def test_init_fdr(self, fetch_formulas_mock):
ds_config = {
'analysis_version': 1,
'fdr': {
'decoy_sample_size': 20
},
'isotope_generation': BASIC_ISOTOPE_GENERATION_CONFIG,
}
moldb_fdr_list = init_fdr(ds_config,
[MolecularDB(0, 'test_db', 'version')])
assert len(moldb_fdr_list) == 1
_, fdr = moldb_fdr_list[0]
assert not fdr.td_df.empty
def test_compute_fdr(spark_context, ds_config):
moldb_fdr_list = init_fdr(ds_config, [MolecularDB(0, 'test_db', 'version')])
_, fdr = moldb_fdr_list[0]
formula_map_df = collect_ion_formulas(spark_context, moldb_fdr_list).drop('moldb_id', axis=1)
formula_metrics_df = pd.DataFrame(
[(10, 'H3O', 0.99), (11, 'C5H4O', 0.5), (12, 'H2ONa', 0.1)],
columns=['formula_i', 'ion_formula', 'msm'],
).set_index('formula_i')
metrics_df = compute_fdr(fdr, formula_metrics_df, formula_map_df, None)
assert len(metrics_df) == 3
assert sorted(metrics_df.columns.tolist()) == sorted(
['ion_formula', 'msm', 'formula', 'modifier', 'fdr']
)
def test_decoy_sample_size_30(self, fetch_formulas_mock, spark_context):
ds_config = {
'analysis_version': 1,
'fdr': {
'decoy_sample_size': 30
},
'isotope_generation': BASIC_ISOTOPE_GENERATION_CONFIG,
}
moldb_fdr_list = init_fdr(ds_config,
[MolecularDB(0, 'test_db', 'version')])
df = collect_ion_formulas(spark_context, moldb_fdr_list)
assert df.columns.tolist() == [
'moldb_id', 'ion_formula', 'formula', 'modifier'
]
assert df.shape == (62, 4)
def test_neutral_losses_and_chem_mods(self, fetch_formulas_mock,
spark_context):
ds_config = {
'analysis_version': 1,
'fdr': {
'decoy_sample_size': 1
},
'isotope_generation': FULL_ISOTOPE_GENERATION_CONFIG,
}
moldb_fdr_list = init_fdr(ds_config,
[MolecularDB(0, 'test_db', 'version')])
df = collect_ion_formulas(spark_context, moldb_fdr_list)
assert df.columns.tolist() == [
'moldb_id', 'ion_formula', 'formula', 'modifier'
]
# 2 formulas * (4 target adducts + (4 target adducts * 1 decoy adducts per target adduct)
# * (no loss + 2 neutral losses) * (no mod + 1 chem mod) = 2 * (4 + 4) * 3 * 2 = 96
assert df.shape == (96, 4)