def compositional_effect_size(max_alpha, reps, intervals, n_species, n_diff,
n_contaminants, lam, library_size, asymmetry,
fold_balance, template_biom,
template_sample_name, output_dir):
if template_biom is not None:
templ = load_table(template_biom)
template = templ.data(id=template_sample_name, axis='sample')
else:
template = None
os.mkdir(output_dir)
gen = compositional_effect_size_generator(max_alpha,
reps,
intervals,
n_species,
n_diff,
n_contaminants,
lam,
library_size=library_size,
asymmetry=asymmetry,
fold_balance=fold_balance,
template=template)
for i, g in enumerate(gen):
table, groups, truth = g
output_table = "%s/table.%d.biom" % (output_dir, i)
output_groups = "%s/metadata.%d.txt" % (output_dir, i)
output_truth = "%s/truth.%d.csv" % (output_dir, i)
deposit(table, groups, truth, output_table, output_groups,
output_truth)
def test_composition_effect_size_balanced(self):
gen = compositional_effect_size_generator(max_alpha=1, reps=5,
intervals=2, n_species=5, n_diff=1,
n_contaminants=2, lam=0.1,
balanced=False)
table, metadata, truth = next(gen)
table, metadata, truth = next(gen)
exp_table = pd.DataFrame({
'S0': [0.100000, 0.100000, 0.100000, 0.100000, 0.100000,
0.499977, 0.0000226989],
'S1': [0.100000, 0.100000, 0.100000, 0.100000, 0.100000,
0.499977, 0.0000226989],
'S2': [0.100000, 0.100000, 0.100000, 0.100000, 0.100000,
0.499977, 0.0000226989],
'S3': [0.100000, 0.100000, 0.100000, 0.100000, 0.100000,
0.499977, 0.0000226989],
'S4': [0.100000, 0.100000, 0.100000, 0.100000, 0.100000,
0.499977, 0.0000226989],
'S5': [0.003817, 0.038168, 0.038168, 0.038168, 0.381679,
0.499977, 0.0000226989],
'S6': [0.003817, 0.038168, 0.038168, 0.038168, 0.381679,
0.499977, 0.0000226989],
'S7': [0.003817, 0.038168, 0.038168, 0.038168, 0.381679,
0.499977, 0.0000226989],
'S8': [0.003817, 0.038168, 0.038168, 0.038168, 0.381679,
0.499977, 0.0000226989],
'S9': [0.003817, 0.038168, 0.038168, 0.038168, 0.381679,
0.499977, 0.0000226989]},
index=['F0', 'F1', 'F2', 'F3', 'F4', 'X0', 'X1']
).T
pdt.assert_frame_equal(table, exp_table, check_less_precise=True)
exp_metadata = pd.DataFrame(
{'group': [0] * 5 + [1] * 5,
'n_diff': [2] * 10,
'effect_size': [10.0] * 10,
'library_size': [10000] * 10
},
index = ['S0', 'S1', 'S2', 'S3', 'S4',
'S5', 'S6', 'S7', 'S8', 'S9'],
)
metadata = metadata.reindex_axis(sorted(metadata.columns), axis=1)
exp_metadata = exp_metadata.reindex_axis(sorted(exp_metadata.columns), axis=1)
pdt.assert_frame_equal(metadata, exp_metadata)
exp_truth = ['F0', 'F4']
self.assertListEqual(truth, exp_truth)
def compositional_effect_size(max_alpha, reps, intervals, n_species, n_diff,
n_contaminants, lam, library_size, balanced,
output_dir):
os.mkdir(output_dir)
gen = compositional_effect_size_generator(max_alpha,
reps,
intervals,
n_species,
n_diff,
n_contaminants,
lam,
library_size=library_size,
balanced=balanced)
for i, g in enumerate(gen):
table, groups, truth = g
output_table = "%s/table.%d.biom" % (output_dir, i)
output_groups = "%s/metadata.%d.txt" % (output_dir, i)
output_truth = "%s/truth.%d.csv" % (output_dir, i)
deposit(table, groups, truth, output_table, output_groups,
output_truth)
def test_composition_effect_template(self):
# test template
np.random.seed(0)
gen = compositional_effect_size_generator(
max_alpha=1, reps=5,
intervals=2, n_species=5, n_diff=1,
n_contaminants=2, lam=0.1,
fold_balance=False,
template=np.array([7.0, 3.0, 1.0, 1.0, 2.0, 4.0, 6.0, 1.0, 10.0]))
table, metadata, truth = next(gen)
table, metadata, truth = next(gen)
exp_table = pd.DataFrame(
np.array([
[0.227273, 0.045455, 0.022727, 0.079545,
0.125, 0.499977, 0.000023],
[0.227273, 0.045455, 0.022727, 0.079545,
0.125, 0.499977, 0.000023],
[0.227273, 0.045455, 0.022727, 0.079545,
0.125, 0.499977, 0.000023],
[0.227273, 0.045455, 0.022727, 0.079545,
0.125, 0.499977, 0.000023],
[0.227273, 0.045455, 0.022727, 0.079545,
0.125, 0.499977, 0.000023],
[0.008000, 0.016000, 0.008000, 0.028000,
0.440, 0.499977, 0.000023],
[0.008000, 0.016000, 0.008000, 0.028000,
0.440, 0.499977, 0.000023],
[0.008000, 0.016000, 0.008000, 0.028000,
0.440, 0.499977, 0.000023],
[0.008000, 0.016000, 0.008000, 0.028000,
0.440, 0.499977, 0.000023],
[0.008000, 0.016000, 0.008000, 0.028000,
0.440, 0.499977, 0.000023]]),
index = ['S0', 'S1', 'S2', 'S3', 'S4',
'S5', 'S6', 'S7', 'S8', 'S9'],
columns = ['F0', 'F1', 'F2', 'F3', 'F4', 'X0', 'X1']
)
npt.assert_allclose(table.values, exp_table.values,
atol=1e-2, rtol=1e-2)
def test_composition_effect_size_simple(self):
gen = compositional_effect_size_generator(max_alpha=1, reps=5,
intervals=2, n_species=5, n_diff=1,
n_contaminants=2, lam=0.1)
table, metadata, truth = next(gen)
exp_table = pd.DataFrame(
np.vstack((
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269]),
np.array([.1, .1, .1, .1, .1, 0.499977, 0.00002269])
)),
index = ['S0', 'S1', 'S2', 'S3', 'S4',
'S5', 'S6', 'S7', 'S8', 'S9'],
columns = ['F0', 'F1', 'F2', 'F3', 'F4', 'X0', 'X1']
)
pdt.assert_frame_equal(table, exp_table, check_less_precise=True)
exp_metadata = pd.DataFrame(
{'group': [0] * 5 + [1] * 5,
'n_diff': [2] * 10,
'effect_size': [1.0] * 10,
'library_size': [10000] * 10
},
index = ['S0', 'S1', 'S2', 'S3', 'S4',
'S5', 'S6', 'S7', 'S8', 'S9'],
)
metadata = metadata.reindex_axis(sorted(metadata.columns), axis=1)
exp_metadata = exp_metadata.reindex_axis(sorted(exp_metadata.columns), axis=1)
pdt.assert_frame_equal(metadata, exp_metadata)
exp_truth = ['F0', 'F4']
self.assertListEqual(truth, exp_truth)
# test to see if the groups are different
table, metadata, truth = next(gen)
exp_table = pd.DataFrame(
closure(
np.vstack((
np.array([0.357143] + [0.035714]*4 +
[0.499977, 0.00002269]),
np.array([0.357143] + [0.035714]*4 +
[0.499977, 0.00002269]),
np.array([0.357143] + [0.035714]*4 +
[0.499977, 0.00002269]),
np.array([0.357143] + [0.035714]*4 +
[0.499977, 0.00002269]),
np.array([0.357143] + [0.035714]*4 +
[0.499977, 0.00002269]),
np.array([0.035714]*4 + [0.357143] +
[0.499977, 0.00002269]),
np.array([0.035714]*4 + [0.357143] +
[0.499977, 0.00002269]),
np.array([0.035714]*4 + [0.357143] +
[0.499977, 0.00002269]),
np.array([0.035714]*4 + [0.357143] +
[0.499977, 0.00002269]),
np.array([0.035714]*4 + [0.357143] +
[0.499977, 0.00002269])
))),
index = ['S0', 'S1', 'S2', 'S3', 'S4',
'S5', 'S6', 'S7', 'S8', 'S9'],
columns = ['F0', 'F1', 'F2', 'F3', 'F4', 'X0', 'X1']
)
pdt.assert_frame_equal(table, exp_table, check_less_precise=True)
exp_metadata = pd.DataFrame(
{'group': [0] * 5 + [1] * 5,
'n_diff': [2] * 10,
'effect_size': [10.0] * 10,
'library_size': [10000] * 10
},
index = ['S0', 'S1', 'S2', 'S3', 'S4',
'S5', 'S6', 'S7', 'S8', 'S9'],
)
metadata = metadata.reindex_axis(sorted(metadata.columns), axis=1)
exp_metadata = exp_metadata.reindex_axis(sorted(exp_metadata.columns), axis=1)
pdt.assert_frame_equal(metadata, exp_metadata)
exp_truth = ['F0', 'F4']
self.assertListEqual(truth, exp_truth)
