def test_lmer_basic_filtering():
groups = ['control', 'treatment', 'control', 'treatment']
repetitions = [1, 1, 2, 2]
features = [
[100, 99, 80, 120, 140, 150, 100, 100, 110, 111, 140, 145],
[115, 110, 90, 110, 145, 155, 110, 120, 115, 120, 120, 150,
100, 90, 100],
[150, 150, 130, 170, 190, 250, 150, 150, 160,
161, 180, 195, 130, 120, 125, 130, 125],
[155, 155, 135, 175, 195, 255, 155, 155, 165, 165,
185, 200, 135, 125, 130, 135, 140, 150, 135, 140]
]
datasets = []
for ff in features:
ds = dclab.new_dataset({"deform": ff})
ds.config["setup"]["chip region"] = "channel"
# apply some filters
datasets.append(ds)
rlme4 = Rlme4(model="lmer", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["anova p-value"], 0.8434625179432129)
# filters should have an effect
for ds in datasets:
ds.filter.manual[:4] = False
ds.apply_filter()
res2 = rlme4.fit()
assert not np.allclose(res["anova p-value"], res2["anova p-value"])
def test_lmer_basic_nan():
groups = ['control', 'treatment', 'control', 'treatment']
repetitions = [1, 1, 2, 2]
features = [
[100, np.inf, 80, np.nan, 140, 150, 100, 100, 110, 111, 140, 145],
[115, 110, 90, 110, 145, 155, 110, 120, 115, 120, 120, 150,
100, 90, 100],
[150, 150, 130, 170, 190, 250, 150, 150, 160,
161, 180, 195, 130, 120, 125, 130, 125],
[155, 155, 135, 175, 195, 255, 155, 155, 165, 165,
185, 200, 135, 125, 130, 135, 140, 150, 135, 140]
]
datasets = []
for ff in features:
ds = dclab.new_dataset({"deform": ff})
ds.config["setup"]["chip region"] = "channel"
assert ds.config["setup"]["chip region"] == "channel"
datasets.append(ds)
rlme4 = Rlme4(model="lmer", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["fixed effects intercept"], 137.37179302516199186)
def test_fail_add_same_dataset():
datasets = standard_datasets(set_region=False)
rlme4 = Rlme4(model="lmer", feature="deform")
rlme4.add_dataset(datasets[0], "control", 1)
with pytest.raises(ValueError, match="has already"):
rlme4.add_dataset(datasets[1], "control", 1)
def test_fail_too_few_dataset():
datasets = standard_datasets(set_region=False)
rlme4 = Rlme4(model="lmer", feature="deform")
rlme4.add_dataset(datasets[0], "control", 1)
rlme4.add_dataset(datasets[1], "treatment", 1)
with pytest.raises(ValueError, match="Linear mixed effects models"
+ " require repeated measurements"):
rlme4.fit()
def test_fail_get_non_existent_data():
datasets = standard_datasets(set_region=False)
rlme4 = Rlme4(model="lmer", feature="deform")
rlme4.add_dataset(datasets[0], "control", 1)
rlme4.add_dataset(datasets[1], "treatment", 1)
rlme4.add_dataset(datasets[2], "control", 2)
rlme4.add_dataset(datasets[3], "treatment", 2)
rlme4.add_dataset(datasets[3], "treatment", 3)
with pytest.raises(ValueError, match="Dataset for group 'control'"):
rlme4.get_feature_data(group="control", repetition=3)
def test_glmer_differential():
"""Differential Deformation in a generalized linear mixed model"""
groups = ['control', 'control', 'control', 'control', 'treatment',
'treatment', 'treatment', 'treatment']
repetitions = [1, 1, 2, 2, 1, 1, 2, 2]
datasets = standard_datasets()
rlme4 = Rlme4(model="glmer+loglink", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["fixed effects intercept"], 93.55789706339498)
assert rlme4.is_differential()
assert rlme4.model == "glmer+loglink"
assert np.allclose(res["anova p-value"], 0.000556063024310929)
assert res["model converged"]
def test_glmer_basic_larger():
"""Original values in a generalized linear mixed model"""
groups = ['treatment', 'control', 'treatment', 'control',
'treatment', 'control', 'treatment', 'control']
repetitions = [1, 1, 2, 2, 3, 3, 4, 4]
datasets = standard_datasets(set_region=False)
rlme4 = Rlme4(model="glmer+loglink", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["fixed effects intercept"], 15.083832084641697)
assert np.allclose(res["anova p-value"], 0.00365675950677214)
assert not rlme4.is_differential()
assert res["model converged"]
def test_lmer_differential():
"""Differential Deformation in a linear mixed model"""
groups = ['control', 'control', 'control', 'control', 'treatment',
'treatment', 'treatment', 'treatment']
repetitions = [1, 1, 2, 2, 1, 1, 2, 2]
datasets = standard_datasets()
rlme4 = Rlme4(model="lmer", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["fixed effects intercept"], 93.693750004463098)
assert rlme4.is_differential()
assert rlme4.model == "lmer"
assert np.allclose(res["anova p-value"], 0.000602622503360039)
assert np.allclose(res["fixed effects intercept"],
np.mean(res["fixed effects repetitions"], axis=1)[0])
assert np.allclose(res["fixed effects treatment"],
np.mean(res["fixed effects repetitions"], axis=1)[1])
assert res["model converged"]
def test_lmer_basic_larger():
"""Original values in a linear mixed model
'Reservoir' measurements are now Controls and 'Channel'
measurements are Treatments. This does not use differential
deformation.
"""
groups = ['treatment', 'control', 'treatment', 'control',
'treatment', 'control', 'treatment', 'control']
repetitions = [1, 1, 2, 2, 3, 3, 4, 4]
datasets = standard_datasets(set_region=False)
rlme4 = Rlme4(model="lmer", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["fixed effects intercept"], 17.171341507432501)
assert np.allclose(res["anova p-value"], 0.000331343267412872)
assert not rlme4.is_differential()
assert res["model converged"]
def test_lmer_basic():
groups = ['control', 'treatment', 'control', 'treatment']
repetitions = [1, 1, 2, 2]
features = [
[100, 99, 80, 120, 140, 150, 100, 100, 110, 111, 140, 145],
[115, 110, 90, 110, 145, 155, 110, 120, 115, 120, 120, 150,
100, 90, 100],
[150, 150, 130, 170, 190, 250, 150, 150, 160,
161, 180, 195, 130, 120, 125, 130, 125],
[155, 155, 135, 175, 195, 255, 155, 155, 165, 165,
185, 200, 135, 125, 130, 135, 140, 150, 135, 140]
]
datasets = []
for ff in features:
ds = dclab.new_dataset({"deform": ff})
ds.config["setup"]["chip region"] = "channel"
assert ds.config["setup"]["chip region"] == "channel"
datasets.append(ds)
rlme4 = Rlme4(model="lmer", feature="deform")
for ii in range(len(datasets)):
rlme4.add_dataset(datasets[ii], groups[ii], repetitions[ii])
res = rlme4.fit()
assert np.allclose(res["anova p-value"], 0.8434625179432129)
assert np.allclose(res["fixed effects intercept"], 136.6365047987075)
assert np.allclose(res["fixed effects treatment"], 1.4085644911191584,
atol=0, rtol=1e-3)
assert np.allclose(res["fixed effects intercept"],
np.mean(res["fixed effects repetitions"], axis=1)[0])
assert np.allclose(res["fixed effects treatment"],
np.mean(res["fixed effects repetitions"], axis=1)[1])
assert not res["is differential"]
assert res["feature"] == "deform"
assert res["model"] == "lmer"
assert res["model converged"]