def get_involved_complex_from_counts(multidatas_counts: pd.DataFrame, clusters_names: list,
complex_expanded: pd.DataFrame, complex_composition: pd.DataFrame) -> (
pd.DataFrame, pd.DataFrame):
"""
Finds the complexes defined in counts and calculates the counts values
"""
proteins_in_complexes = complex_composition['protein_multidata_id'].tolist()
# Remove counts that can't be part of a complex
multidatas_counts_filtered = multidatas_counts[
multidatas_counts['id_multidata'].apply(lambda multidata: multidata in proteins_in_complexes)]
# Find complexes with all components defined in counts
complex_composition_counts = complex_helper.get_involved_complex_from_protein(multidatas_counts_filtered,
complex_expanded,
complex_composition,
drop_duplicates=False)
if complex_composition_counts.empty:
return pd.DataFrame(), pd.DataFrame()
# Remove counts that are not defined in selected complexes
multidatas_counts_filtered = filter_counts_by_genes(multidatas_counts_filtered,
complex_composition_counts['gene'].tolist())
# Set the counts value a complex count. This is the minimum value of the cell component
complex_counts = cluster_counts_helper.merge_complex_counts(clusters_names, complex_composition_counts,
list(complex_expanded.columns.values))
# Removes empty counts
complex_counts = cluster_counts_filter.filter_empty_cluster_counts(complex_counts, clusters_names)
complex_counts.drop(clusters_names, axis=1, inplace=True)
return complex_counts, multidatas_counts_filtered
def test_get_involved_complex_from_protein_empty_result(self):
proteins = pd.read_csv('{}/helper_complex_protein.csv'.format(
self.FIXTURES_SUBPATH))
proteins.drop(proteins.index, inplace=True)
complexes = pd.read_csv('{}/helper_complex_complex.csv'.format(
self.FIXTURES_SUBPATH))
complex_composition = pd.read_csv(
'{}/helper_complex_complex_composition.csv'.format(
self.FIXTURES_SUBPATH))
result = complex_helper.get_involved_complex_from_protein(
proteins, complexes, complex_composition, drop_duplicates=False)
self.assertTrue(
dataframe_functions.dataframes_has_same_data(
result, pd.DataFrame()))
def test_get_involved_complex_from_protein(self):
proteins = pd.read_csv('{}/helper_complex_protein.csv'.format(
self.FIXTURES_SUBPATH))
complexes = pd.read_csv('{}/helper_complex_complex.csv'.format(
self.FIXTURES_SUBPATH))
complex_composition = pd.read_csv(
'{}/helper_complex_complex_composition.csv'.format(
self.FIXTURES_SUBPATH))
result_expected = pd.read_csv('{}/helper_complex_result.csv'.format(
self.FIXTURES_SUBPATH))
result = complex_helper.get_involved_complex_from_protein(
proteins, complexes, complex_composition, drop_duplicates=False)
self.assertTrue(
dataframe_functions.dataframes_has_same_data(
result, result_expected))
def get_complex_involved_in_counts(
multidatas_counts: pd.DataFrame, clusters_names: list,
complex_composition: pd.DataFrame,
complex_expanded: pd.DataFrame) -> pd.DataFrame:
"""
Gets complexes involved in counts
"""
core_logger.debug('Finding Complexes')
complex_counts_composition = complex_helper.get_involved_complex_from_protein(
multidatas_counts,
complex_expanded,
complex_composition,
drop_duplicates=False)
complex_counts = merge_complex_counts(
clusters_names, complex_counts_composition,
list(complex_expanded.columns.values))
complex_counts = filter_empty_cluster_counts(complex_counts,
clusters_names)
complex_counts.reset_index(drop=True, inplace=True)
return complex_counts
