def test_merge_chrom(self):
s1 = SNPs("tests/input/generic.csv")
df = s1.snps.append(
self.create_snp_df(
rsid=["rs100", "rs101", "rs102", "rs103"],
chrom=["Y", "Y", "Y", "Y"],
pos=[100, 101, 102, 103],
genotype=["A", np.nan, "A", "A"],
))
s1._snps = df.copy()
s2 = SNPs()
s2._build = 37
s2._snps = df.copy()
# set values for chrom that will be ignored (that would otherwise result in
# identification of discrepant SNPs or updating genotype)
s2._snps.loc["rs3", "pos"] = 1003 # discrepant position
s2._snps.loc["rs4", "genotype"] = "AA" # discrepant genotype
s2._snps.loc["rs5", "genotype"] = "AA"
# set values for chrom to be merged
s2._snps.loc["rs100", "genotype"] = "T" # discrepant genotype
s2._snps.loc["rs101", "genotype"] = "A"
s2._snps.loc["rs102", "pos"] = 1002 # discrepant position
# set expected values for merge result
df.loc["rs100",
"genotype"] = np.nan # discrepant genotype sets to np.nan
df.loc["rs101", "genotype"] = "A" # updates np.nan
results = s1.merge([s2], chrom="Y")
pd.testing.assert_frame_equal(s1.snps, df, check_exact=True)
self.assert_results(
results,
[{
"merged":
True,
"common_rsids":
pd.Index(["rs100", "rs101", "rs102", "rs103"], name="rsid"),
"discrepant_position_rsids":
pd.Index(["rs102"], name="rsid"),
"discrepant_genotype_rsids":
pd.Index(["rs100"], name="rsid"),
}],
)
self.assertEqual(len(s1.discrepant_merge_positions), 1)
self.assertEqual(len(s1.discrepant_merge_genotypes), 1)
def test_appending_dfs(self):
s = SNPs()
s._snps = self.create_snp_df(rsid=["rs1"],
chrom=["1"],
pos=[1],
genotype=["AA"])
s._duplicate = self.create_snp_df(rsid=["rs1"],
chrom=["1"],
pos=[1],
genotype=["AA"])
s._discrepant_XY = self.create_snp_df(rsid=["rs1"],
chrom=["1"],
pos=[1],
genotype=["AA"])
s.merge([s])
df = self.create_snp_df(rsid=["rs1", "rs1"],
chrom=["1", "1"],
pos=[1, 1],
genotype=["AA", "AA"])
pd.testing.assert_frame_equal(s.duplicate, df, check_exact=True)
pd.testing.assert_frame_equal(s.discrepant_XY, df, check_exact=True)
pd.testing.assert_frame_equal(s.heterozygous_MT,
get_empty_snps_dataframe(),
check_exact=True)
pd.testing.assert_frame_equal(s.discrepant_vcf_position,
get_empty_snps_dataframe(),
check_exact=True)
def test__lookup_build_with_snp_pos_None(self):
snps = SNPs()
snps._snps = self.create_snp_df(rsid=["rs3094315"],
chrom=["1"],
pos=[1],
genotype=["AA"])
self.assertFalse(snps.detect_build())
def simulate_snps(
self,
chrom="1",
pos_start=1,
pos_max=248140902,
pos_step=100,
genotype="AA",
insert_nulls=True,
null_snp_step=101,
complement_genotype_one_chrom=False,
complement_genotype_two_chroms=False,
complement_snp_step=50,
):
s = SNPs()
s._build = 37
positions = np.arange(pos_start, pos_max, pos_step, dtype=np.uint32)
snps = pd.DataFrame(
{"chrom": chrom},
index=pd.Index(["rs" + str(x + 1) for x in range(len(positions))],
name="rsid"),
)
snps["pos"] = positions
snps["genotype"] = genotype
if insert_nulls:
snps.loc[snps.iloc[0::null_snp_step, :].index, "genotype"] = np.nan
indices = snps.iloc[0::complement_snp_step, :].index
if complement_genotype_two_chroms:
snps.loc[indices,
"genotype"] = snps.loc[indices, "genotype"].apply(
self.complement_two_chroms)
elif complement_genotype_one_chrom:
snps.loc[indices,
"genotype"] = snps.loc[indices, "genotype"].apply(
self.complement_one_chrom)
s._snps = snps
return s
def test_save_snps_vcf_discrepant_pos(self):
with tempfile.TemporaryDirectory() as tmpdir1:
s = SNPs("tests/input/testvcf.vcf", output_dir=tmpdir1)
r = Resources()
r._reference_sequences["GRCh37"] = {}
output = os.path.join(tmpdir1, "vcf_GRCh37.vcf")
with tempfile.TemporaryDirectory() as tmpdir2:
dest = os.path.join(tmpdir2, "generic.fa.gz")
gzip_file("tests/input/generic.fa", dest)
seq = ReferenceSequence(ID="1", path=dest)
r._reference_sequences["GRCh37"]["1"] = seq
# create discrepant SNPs by setting positions outside reference sequence
s._snps.loc["rs1", "pos"] = 0
s._snps.loc["rs17", "pos"] = 118
# esnure this is the right type after manual tweaking
s._snps = s._snps.astype({"pos": np.uint32})
self.assertEqual(s.save(vcf=True), output)
pd.testing.assert_frame_equal(
s.discrepant_vcf_position,
self.create_snp_df(
rsid=["rs1", "rs17"],
chrom=["1", "1"],
pos=[0, 118],
genotype=["AA", np.nan],
),
check_exact=True,
)
expected = self.generic_snps_vcf().drop(["rs1", "rs17"])
self.run_parsing_tests_vcf(output, snps_df=expected)
def test_merging_files_discrepant_snps(self):
with tempfile.TemporaryDirectory() as tmpdir:
dest1 = os.path.join(tmpdir, "discrepant_snps1.csv")
dest2 = os.path.join(tmpdir, "discrepant_snps2.csv")
df = pd.read_csv(
"tests/input/discrepant_snps.csv",
skiprows=1,
na_values="--",
names=[
"rsid",
"chrom",
"pos_file1",
"pos_file2",
"genotype_file1",
"genotype_file2",
"discrepant_position",
"discrepant_genotype",
"expected_position",
"expected_genotype",
],
index_col=0,
dtype={
"chrom": object,
"pos_file1": np.uint32,
"pos_file2": np.uint32,
"discrepant_position": bool,
"discrepant_genotype": bool,
"expected_position": np.uint32,
},
)
df1 = df[["chrom", "pos_file1", "genotype_file1"]]
df2 = df[["chrom", "pos_file2", "genotype_file2"]]
df1.to_csv(dest1,
na_rep="--",
header=["chromosome", "position", "genotype"])
df2.to_csv(dest2,
na_rep="--",
header=["chromosome", "position", "genotype"])
s = SNPs()
s.merge([SNPs(dest1), SNPs(dest2)])
expected = df[[
"chrom",
"discrepant_position",
"discrepant_genotype",
"expected_position",
"expected_genotype",
]]
expected = expected.rename(columns={
"expected_position": "pos",
"expected_genotype": "genotype"
})
expected_snps = SNPs()
expected_snps._snps = expected
expected_snps.sort()
expected = expected_snps.snps
pd.testing.assert_index_equal(
s.discrepant_merge_positions.index,
expected.loc[expected["discrepant_position"] == True].index,
)
pd.testing.assert_index_equal(
s.discrepant_merge_genotypes.index,
expected.loc[expected["discrepant_genotype"] == True].index,
)
pd.testing.assert_series_equal(s.snps["pos"], expected["pos"])
pd.testing.assert_series_equal(s.snps["genotype"],
expected["genotype"])
def test__lookup_build_with_snp_pos_None(self):
snps = SNPs()
snps._snps = self.snps_discrepant_pos()
assert not snps.detect_build()
def test_merging_files_discrepant_snps(self):
df = pd.read_csv(
"tests/input/discrepant_snps.csv",
skiprows=1,
na_values="--",
names=[
"rsid",
"chrom",
"pos_file1",
"pos_file2",
"genotype_file1",
"genotype_file2",
"discrepant_position",
"discrepant_genotype",
"expected_position",
"expected_genotype",
],
index_col=0,
dtype={
"chrom": object,
"pos_file1": np.int64,
"pos_file2": np.int64,
"discrepant_position": bool,
"discrepant_genotype": bool,
},
)
df1 = df[["chrom", "pos_file1", "genotype_file1"]]
df2 = df[["chrom", "pos_file2", "genotype_file2"]]
df1.to_csv(
"tests/input/discrepant_snps1.csv",
na_rep="--",
header=["chromosome", "position", "genotype"],
)
df2.to_csv(
"tests/input/discrepant_snps2.csv",
na_rep="--",
header=["chromosome", "position", "genotype"],
)
sc = SNPsCollection([
"tests/input/discrepant_snps1.csv",
"tests/input/discrepant_snps2.csv"
])
expected = df[[
"chrom",
"discrepant_position",
"discrepant_genotype",
"expected_position",
"expected_genotype",
]]
expected = expected.rename(columns={
"expected_position": "pos",
"expected_genotype": "genotype"
})
expected_snps = SNPs()
expected_snps._snps = expected
expected_snps.sort_snps()
expected = expected_snps.snps
pd.testing.assert_index_equal(
sc.discrepant_positions.index,
expected.loc[expected["discrepant_position"] == True].index,
)
pd.testing.assert_index_equal(
sc.discrepant_genotypes.index,
expected.loc[expected["discrepant_genotype"] == True].index,
)
pd.testing.assert_series_equal(sc.snps["pos"], expected["pos"])
pd.testing.assert_series_equal(sc.snps["genotype"],
expected["genotype"])