def test_amusr_regression(self):
des = [np.array([[1, 1, 3], [0, 0, 2], [0, 0, 1]]).astype(float),
np.array([[1, 1, 3], [0, 0, 2], [0, 0, 1]]).astype(float)]
res = [np.array([1, 2, 3]).reshape(-1, 1).astype(float),
np.array([1, 2, 3]).reshape(-1, 1).astype(float)]
tfs = ['tf1', 'tf2', 'tf3']
targets = ['gene1', 'gene2']
priors = [pd.DataFrame([[0, 1, 1], [1, 0, 1]], index=targets, columns=tfs),
pd.DataFrame([[0, 0, 1], [1, 0, 1]], index=targets, columns=tfs)]
gene1_prior = amusr_regression.format_prior(priors, 'gene1', [0, 1], 1.)
gene2_prior = amusr_regression.format_prior(priors, 'gene2', [0, 1], 1.)
output = [amusr_regression.run_regression_EBIC(des, res, ['tf1', 'tf2', 'tf3'], [0, 1], 'gene1', gene1_prior,
scale_data=True, use_numba=self.use_numba),
amusr_regression.run_regression_EBIC(des, res, ['tf1', 'tf2', 'tf3'], [0, 1], 'gene2', gene2_prior,
scale_data=True, use_numba=self.use_numba)]
out0 = pd.DataFrame([['tf3', 'gene1', -1, 1],
['tf3', 'gene1', -1, 1]],
index=pd.MultiIndex(levels=[[0, 1], [0]],
codes=[[0, 1], [0, 0]]),
columns=['regulator', 'target', 'weights', 'resc_weights'])
out1 = pd.DataFrame([['tf3', 'gene2', -1, 1],
['tf3', 'gene2', -1, 1]],
index=pd.MultiIndex(levels=[[0, 1], [0]],
codes=[[0, 1], [0, 0]]),
columns=['regulator', 'target', 'weights', 'resc_weights'])
pdt.assert_frame_equal(pd.concat(output[0]), out0, check_dtype=False)
pdt.assert_frame_equal(pd.concat(output[1]), out1, check_dtype=False)
def test_format_priors_pweight(self):
tfs = ['tf1', 'tf2']
priors = [pd.DataFrame([[0, 1], [1, 0]], index=['gene1', 'gene2'], columns=tfs),
pd.DataFrame([[0, 0], [1, 0]], index=['gene1', 'gene2'], columns=tfs)]
gene1_prior = amusr_regression.format_prior(priors, 'gene1', [0, 1], 1.2)
gene2_prior = amusr_regression.format_prior(priors, 'gene2', [0, 1], 1.2)
npt.assert_almost_equal(gene1_prior, np.array([[1.09090909, 1.], [0.90909091, 1.]]))
npt.assert_almost_equal(gene2_prior, np.array([[0.90909091, 0.90909091], [1.09090909, 1.09090909]]))
def test_format_priors_noweight(self):
runner = amusr_regression.AMuSR_regression([pd.DataFrame()],
[pd.DataFrame()], None)
tfs = ['tf1', 'tf2']
priors = [
pd.DataFrame([[0, 1], [1, 0]],
index=['gene1', 'gene2'],
columns=tfs),
pd.DataFrame([[0, 0], [1, 0]],
index=['gene1', 'gene2'],
columns=tfs)
]
gene1_prior = amusr_regression.format_prior(priors, 'gene1', [0, 1], 1)
gene2_prior = amusr_regression.format_prior(priors, 'gene2', [0, 1], 1)
npt.assert_almost_equal(gene1_prior, np.array([[1., 1.], [1., 1.]]))
npt.assert_almost_equal(gene2_prior, np.array([[1., 1.], [1., 1.]]))
def regression_maker(j, x_df, y_list, prior, tf):
level = 0 if j % 100 == 0 else 2
utils.Debug.allprint(base_regression.PROGRESS_STR.format(gn=genes[j],
i=j,
total=G),
level=level)
gene = genes[j]
x, y, tasks = [], [], []
if remove_autoregulation:
tf = [t for t in tf if t != gene]
else:
pass
for k, y_data in y_list:
x.append(x_df[k].get_gene_data(tf)) # list([N, K])
y.append(y_data)
tasks.append(k) # [T,]
prior = format_prior(prior, gene, tasks, prior_weight, tfs=tf)
return j, regression_function(x,
y,
tf,
tasks,
gene,
prior,
lambda_Bs=lambda_Bs,
lambda_Ss=lambda_Ss,
Cs=Cs,
Ss=Ss,
tol=tol,
rel_tol=rel_tol,
use_numba=use_numba)
def test_amusr_regression(self):
des = [
np.array([[1, 1, 3], [0, 0, 2], [0, 0, 1]]).astype(float),
np.array([[1, 1, 3], [0, 0, 2], [0, 0, 1]]).astype(float)
]
res = [
np.array([1, 2, 3]).reshape(-1, 1).astype(float),
np.array([1, 2, 3]).reshape(-1, 1).astype(float)
]
tfs = ['tf1', 'tf2', 'tf3']
targets = ['gene1', 'gene2']
priors = [
pd.DataFrame([[0, 1, 1], [1, 0, 1]], index=targets, columns=tfs),
pd.DataFrame([[0, 0, 1], [1, 0, 1]], index=targets, columns=tfs)
]
runner = amusr_regression.AMuSR_regression(
[pd.DataFrame(des[0], columns=tfs)],
[pd.DataFrame(res[0], columns=["gene1"])], None)
gene1_prior = amusr_regression.format_prior(priors, 'gene1', [0, 1],
1.)
gene2_prior = amusr_regression.format_prior(priors, 'gene2', [0, 1],
1.)
output = []
output.append(
amusr_regression.run_regression_EBIC(des, res,
['tf1', 'tf2', 'tf3'], [0, 1],
'gene1', gene1_prior))
output.append(
amusr_regression.run_regression_EBIC(des, res,
['tf1', 'tf2', 'tf3'], [0, 1],
'gene2', gene2_prior))
out0 = pd.DataFrame(
[['tf3', 'gene1', -1, 1], ['tf3', 'gene1', -1, 1]],
index=pd.MultiIndex(levels=[[0, 1], [0]], labels=[[0, 1], [0, 0]]),
columns=['regulator', 'target', 'weights', 'resc_weights'])
out1 = pd.DataFrame(
[['tf3', 'gene2', -1, 1], ['tf3', 'gene2', -1, 1]],
index=pd.MultiIndex(levels=[[0, 1], [0]], labels=[[0, 1], [0, 0]]),
columns=['regulator', 'target', 'weights', 'resc_weights'])
pdt.assert_frame_equal(pd.concat(output[0]), out0, check_dtype=False)
pdt.assert_frame_equal(pd.concat(output[1]), out1, check_dtype=False)
def regression_maker(j, x_df, y_list, prior, tf):
level = 0 if j % 100 == 0 else 2
utils.Debug.allprint(base_regression.PROGRESS_STR.format(gn=genes[j],
i=j,
total=G),
level=level)
gene = genes[j]
x, y, tasks = [], [], []
if remove_autoregulation:
tf = [t for t in tf if t != gene]
else:
pass
for k, y_data in y_list:
x.append(x_df[k].loc[:, tf].values) # list([N, K])
y.append(y_data)
tasks.append(k) # [T,]
del y_list
prior = format_prior(prior, gene, tasks, prior_weight)
return j, run_regression_EBIC(x, y, tf, tasks, gene, prior)