def test_genome_wide_norm(self):
"""Tests functionality to extract Obs/Exp values
genome-wide with median normalization from synthetic Hi-C data."""
position_frame = pd.read_csv("testFiles/posPileups.csv")
position_frame.loc[:, "mid"] = position_frame["pos"]
arms = pd.DataFrame({
"chrom": "chrSyn",
"start": 0,
"end": 4990000
},
index=[0])
cooler_file = cooler.Cooler(
"testFiles/test2.mcool::/resolutions/10000")
exp_f = HT.get_expected(cooler_file, arms, ignore_diagonals=0)
pairing_score = HT.get_pairing_score_obs_exp(cooler_file,
exp_f,
50000,
arms=arms,
norm=True)
expected = pd.read_csv(
"testFiles/test_pairingScore_obsExp_genomeWide_Norm.csv",
dtype={
name: pairing_score.dtypes[name]
for name in pairing_score.dtypes.index
},
)
assert_frame_equal(pairing_score, expected)
def test_no_collapse_real_data(self):
positions = pd.read_csv("testFiles/testAssignRegions.csv")
arms = HT.get_arms_hg19()
cooler_file = cooler.Cooler(
"testFiles/test3_realdata.mcool::resolutions/50000")
expected = HT.get_expected(cooler_file, arms)
result = HT.do_pileup_obs_exp(cooler_file,
expected,
positions,
proc=1,
collapse=False)
expected = np.load(
"testFiles/real_data_obsexp_pileup_not_collapsed.npy")
self.assertTrue(np.allclose(result, expected, equal_nan=True))
def test_synthetic_data(self):
"""Tests expected counts for synthetic Hi-C data.
Known values were provided and expected counts for each diagonal
calculated."""
result = HT.get_expected(self.cooler,
self.arms,
proc=1,
ignore_diagonals=0)
check = pd.read_csv("testFiles/test_expected_chrSyn.csv")
# merge regions for new expected format
check.loc[:, "region"] = check.apply(
lambda x: f"{x['chrom']}:{x['start']}-{x['end']}", axis=1)
check_final = check.drop(
columns=["chrom", "start", "end"])[result.columns]
assert_frame_equal(result, check_final)
def test_synthetic_data_mult_chroms(self):
"""Tests expected counts for synthetic Hi-C data
with multiple chromosomal arms. Known values were
provided and expected counts for each diagonal
calculated."""
arms = pd.DataFrame({
"chrom": ["chrSyn", "chrSyn"],
"start": [0, 2000000],
"end": [2000000, 4990000],
})
result = HT.get_expected(self.cooler, arms, proc=1, ignore_diagonals=0)
check = pd.read_csv("testFiles/test_expected_multiple_chroms.csv")
# merge regions for new expected format
check.loc[:, "region"] = check.apply(
lambda x: f"{x['chrom']}:{x['start']}-{x['end']}", axis=1)
check_final = check.drop(
columns=["chrom", "start", "end"])[result.columns]
assert_frame_equal(result, check_final)
def test_collapse(self):
"""Tests pileup of synthetic Hi-C data, with collapsing results."""
position_frame = pd.read_csv("testFiles/posPileups.csv")
arms = pd.DataFrame({
"chrom": "chrSyn",
"start": 0,
"end": 4990000
},
index=[0])
assigned = HT.assign_regions(50000, 10000, position_frame["chrom"],
position_frame["pos"], arms)
cooler_file = cooler.Cooler(
"testFiles/test2.mcool::/resolutions/10000")
exp_f = HT.get_expected(cooler_file, arms, ignore_diagonals=0)
result = HT.do_pileup_obs_exp(cooler_file,
exp_f,
assigned,
proc=1,
collapse=True)
expected = np.load("testFiles/test_pileups_obsExp_collapse.npy")
self.assertTrue(np.allclose(result, expected))
def test_expected_real_data(self):
"""Tests expected counts for real Hi-C data
with multiple chromosomal arms. Known values were
provided and expected counts for each diagonal
calculated."""
arms = HT.get_arms_hg19()
cooler_file = cooler.Cooler(
"testFiles/test3_realdata.mcool::/resolutions/50000")
result = HT.get_expected(
cooler_file, arms, proc=1,
ignore_diagonals=0).drop(columns=["count.sum"])
check = pd.read_csv("testFiles/test_expected_realdata.csv")
# merge regions for new expected format
check.loc[:, "region"] = check.apply(
lambda x: f"{x['chrom']}:{x['start']}-{x['end']}", axis=1)
check_final = (check.drop(
columns=["chrom", "start", "end"])[result.columns].sort_values(
by=["region", "diag"]).reset_index(drop=True))
sorted_result = result.sort_values(by=["region", "diag"]).reset_index(
drop=True)
assert_frame_equal(sorted_result, check_final)
def test_specific_regions(self):
"""Tests functionality to extract Obs/Exp values
at specific regions from synthetic Hi-C data."""
position_frame = pd.read_csv("testFiles/posPileups.csv")
position_frame.loc[:, "mid"] = position_frame["pos"]
arms = pd.DataFrame({
"chrom": "chrSyn",
"start": 0,
"end": 4990000
},
index=[0])
cooler_file = cooler.Cooler(
"testFiles/test2.mcool::/resolutions/10000")
exp_f = HT.get_expected(cooler_file, arms, ignore_diagonals=0)
pairing_score = HT.get_pairing_score_obs_exp(cooler_file,
exp_f,
50000,
regions=position_frame,
arms=arms,
norm=False)
expected = pd.read_csv(
"testFiles/test_pairingScore_obsExp_specificRegions.csv")
assert_frame_equal(pairing_score, expected)
def test_equal_sized_windows(self):
"""Test flexible pileup with equally sized windows (obs/exp)."""
position_frame = pd.read_csv("testFiles/posPileupSymmetric.csv")
arms = pd.DataFrame({
"chrom": "chrSyn",
"start": 0,
"end": 4990000
},
index=[0])
cooler_file = cooler.Cooler(
"testFiles/test2.mcool::/resolutions/10000")
expected = HT.get_expected(cooler_file,
arms,
proc=2,
ignore_diagonals=0)
result = HT.extract_windows_different_sizes_obs_exp(
position_frame, arms, cooler_file, expected)
# load expected extracted windows
with open("testFiles/test_pilesup_symmetric_obs_exp.pickle",
"rb") as file_pointer:
expected = pickle.load(file_pointer)
self.assertTrue(
all(np.allclose(i, j) for i, j in zip(result, expected)))
def test_wrong_parameters(self):
"""Tests raising of error when specific
region pileup is done with the norm parameter set to True."""
position_frame = pd.read_csv("testFiles/posPileups.csv")
position_frame.loc[:, "mid"] = position_frame["pos"]
arms = pd.DataFrame({
"chrom": "chrSyn",
"start": 0,
"end": 4990000
},
index=[0])
cooler_file = cooler.Cooler(
"testFiles/test2.mcool::/resolutions/10000")
exp_f = HT.get_expected(cooler_file, arms, ignore_diagonals=0)
bad_call = partial(
HT.get_pairing_score_obs_exp,
cooler_file,
exp_f,
50000,
regions=position_frame,
arms=arms,
norm=True,
)
self.assertRaises(ValueError, bad_call)