def test_run_with_kegg_returns_dicts_of_dataframes_with_shape(self):
results = _.all(expression_table=self.expression_table, )
self.assertIsInstance(results, dict)
self.assertIsInstance(results['harmonic'], pd.DataFrame)
self.assertEqual(results['harmonic'].shape, (324, 3))
self.assertEqual(results['geometric'].shape, (324, 3))
self.assertEqual(results['min_p_val'].shape, (324, 3))
def test_pathway_assessor_returns_none_if_statistic_is_set_to_false(self):
results = _.all(
expression_table=self.expression_table,
pathways=self.user_pathway_db,
geometric=False,
min_p_val=False,
)
self.assertIsNone(results['geometric'])
self.assertIsNone(results['min_p_val'])
def test_pathway_assessor_returns_three_dataframes_of_expected_values(
self):
user_pathways = {
'Sample_pathway': ['SLC2A6', 'PHOSPHO1', 'PIKFYVE', 'VHL']
}
results = _.all(expression_table=self.expression_table,
pathways=user_pathways)
harmonic_avg = results['harmonic'].loc['Sample_pathway'].to_dict()
geometric_avg = results['geometric'].loc['Sample_pathway'].to_dict()
min_p_vals = results['min_p_val'].loc['Sample_pathway'].to_dict()
expected_harmonic_avg = {
'Sample_A': 8.610084236222475,
'Sample_B': 9.519349644266763,
'Sample_C': 6.02244237008846
}
expected_geometric_avg = {
'Sample_A': 5.916555748731008,
'Sample_B': 7.153859966001466,
'Sample_C': 4.75439878004548
}
expected_min_p_vals = {
'Sample_A': 9.690912952571226,
'Sample_B': 10.576744476960645,
'Sample_C': 6.695180679017199
}
self.assertEqual(len(results), 3)
self.assertAlmostEqual(harmonic_avg['Sample_A'],
expected_harmonic_avg['Sample_A'])
self.assertAlmostEqual(harmonic_avg['Sample_B'],
expected_harmonic_avg['Sample_B'])
self.assertAlmostEqual(harmonic_avg['Sample_C'],
expected_harmonic_avg['Sample_C'])
self.assertAlmostEqual(geometric_avg['Sample_A'],
expected_geometric_avg['Sample_A'])
self.assertAlmostEqual(geometric_avg['Sample_B'],
expected_geometric_avg['Sample_B'])
self.assertAlmostEqual(geometric_avg['Sample_C'],
expected_geometric_avg['Sample_C'])
self.assertAlmostEqual(min_p_vals['Sample_A'],
expected_min_p_vals['Sample_A'])
self.assertAlmostEqual(min_p_vals['Sample_B'],
expected_min_p_vals['Sample_B'])
self.assertAlmostEqual(min_p_vals['Sample_C'],
expected_min_p_vals['Sample_C'])
def write_table(score_type):
if ascending:
direction = 'suppression'
else:
direction = 'activation'
output_f = '{}/scores/{}_{}_{}_{}'.format(examples_dir, tumor, pw_name, direction, score_type)
scores[score_type].to_csv(output_f, sep='\t')
print('Finished: {}'.format(output_f))
if __name__ == '__main__':
pa_home = os.getenv('PATHWAY_ASSESSOR_HOME')
tumor = 'blca_normal_slim'
pw_name = 'kegg_immune'
db_dir = '{}/pathway_assessor/databases/as_tables'.format(pa_home)
ascending = True
examples_dir = os.path.dirname(os.path.abspath(__file__))
expression_table_f = '{}/{}'.format(examples_dir, tumor)
pathways_f = '{}/{}.tsv'.format(db_dir, pw_name)
expression_table = pd.read_csv(expression_table_f, sep='\t', index_col=0)
pathways = pathways_dict(pathways_f)
scores = pa.all(expression_table=expression_table, ascending=ascending, pathways=pathways)
for method in scores:
write_table(method)
else:
user_pw_db_f = None
output_dir = output_dir_path(base_dir, expression_table_f, pathway_db_choice, ascending, rank_method)
os.makedirs(output_dir, exist_ok=True)
expression_df = pd.read_csv(expression_table_f, sep='\t', index_col=0)
if pathway_db_choice == 'xcell' \
or pathway_db_choice == 'xcell_complete_signatures' \
or pathway_db_choice == 'all' \
or pathway_db_choice == 'wikipathways' \
or pathway_db_choice == 'immune_all':
user_pathways = pickle.load(open('databases/{}.pkl'.format(pathway_db_choice), 'rb'))
scores = pa.all(expression_table=expression_df,
pathways=user_pathways,
ascending=ascending,
rank_method=rank_method)
elif not user_pw_db_f:
scores = pa.all(expression_table=expression_df,
db=pathway_db_choice,
ascending=ascending,
rank_method=rank_method)
else:
user_pathways = user_pathways_dict(user_pw_db_f)
scores = pa.all(expression_table=expression_df,
pathways=user_pathways,
ascending=ascending,
rank_method=rank_method)
for score_type in scores:
write_score_table(output_dir_path=output_dir, score_type=score_type, df=scores[score_type])
import pandas as pd
import pathway_assessor as pa
expression_table = "C:\\Users\\Boris\\Desktop\\PW_A_Local\\ucec.normal.txt"
expression_df = pd.read_csv(expression_table, sep='\t', index_col=0)
scores = pa.all(expression_table=expression_df,
db='hallmark',
ascending=True,
rank_method='max')
print(scores)