def _run_statistical_test_groups(self, df: pd.DataFrame, group_col: str,
stats_test: str, correction_method: str,
structure_pval: str, sym: bool):
if correction_method is not None:
pval = statistical_test_groups_comparison(
df[self.DIVERSITY_INDEXES_NAME],
df[group_col],
stats_test,
output='series',
sym=False)
# dropping NaN (= comparison that couldn't have been generated due to too few samples in one or both groups)
corrected_pval = pd.Series(
multipletests(pval.dropna(),
alpha=0.05,
method=correction_method)[1])
corrected_pval.index = pval.dropna(
).index # postulate that the order hasn't changed
if pval[pval.isnull()].size > 0:
corrected_pval = corrected_pval.append(pval[pval.isnull()])
# remodelling of p-values output
corrected_pval = self._structure_remodelling(
corrected_pval, structure=structure_pval, sym=sym)
return corrected_pval
else:
pval = statistical_test_groups_comparison(
df[self.DIVERSITY_INDEXES_NAME],
df[group_col],
stats_test,
output=structure_pval,
sym=sym)
return pval
def test_mann_whitney_u_groups_wrong_metadata(self):
metadata_df = pd.DataFrame(
[['F', 2], ['M', 1]],
columns=['sex', 'group'],
index=['sample6', 'sample7'],
)
with self.assertRaises(RuntimeError) as cm:
statistical_test_groups_comparison(self.test_df,
metadata_df['group'],
stat_test='mann_whitney_u')
the_exception = cm.exception
self.assertEqual(
the_exception.__str__(),
"All groups have been dropped: not enough observations by group.")
def test_chi2_contingency_groups(self):
expected_df = pd.DataFrame(
[[np.nan, 0.06720551273974977, 0.9310443064194389],
[0.06720551273974977, np.nan, 0.152090],
[0.9310443064194389, 0.152090, np.nan]],
columns=[1, 2, 3],
index=[1, 2, 3])
matrix = statistical_test_groups_comparison(
self.test_df, self.metadata_df, stat_test='chi2_contingency')
pd.testing.assert_frame_equal(matrix, expected_df, check_dtype=False)
def test_ttest_independance_groups(self):
metadata_df = pd.DataFrame(
[['F', 2], ['M', 1], ['F', 1], ['M', 1], ['M', 2]],
columns=['sex', 'group'],
index=['sample1', 'sample2', 'sample3', 'sample4', 'sample5'],
)
expected_df = pd.DataFrame([[np.nan, 0.483379], [0.483379, np.nan]],
columns=[1, 2],
index=[1, 2])
matrix = statistical_test_groups_comparison(
self.test_df, metadata_df['group'], stat_test='ttest_independence')
pd.testing.assert_frame_equal(matrix, expected_df, check_dtype=False)
def test_mann_whitney_u_groups_nonsym_series_output(self):
metadata_df = pd.DataFrame(
[['F', 2], ['M', 1], ['F', 1], ['M', 1], ['M', 2]],
columns=['sex', 'group'],
index=['sample1', 'sample2', 'sample3', 'sample4', 'sample5'],
)
expected_df = pd.Series({(1, 2): 0.3742593192802244})
expected_df.index.names = ['Group', 'Group']
matrix = statistical_test_groups_comparison(self.test_df,
metadata_df['group'],
stat_test='mann_whitney_u',
output='series',
sym=False)
pd.testing.assert_series_equal(matrix, expected_df, check_dtype=False)
def test_mann_whitney_u_groups_nonsym(self):
metadata_df = pd.DataFrame(
[['F', 2], ['M', 1], ['F', 1], ['M', 1], ['M', 2]],
columns=['sex', 'group'],
index=['sample1', 'sample2', 'sample3', 'sample4', 'sample5'],
)
expected_df = pd.DataFrame([[np.nan, 0.374259], [np.nan, np.nan]],
columns=[1, 2],
index=[1, 2])
matrix = statistical_test_groups_comparison(self.test_df,
metadata_df['group'],
stat_test='mann_whitney_u',
output='dataframe',
sym=False)
pd.testing.assert_frame_equal(matrix, expected_df, check_dtype=False)