == pytest.approx(min_score + dy)