def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
qual = "1000"
filt = "PASS"
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;" \
"CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;" \
"MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
keys = "inheritance:DP"
values = "deNovo:50"
sex = "F"
# set up a CNV object
self.var = CNV(chrom,
pos,
snp_id,
ref,
alt,
qual,
filt,
info=info,
format=keys,
sample=values,
gender=sex)
def create_cnv(self, gender, inh, chrom, pos, cq=None):
""" create a default variant
"""
snp_id = "."
ref = "A"
alt = "<DEL>"
filt = "PASS"
if cq is None:
cq = "transcript_ablation"
# set up a SNV object, since SNV inherits VcfInfo
var = CNV(chrom, pos, snp_id, ref, alt, filt)
info = "CQ={};HGNC=TEST;HGNC_ALL=TEST;END=16000000;SVLEN=5000".format(
cq)
format_keys = "INHERITANCE:DP"
sample_values = inh + ":50"
var.add_info(info)
var.add_format(format_keys, sample_values)
var.set_gender(gender)
var.set_genotype()
return var
def create_cnv(self,
chrom,
info=None,
pos='15000000',
snp_id='.',
ref='A',
alt='<DUP>',
qual='1000',
filt='PASS',
**kwargs):
if info is None:
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;CNSOLIDATE;' \
'WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;MEANLR2=0.5;' \
'MADL2R=0.02;END=16000000;SVLEN=1000000"
keys = "inheritance:DP"
values = "deNovo:50"
return CNV(chrom,
pos,
snp_id,
ref,
alt,
qual,
filt,
info=info,
format=keys,
sample=values,
gender='male',
**kwargs)
def test_construct_variant(self):
""" test that construct_variant() works correctly
"""
# check that construct variant works for SNVs
line = ["1", "100", ".", "T", "G", "1000", "PASS", ".", "GT", "0/1"]
gender = "M"
test_var = SNV(*line[:6])
variant = self.vcf_loader.construct_variant(line, gender)
self.assertEqual(variant.get_key(), test_var.get_key())
# initally constructing a SNV shouldn't affect the format variable
self.assertEqual(variant.format, None)
# check that construct variant works for CNVs
line = [
"1", "100", ".", "T", "<DEL>", "1000", "PASS", "END=200", "GT",
"0/1"
]
gender = "M"
test_var = CNV(*line[:6])
test_var.add_info(line[7])
variant = self.vcf_loader.construct_variant(line, gender)
self.assertEqual(variant.get_key(), test_var.get_key())
self.assertNotEqual(variant.format, None)
def test_construct_variant(self):
""" test that construct_variant() works correctly
"""
# check that construct variant works for SNVs
line = ["1", "100", ".", "T", "G", "1000", "PASS", ".", "GT", "0/1"]
gender = "M"
test_var = SNV(*line, gender=gender)
variant = construct_variant(line, gender)
self.assertEqual(variant.get_key(), test_var.get_key())
self.assertEqual(variant.format, {'GT': '0/1'})
# check that construct variant works for CNVs
line = [
"1", "100", ".", "T", "<DEL>", "1000", "PASS", "END=200", "GT",
"0/1"
]
gender = "M"
test_var = CNV(*line, gender=gender)
variant = construct_variant(line, gender)
self.assertEqual(variant.get_key(), test_var.get_key())
self.assertEqual(variant.format, {'GT': '0/1'})
def test_include_variant(self):
""" check that include_variant() works correctly
"""
mnvs = {}
child_variants = False
gender = "M"
# make a child var which passes the filters
line = ["1", "100", ".", "T", "A", "1000", "PASS", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertTrue(self.vcf_loader.include_variant(line, child_variants, gender, mnvs))
# make a child var that fails the filters, which should return False
line = ["1", "100", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertFalse(self.vcf_loader.include_variant(line, child_variants, gender, mnvs))
# now check for parents variants
child_variants = True
# check a parents var, where we have a matching child var
self.vcf_loader.child_keys = set([("1", 100), ("X", 200)])
line = ["1", "100", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertTrue(self.vcf_loader.include_variant(line, child_variants, gender, mnvs))
# check a parents var, where we don't have a matching child var
line = ["1", "200", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertFalse(self.vcf_loader.include_variant(line, child_variants, gender, mnvs))
# and check parental CNVs
line = ["1", "100", ".", "T", "<DEL>", "1000", "PASS", "END=200", "GT", "0/1"]
gender = "M"
test_var = CNV(*line[:6])
test_var.add_info(line[7])
# in this function we look for overlap in CNVs. Set up a child CNV
# that the parents CNV must match.
self.assertTrue(self.vcf_loader.include_variant(line, child_variants, gender, mnvs))
# check that a parental CNV without any overlap to any childs CNVs,
# fails to pass
line = ["1", "300", ".", "T", "<DEL>", "1000", "PASS", "END=400", "GT", "0/1"]
gender = "M"
self.assertFalse(self.vcf_loader.include_variant(line, child_variants, gender, mnvs))
def test_get_parental_var_cnv(self):
''' check that get_parental_var() works correctly for CNVs
'''
sex = 'F'
var = create_cnv(sex, 'deNovo')
mom = Person('fam_id', 'mom', '0', '0', 'F', '1', '/PATH')
parental_vars = []
self.assertEqual(
get_parental_var(var, parental_vars, mom),
CNV(chrom="1",
position=150,
id=".",
ref="A",
alts="<REF>",
qual='1000',
filter="PASS",
info=str(var.info),
format='INHERITANCE',
sample='uncertain',
gender="female",
mnv_code=None))
# check that even if a CNV exist in the parent at a matching site, we
# still create a new CNV objectr for the parent
mother_var = create_cnv(sex, 'uncertain')
self.assertEqual(
get_parental_var(var, [mother_var], mom),
CNV(chrom="1",
position=150,
id=".",
ref="A",
alts="<REF>",
qual='1000',
filter="PASS",
info=str(var.info),
format='INHERITANCE',
sample='uncertain',
gender="female",
mnv_code=None))
def test_include_variant(self):
""" check that include_variant() works correctly
"""
child_variants = False
gender = "M"
# make a child var which passes the filters
line = ["1", "100", ".", "T", "A", "1000", "PASS", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertTrue(self.vcf_loader.include_variant(line, child_variants, gender))
# make a child var that fails the filters, which should return False
line = ["1", "100", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertFalse(self.vcf_loader.include_variant(line, child_variants, gender))
# now check for parents variants
child_variants = True
# check a parents var, where we have a matching child var
self.vcf_loader.child_keys = set([("1", 100), ("X", 200)])
line = ["1", "100", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertTrue(self.vcf_loader.include_variant(line, child_variants, gender))
# check a parents var, where we don't have a matching child var
line = ["1", "200", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertFalse(self.vcf_loader.include_variant(line, child_variants, gender))
# and check parental CNVs
line = ["1", "100", ".", "T", "<DEL>", "1000", "PASS", "END=200", "GT", "0/1"]
gender = "M"
test_var = CNV(*line[:6])
test_var.add_info(line[7])
# in this function we look for overlap in CNVs. Set up a child CNV
# that the parents CNV must match.
self.vcf_loader.cnv_matcher = MatchCNVs([test_var])
self.assertTrue(self.vcf_loader.include_variant(line, child_variants, gender))
# check that a parental CNV without any overlap to any childs CNVs,
# fails to pass
line = ["1", "300", ".", "T", "<DEL>", "1000", "PASS", "END=400", "GT", "0/1"]
gender = "M"
self.assertFalse(self.vcf_loader.include_variant(line, child_variants, gender))
def create_cnv(self, chrom):
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
# set up a SNV object, since SNV inherits VcfInfo
var = CNV(chrom, pos, snp_id, ref, alt, filt)
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
format_keys = "inheritance:DP"
sample_values = "deNovo:50"
var.add_info(info)
var.add_format(format_keys, sample_values)
var.set_gender("F")
var.set_genotype()
return var
def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
# set up a SNV object, since SNV inherits VcfInfo
self.var = CNV(chrom, pos, snp_id, ref, alt, filt)
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
format_keys = "inheritance:DP"
sample_values = "deNovo:50"
self.var.add_info(info)
self.var.add_format(format_keys, sample_values)
self.var.set_gender("F")
def create_cnv(self, gender, inh, chrom, pos):
""" create a default variant
"""
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
# set up a SNV object, since SNV inherits VcfInfo
var = CNV(chrom, pos, snp_id, ref, alt, filt)
info = "HGNC=TEST;HGNC_ALL=TEST;END=16000000;SVLEN=5000"
format_keys = "INHERITANCE:DP"
sample_values = inh + ":50"
var.add_info(info)
var.add_format(format_keys, sample_values)
var.set_gender(gender)
var.set_genotype()
return var
def test_get_parental_var_cnv_maternally_inherited(self):
''' test that we can construct a maternally inherited CNV
'''
sex = 'F'
mom = Person('fam_id', 'mom', '0', '0', 'F', '1', '/PATH')
# check that even if a CNV exist in the parent at a matching site, we
# still create a new CNV object for the parent
var = create_cnv(sex, 'maternal')
self.assertEqual(get_parental_var(var, [], mom),
CNV(chrom="1", position=150, id=".", ref="A",
alts="<DUP>", qual='1000',filter="PASS", info=str(var.info),
format='INHERITANCE', sample='uncertain', gender="female",
mnv_code=None))
def create_cnv(self, gender, inh, cifer, chrom, pos):
""" create a default variant
"""
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
# set up a SNV object, since SNV inherits VcfInfo
var = CNV(chrom, pos, snp_id, ref, alt, filt)
info = "HGNC=TEST;HGNC_ALL=TEST;END=16000000;SVLEN=5000;CNS=3"
format_keys = "CIFER:INHERITANCE:DP"
sample_values = cifer + ":" + inh + ":50"
var.add_info(info)
var.add_format(format_keys, sample_values)
var.set_gender(gender)
var.set_genotype()
return var
def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;" \
"CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;" \
"MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
keys = "inheritance:DP"
values = "deNovo:50"
sex = "F"
# set up a CNV object
self.var = CNV(chrom, pos, snp_id, ref, alt, filt, info=info,
format=keys, sample=values, gender=sex)
def test_include_variant(self):
""" check that include_variant() works correctly
"""
mnvs = {}
child_keys = None
gender = "M"
sum_x_l2r = {}
# make a child var which passes the filters
line = ["1", "100", ".", "T", "A", "1000", "PASS", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertTrue(include_variant(line, child_keys, gender, mnvs, sum_x_l2r))
# make a child var that fails the filters, which should return False
line = ["1", "100", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertFalse(include_variant(line, child_keys, gender, mnvs, sum_x_l2r))
# now check for parents variants
# check a parents var, where we have a matching child var
child_keys = set([("1", 100), ("X", 200)])
line = ["1", "100", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertTrue(include_variant(line, child_keys, gender, mnvs, sum_x_l2r))
# check a parents var, where we don't have a matching child var
line = ["1", "200", ".", "T", "A", "1000", "FAIL", "CQ=missense_variant;HGNC=ATRX", "GT", "0/1"]
self.assertFalse(include_variant(line, child_keys, gender, mnvs, sum_x_l2r))
# and check parental CNVs
line = ["1", "100", ".", "T", "<DEL>", "1000", "PASS", "END=200", "GT", "0/1"]
gender = "M"
test_var = CNV(*line)
# in this function we look for overlap in CNVs. Set up a child CNV
# that the parents CNV must match.
self.assertTrue(include_variant(line, child_keys, gender, mnvs, sum_x_l2r))
# check that a parental CNV without any overlap to any childs CNVs,
# fails to pass
line = ["1", "300", ".", "T", "<DEL>", "1000", "PASS", "END=400", "GT", "0/1"]
gender = "M"
self.assertFalse(include_variant(line, child_keys, gender, mnvs, sum_x_l2r))
def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
# set up a SNV object, since SNV inherits VcfInfo
cnv = CNV(chrom, pos, snp_id, ref, alt, filt)
self.var = ExomeCNV(cnv)
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;CONVEX;RC50INTERNALFREQ=0.005;COMMONFORWARDS=0.000;MEANLR2=0.5;MADL2R=0.02"
format_keys = "inheritance:DP"
sample_values = "deNovo:50"
self.var.cnv.add_info(info)
self.var.cnv.add_format(format_keys, sample_values)
self.var.cnv.set_gender("F")
def get_parental_var(self, var, parental_vars, gender, matcher):
""" get the corresponding parental variant to a childs variant, or
create a default variant with reference genotype.
Args:
var: childs var, as Variant object
parental_vars: list of parental variants
gender: gender of the parent
matcher: cnv matcher for parent
Returns:
returns a Variant object, matched to the proband's variant
"""
key = var.get_key()
# if the variant is a CNV, the corresponding variant might not match
# the start site, so we look a variant that overlaps
if isinstance(var, CNV) and matcher.has_match(var):
key = matcher.get_overlap_key(key)
for parental in parental_vars:
if key == parental.get_key():
return parental
# if the childs variant does not exist in the parents VCF, then we
# create a default variant for the parent
if isinstance(var, CNV):
parental = CNV(var.chrom, var.position, var.variant_id,
var.ref_allele, var.alt_allele, var.filter)
else:
parental = SNV(var.chrom, var.position, var.variant_id,
var.ref_allele, var.alt_allele, var.filter)
parental.set_gender(gender)
parental.set_default_genotype()
return parental
def get_parental_var(self, var, parental_vars, gender, matcher):
""" get the corresponding parental variant to a childs variant, or
create a default variant with reference genotype.
Args:
var: childs var, as Variant object
parental_vars: list of parental variants
gender: gender of the parent
matcher: cnv matcher for parent
Returns:
returns a Variant object, matched to the proband's variant
"""
key = var.get_key()
# if the variant is a CNV, the corresponding variant might not match
# the start site, so we look a variant that overlaps
if isinstance(var, CNV) and matcher.has_match(var):
key = matcher.get_overlap_key(key)
for parental in parental_vars:
if key == parental.get_key():
return parental
# if the childs variant does not exist in the parents VCF, then we
# create a default variant for the parent
if isinstance(var, CNV):
parental = CNV(var.chrom, var.position, var.variant_id, var.ref_allele, var.alt_allele, var.filter)
else:
parental = SNV(var.chrom, var.position, var.variant_id, var.ref_allele, var.alt_allele, var.filter)
parental.set_gender(gender)
parental.set_default_genotype()
return parental
def construct_variant(self, line, gender):
""" constructs a Variant object for a VCF line, specific to the variant type
Args:
line: list of elements of a single sample VCF line:
[chrom, position, snp_id, ref_allele, alt_allele, quality,
filter_value, info, format_keys, format_values]
gender: gender of the individual to whom the variant line belongs
(eg "1" or "M" for male, "2", or "F" for female).
Returns:
returns a Variant object
"""
# CNVs are found by their alt_allele values, as either <DUP>, or <DEL>
if line[4] == "<DUP>" or line[4] == "<DEL>":
var = CNV(line[0], line[1], line[2], line[3], line[4], line[6])
var.add_info(line[7])
# CNVs require the format values for filtering
var.set_gender(gender)
var.add_format(line[8], line[9])
if self.known_genes is not None:
var.fix_gene_IDs()
else:
var = SNV(line[0], line[1], line[2], line[3], line[4], line[6])
var.add_info(line[7])
return var
class TestVariantCnvPy(unittest.TestCase):
"""
"""
def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
# set up a SNV object, since SNV inherits VcfInfo
self.var = CNV(chrom, pos, snp_id, ref, alt, filt)
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
format_keys = "inheritance:DP"
sample_values = "deNovo:50"
self.var.add_info(info)
self.var.add_format(format_keys, sample_values)
self.var.set_gender("F")
def test_set_genotype(self):
""" test that set_genotype() operates correctly
"""
# check that DUPs are set correctly
self.var.alt_allele = "<DUP>"
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DUP")
# check that DELs are set correctly
self.var.alt_allele = "<DEL>"
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DEL")
# check that other genotypes raise an error
self.var.alt_allele = "G"
with self.assertRaises(ValueError):
self.var.set_genotype()
# and check that we raise an error for female Y chrom CNVs
self.var.chrom = "Y"
self.var.set_gender("F")
with self.assertRaises(ValueError):
self.var.set_genotype()
def test_set_genotype_pseudoautosomal(self):
""" test that set_genotype() works correctly in pseudoautosomal regions
"""
pseudoautosomal_region_start = 60002
pseudoautosomal_region_end = 2699520
# set a CNV that lies within a pseudoautosomal region
self.var.chrom = "X"
self.var.position = pseudoautosomal_region_start + 1000
self.var.info["END"] = pseudoautosomal_region_end - 1000
self.var.set_gender("F")
self.var.alt_allele = "<DUP>"
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DUP")
self.assertEqual(self.var.get_inheritance_type(), "autosomal")
def test_get_range(self):
""" test that get_range() operates correctly
"""
# check that range is set correctly under normal function
self.var.position = 1000
self.var.info["END"] = "2000"
self.assertEqual(self.var.get_range(), (1000, 2000))
# check that range is set correctly when no info available
self.var.info = {}
self.assertEqual(self.var.get_range(), (1000, 11000))
def test_fix_gene_IDs(self):
""" test that fix_gene_IDs() works correctly
"""
self.var.known_genes = {"TEST": {"start": 1000, "end": 2000, "chrom": "5"}}
# make a CNV that will overlap with the known gene set
self.var.genes = ["TEST"]
self.var.position = 1000
self.var.info["END"] = "1500"
# check that fixing gene names does not alter anything for a CNV in a
# single known gene
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, ["TEST"])
# check that fixing gene names does not alter names not in the gene dict
self.var.genes = ["TEST", "TEST2"]
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, ["TEST", "TEST2"])
# check that fixing gene names drop name of genes where the name is in
# the known genes dict, and the CNV and gene do not overlap
self.var.position = 900
self.var.info["END"] = "950"
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, [".", "TEST2"])
# check that when we do not have any known genes, the gene names are
# unaltered
self.var.genes = ["TEST", "TEST2"]
self.var.known_genes = None
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, ["TEST", "TEST2"])
def test_set_gene_from_info_cnv(self):
""" test that set_add_gene_from_info() works correctly
"""
# make sure the known genes are None, otherwise sometimes the values
# from test_variant_info.py unit tests can bleed through. I'm not sure
# why!
self.var.known_genes = None
# check that HGNC takes precedence
self.var.info["HGNC"] = "A"
self.var.info["HGNC_ALL"] = "B"
self.var.set_gene_from_info()
self.assertEqual(self.var.genes, ["A"])
# check that HGNC is used in the absence of HGNC_ALL
del self.var.info["HGNC"]
self.var.set_gene_from_info()
self.assertEqual(self.var.genes, ["B"])
# check that when HGNC and HGNC_ALL are undefined, we can still include
# CNVs overlapping genes through NUMBERGENES > 0.
del self.var.info["HGNC_ALL"]
# first test for NUMBERGENES = 0
self.var.info["NUMBERGENES"] = 0
self.var.set_gene_from_info()
self.assertIsNone(self.var.genes)
# and then make sure we are correct for NUMBERGENES > 0
self.var.info["NUMBERGENES"] = 1
self.var.set_gene_from_info()
self.assertEqual(self.var.genes, ["."])
# finally check for no HGNC, HGNC_ALL, or NUMBERGENES
del self.var.info["NUMBERGENES"]
self.var.set_gene_from_info()
self.assertEqual(self.var.genes, "1:15000000")
def test_get_genes(self):
""" test that get_genes() works correctly
"""
self.var.genes = None
self.assertEqual(self.var.get_genes(), [])
self.var.genes = ["TEST"]
self.assertEqual(self.var.get_genes(), ["TEST"])
self.var.genes = ["TEST1", "TEST2"]
self.assertEqual(self.var.get_genes(), ["TEST1", "TEST2"])
self.var.genes = ["."]
self.assertEqual(self.var.get_genes(), ["."])
def test_fails_y_chrom_female(self):
""" test that passes_filters() works correctly for female Y chrom CNVs
"""
self.var.chrom = "Y"
self.var.set_gender("F")
self.assertFalse(self.var.passes_filters())
def construct_variant(self, line, gender):
""" constructs a Variant object for a VCF line, specific to the variant type
Args:
line: list of elements of a single sample VCF line:
[chrom, position, snp_id, ref_allele, alt_allele, quality,
filter_value, info, format_keys, format_values]
gender: gender of the individual to whom the variant line belongs
(eg "1" or "M" for male, "2", or "F" for female).
Returns:
returns a Variant object
"""
# CNVs are found by their alt_allele values, as either <DUP>, or <DEL>
if line[4] == "<DUP>" or line[4] == "<DEL>":
var = CNV(line[0], line[1], line[2], line[3], line[4], line[6])
var.add_info(line[7])
# CNVs require the format values for filtering
var.set_gender(gender)
var.add_format(line[8], line[9])
if self.known_genes is not None:
var.fix_gene_IDs()
else:
var = SNV(line[0], line[1], line[2], line[3], line[4], line[6])
var.add_info(line[7])
return var
class TestVariantCnvPy(unittest.TestCase):
""" unit testing of the CNV class
"""
def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
qual = "1000"
filt = "PASS"
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;" \
"CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;" \
"MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
keys = "inheritance:DP"
values = "deNovo:50"
sex = "F"
# set up a CNV object
self.var = CNV(chrom,
pos,
snp_id,
ref,
alt,
qual,
filt,
info=info,
format=keys,
sample=values,
gender=sex)
def test_set_genotype(self):
""" test that set_genotype() operates correctly
"""
# check that DUPs are set correctly
self.var.alt_alleles = ["<DUP>"]
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DUP")
# check that DELs are set correctly
self.var.alt_alleles = ["<DEL>"]
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DEL")
# check that other genotypes raise an error
self.var.alt_alleles = ["G"]
with self.assertRaises(ValueError):
self.var.set_genotype()
# and check that we raise an error for female Y chrom CNVs
self.var.chrom = "Y"
self.var._set_gender("F")
with self.assertRaises(ValueError):
self.var.set_genotype()
def test_set_genotype_pseudoautosomal(self):
""" test that set_genotype() works correctly in pseudoautosomal regions
"""
pseudoautosomal_region_start = 60002
pseudoautosomal_region_end = 2699520
# set a CNV that lies within a pseudoautosomal region
self.var.chrom = "X"
self.var.position = pseudoautosomal_region_start + 1000
self.var.info["END"] = pseudoautosomal_region_end - 1000
self.var._set_gender("F")
self.var.alt_alleles = ["<DUP>"]
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DUP")
self.assertEqual(self.var.get_inheritance_type(), "autosomal")
def test_get_range(self):
""" test that get_range() operates correctly
"""
# check that range is set correctly under normal function
self.var.position = 1000
self.var.info["END"] = "2000"
self.assertEqual(self.var.get_range(), (1000, 2000))
# check that range is set correctly when no info available
self.var.info = {}
self.assertEqual(self.var.get_range(), (1000, 11000))
def test_fix_gene_IDs(self):
""" test that fix_gene_IDs() works correctly
"""
self.var.known_genes = {
"TEST": {
"start": 1000,
"end": 2000,
"chrom": "5"
}
}
# make a CNV that will overlap with the known gene set
self.var.info.symbols = [Symbols(info={'HGNC_ID': 'TEST'}, idx=0)]
self.var.position = 1000
self.var.info["END"] = "1500"
# check that fixing gene names does not alter anything for a CNV in a
# single known gene
self.var.fix_gene_IDs()
self.assertEqual(self.var.info.get_genes(), [['TEST']])
# check that fixing gene names does not alter names not in the gene dict
self.var.info.symbols = [
Symbols(info={'HGNC_ID': 'TEST|TEST2'}, idx=0)
]
self.var.fix_gene_IDs()
self.assertEqual(self.var.info.get_genes(), [['TEST', 'TEST2']])
# check that fixing gene names drop name of genes where the name is in
# the known genes dict, and the CNV and gene do not overlap
self.var.position = 900
self.var.info["END"] = "950"
self.var.fix_gene_IDs()
self.assertEqual(self.var.info.get_genes(), [[None, 'TEST2']])
# check that when we do not have any known genes, the gene names are
# unaltered
self.var.info.symbols = [
Symbols(info={'HGNC_ID': 'TEST|TEST2'}, idx=0)
]
self.var.known_genes = None
self.var.fix_gene_IDs()
self.assertEqual(self.var.info.get_genes(), [['TEST', 'TEST2']])
def test_set_gene_from_info_cnv(self):
""" test that set_add_gene_from_info() works correctly
"""
# make sure the known genes are None, otherwise sometimes the values
# from test_variant_info.py unit tests can bleed through. I'm not sure
# why!
self.var.known_genes = None
# check that HGNC takes precedence
self.var.info["HGNC"] = "A"
genes = self.var.info.parse_gene_symbols(self.var.alt_alleles, [])
self.assertEqual(genes, [Symbols(info={'HGNC': 'A'}, idx=0)])
# check that HGNC_ALL doesn't affect anything
self.var.info["HGNC_ALL"] = "B"
del self.var.info["HGNC"]
genes = self.var.info.parse_gene_symbols(self.var.alt_alleles, [])
self.assertEqual(genes, [Symbols(info={}, idx=0)])
def test_get_genes(self):
""" test that get_genes() works correctly
"""
self.var.info.symbols = [Symbols(info={}, idx=0)]
self.assertEqual(self.var.info.get_genes(), [[]])
self.var.info.symbols = [Symbols(info={'HGNC': 'TEST'}, idx=0)]
self.assertEqual(self.var.info.get_genes(), [["TEST"]])
self.var.info.symbols = [Symbols(info={'HGNC': 'TEST1|TEST2'}, idx=0)]
self.assertEqual(self.var.info.get_genes(), [["TEST1", "TEST2"]])
self.var.info.symbols = [Symbols(info={'HGNC': '.'}, idx=0)]
self.assertEqual(self.var.info.get_genes(), [[None]])
def test_fails_y_chrom_female(self):
""" test that passes_filters() works correctly for female Y chrom CNVs
"""
self.var.chrom = "Y"
self.var._set_gender("F")
self.assertFalse(self.var.passes_filters())
class TestVariantCnvPy(unittest.TestCase):
""" unit testing of the CNV class
"""
def setUp(self):
""" define a default VcfInfo object
"""
chrom = "1"
pos = "15000000"
snp_id = "."
ref = "A"
alt = "<DUP>"
filt = "PASS"
info = "HGNC=TEST;HGNC_ALL=TEST,OR5A1;CQ=missense_variant;" \
"CNSOLIDATE;WSCORE=0.5;CALLP=0.000;COMMONFORWARDS=0.000;" \
"MEANLR2=0.5;MADL2R=0.02;END=16000000;SVLEN=1000000"
keys = "inheritance:DP"
values = "deNovo:50"
sex = "F"
# set up a CNV object
self.var = CNV(chrom, pos, snp_id, ref, alt, filt, info=info,
format=keys, sample=values, gender=sex)
def test_set_genotype(self):
""" test that set_genotype() operates correctly
"""
# check that DUPs are set correctly
self.var.alt_alleles = ["<DUP>"]
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DUP")
# check that DELs are set correctly
self.var.alt_alleles = ["<DEL>"]
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DEL")
# check that other genotypes raise an error
self.var.alt_alleles = ["G"]
with self.assertRaises(ValueError):
self.var.set_genotype()
# and check that we raise an error for female Y chrom CNVs
self.var.chrom = "Y"
self.var._set_gender("F")
with self.assertRaises(ValueError):
self.var.set_genotype()
def test_set_genotype_pseudoautosomal(self):
""" test that set_genotype() works correctly in pseudoautosomal regions
"""
pseudoautosomal_region_start = 60002
pseudoautosomal_region_end = 2699520
# set a CNV that lies within a pseudoautosomal region
self.var.chrom = "X"
self.var.position = pseudoautosomal_region_start + 1000
self.var.info["END"] = pseudoautosomal_region_end - 1000
self.var._set_gender("F")
self.var.alt_alleles = ["<DUP>"]
self.var.set_genotype()
self.assertEqual(self.var.genotype, "DUP")
self.assertEqual(self.var.get_inheritance_type(), "autosomal")
def test_get_range(self):
""" test that get_range() operates correctly
"""
# check that range is set correctly under normal function
self.var.position = 1000
self.var.info["END"] = "2000"
self.assertEqual(self.var.get_range(), (1000, 2000))
# check that range is set correctly when no info available
self.var.info = {}
self.assertEqual(self.var.get_range(), (1000, 11000))
def test_fix_gene_IDs(self):
""" test that fix_gene_IDs() works correctly
"""
self.var.known_genes = {"TEST": {"start": 1000, "end": 2000, "chrom": "5"}}
# make a CNV that will overlap with the known gene set
self.var.genes = [["TEST"]]
self.var.position = 1000
self.var.info["END"] = "1500"
# check that fixing gene names does not alter anything for a CNV in a
# single known gene
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, [["TEST"]])
# check that fixing gene names does not alter names not in the gene dict
self.var.genes = [["TEST", "TEST2"]]
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, [["TEST", "TEST2"]])
# check that fixing gene names drop name of genes where the name is in
# the known genes dict, and the CNV and gene do not overlap
self.var.position = 900
self.var.info["END"] = "950"
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, [[".", "TEST2"]])
# check that when we do not have any known genes, the gene names are
# unaltered
self.var.genes = [["TEST", "TEST2"]]
self.var.known_genes = None
self.var.fix_gene_IDs()
self.assertEqual(self.var.genes, [["TEST", "TEST2"]])
def test_set_gene_from_info_cnv(self):
""" test that set_add_gene_from_info() works correctly
"""
# make sure the known genes are None, otherwise sometimes the values
# from test_variant_info.py unit tests can bleed through. I'm not sure
# why!
self.var.known_genes = None
# check that HGNC takes precedence
self.var.info["HGNC"] = "A"
self.var.info["HGNC_ALL"] = "B"
genes = self.var.get_gene_from_info(self.var.info, self.var.alt_alleles, [])
self.assertEqual(genes, [["A"]])
# check that HGNC is used in the absence of HGNC_ALL
del self.var.info["HGNC"]
genes = self.var.get_gene_from_info(self.var.info, self.var.alt_alleles, [])
self.assertEqual(genes, [["B"]])
# check that when HGNC and HGNC_ALL are undefined, we can still include
# CNVs overlapping genes through NUMBERGENES > 0.
del self.var.info["HGNC_ALL"]
# first test for NUMBERGENES = 0
self.var.info["NUMBERGENES"] = 0
genes = self.var.get_gene_from_info(self.var.info, self.var.alt_alleles, [])
self.assertIsNone(genes)
# and then make sure we are correct for NUMBERGENES > 0
self.var.info["NUMBERGENES"] = 1
genes = self.var.get_gene_from_info(self.var.info, self.var.alt_alleles, [])
self.assertEqual(genes, [["."]])
# finally check for no HGNC, HGNC_ALL, or NUMBERGENES
del self.var.info["NUMBERGENES"]
genes = self.var.get_gene_from_info(self.var.info, self.var.alt_alleles, [])
self.assertEqual(genes, [["1:15000000"]])
def test_get_genes(self):
""" test that get_genes() works correctly
"""
self.var.genes = None
self.assertEqual(self.var.get_genes(), [])
self.var.genes = ["TEST"]
self.assertEqual(self.var.get_genes(), ["TEST"])
self.var.genes = ["TEST1", "TEST2"]
self.assertEqual(self.var.get_genes(), ["TEST1", "TEST2"])
self.var.genes = ["."]
self.assertEqual(self.var.get_genes(), ["."])
def test_fails_y_chrom_female(self):
""" test that passes_filters() works correctly for female Y chrom CNVs
"""
self.var.chrom = "Y"
self.var._set_gender("F")
self.assertFalse(self.var.passes_filters())