def test_read_airr():
# Test that reading the files one-by-one or at once yields the same results
anndata_tra = read_airr(TESTDATA / "airr/rearrangement_tra.tsv")
anndata_trb = read_airr(TESTDATA / "airr/rearrangement_trb.tsv")
anndata_ig = read_airr(TESTDATA / "airr/rearrangement_ig.tsv")
anndata = read_airr([
TESTDATA / "airr/rearrangement_tra.tsv",
TESTDATA / "airr/rearrangement_trb.tsv",
TESTDATA / "airr/rearrangement_ig.tsv",
])
tra_cols = [
"IR_VJ_1_junction_aa",
"IR_VJ_1_junction",
"IR_VJ_1_v_call",
"IR_VJ_1_d_call",
"IR_VJ_1_j_call",
"IR_VJ_1_c_call",
"IR_VJ_1_consensus_count",
]
trb_cols = [x.replace("IR_VJ", "IR_VDJ") for x in tra_cols]
ig_cols = tra_cols + trb_cols
pdt.assert_frame_equal(
anndata.obs.loc[anndata.obs["IR_VJ_1_locus"] == "TRA", tra_cols],
anndata_tra.obs.loc[:, tra_cols],
check_categorical=False, # categories differ, obviously
check_dtype=False,
)
pdt.assert_frame_equal(
anndata.obs.loc[anndata.obs["IR_VDJ_1_locus"] == "TRB", trb_cols],
anndata_trb.obs.loc[:, trb_cols],
check_categorical=False, # categories differ, obviously
check_dtype=False,
)
pdt.assert_frame_equal(
anndata.obs.loc[anndata.obs["IR_VDJ_1_locus"] == "IGH", ig_cols],
anndata_ig.obs.loc[:, ig_cols],
check_categorical=False, # categories differ, obviously
check_dtype=False,
)
# test some fundamental values
assert anndata.obs.shape[0] == 5
cell1 = anndata.obs.loc["cell1", :]
cell2 = anndata.obs.loc["cell2", :]
cell3 = anndata.obs.loc["AAACCTGCAGCGTAAG-1", :]
assert cell1.name == "cell1"
assert cell1["IR_VJ_1_junction_aa"] == "CTRPKWESPMVDAFDIW"
assert cell1["IR_VDJ_2_junction_aa"] == "CQQYDNLQITF"
assert cell1["IR_VDJ_1_junction_aa"] == "CQQYYHTPYSF"
assert cell1["IR_VJ_1_locus"] == "TRA"
assert cell1["IR_VDJ_1_locus"] == "TRB"
assert cell2.name == "cell2"
# check that inferring the locus name from genes works
assert cell3["IR_VJ_1_locus"] == "IGL"
assert cell3["IR_VDJ_1_locus"] == "IGH"
def test_airr_roundtrip_conversion(anndata_from_10x_sample, tmp_path):
"""Test that writing and reading to and from AIRR format results in the
identity"""
anndata = anndata_from_10x_sample
tmp_file = tmp_path / "test.airr.tsv"
write_airr(anndata, tmp_file)
anndata2 = read_airr(tmp_file, include_fields=None)
_normalize_df_types(anndata.obs)
_normalize_df_types(anndata2.obs)
pdt.assert_frame_equal(anndata.obs,
anndata2.obs,
check_dtype=False,
check_categorical=False)
def test_read_airr():
# Test that reading the files one-by-one or at once yields the same results
anndata_tra = read_airr("tests/data/airr/rearrangement_tra.tsv")
anndata_trb = read_airr("tests/data/airr/rearrangement_trb.tsv")
anndata = read_airr(
[
"tests/data/airr/rearrangement_tra.tsv",
"tests/data/airr/rearrangement_trb.tsv",
]
)
tra_cols = [
"TRA_1_cdr3",
"TRA_1_cdr3_nt",
"TRA_1_v_gene",
"TRA_1_d_gene",
"TRA_1_j_gene",
"TRA_1_c_gene",
"TRA_1_expr",
]
trb_cols = [x.replace("TRA", "TRB") for x in tra_cols]
pdt.assert_frame_equal(anndata.obs[tra_cols], anndata_tra.obs[tra_cols])
pdt.assert_frame_equal(anndata.obs[trb_cols], anndata_trb.obs[trb_cols])
# test some fundamental values
assert "cell1" in anndata.obs_names and "cell2" in anndata.obs_names
assert anndata.obs.shape[0] == 3
cell1 = anndata.obs.loc["cell1", :]
cell2 = anndata.obs.loc["cell2", :]
assert cell1.name == "cell1"
assert cell1["TRA_1_cdr3"] == "CTRPKWESPMVDAFDIW"
assert cell1["TRB_2_cdr3"] == "CQQYDNLQITF"
assert cell1["TRB_1_cdr3"] == "CQQYYHTPYSF"
assert cell2.name == "cell2"
def test_read_airr_issue280():
"""Test that reading the example shown in issue #280 works."""
anndata = read_airr(TESTDATA / "airr" / "tra_issue_280.tsv")
assert anndata.obs["IR_VDJ_1_junction_aa"][0] == "CASSLGGESQNTLYF"
assert anndata.obs["IR_VJ_1_junction_aa"][0] == "CAARGNRIFF"