def test_sort_variants(self):
sorted_variants = [
Variant(reference_name='a', start=20, end=22),
Variant(reference_name='a', start=20, end=22, quality=20),
Variant(reference_name='b', start=20, end=22),
Variant(reference_name='b', start=21, end=22),
Variant(reference_name='b', start=21, end=23)]
for permutation in permutations(sorted_variants):
self.assertEqual(sorted(permutation), sorted_variants)
def test_info_numbers_and_types(self):
info_headers = [
'##INFO=<ID=HA,Number=A,Type=String,Description="StringInfo_A">\n',
'##INFO=<ID=HG,Number=G,Type=Integer,Description="IntInfo_G">\n',
'##INFO=<ID=HR,Number=R,Type=Character,Description="ChrInfo_R">\n',
'##INFO=<ID=HF,Number=0,Type=Flag,Description="FlagInfo">\n',
'##INFO=<ID=HU,Number=.,Type=Float,Description="FloatInfo_variable">\n'
]
record_lines = [
'19 2 . A T,C . . HA=a1,a2;HG=1,2,3;HR=a,b,c;HF;HU=0.1 GT 1/0 0/1\n',
'19 124 . A T . . HG=3,4,5;HR=d,e;HU=1.1,1.2 GT 0/0 0/1'
]
variant_1 = Variant(reference_name='19',
start=1,
end=2,
reference_bases='A',
alternate_bases=['T', 'C'],
info={
'HA':
VariantInfo(data=['a1', 'a2'],
field_count='A'),
'HG':
VariantInfo(data=[1, 2, 3], field_count='G'),
'HR':
VariantInfo(data=['a', 'b', 'c'],
field_count='R'),
'HF':
VariantInfo(data=True, field_count='0'),
'HU':
VariantInfo(data=[0.1], field_count=None)
})
variant_1.calls.append(VariantCall(name='Sample1', genotype=[1, 0]))
variant_1.calls.append(VariantCall(name='Sample2', genotype=[0, 1]))
variant_2 = Variant(reference_name='19',
start=123,
end=124,
reference_bases='A',
alternate_bases=['T'],
info={
'HG':
VariantInfo(data=[3, 4, 5], field_count='G'),
'HR':
VariantInfo(data=['d', 'e'], field_count='R'),
'HU':
VariantInfo(data=[1.1, 1.2], field_count=None)
})
variant_2.calls.append(VariantCall(name='Sample1', genotype=[0, 0]))
variant_2.calls.append(VariantCall(name='Sample2', genotype=[0, 1]))
read_data = self._create_temp_file_and_read_records(
info_headers + _SAMPLE_HEADER_LINES[1:] + record_lines)
self.assertEqual(2, len(read_data))
self._assert_variants_equal([variant_1, variant_2], read_data)
def test_use_of_representative_header_two_files(self):
# Info field `HU` is defined as Float in file header while data is String.
# This results in parser failure. We test if parser completes successfully
# when a representative headers with String definition for field `HU` is
# given.
file_content_1 = [
'##INFO=<ID=HU,Number=.,Type=Float,Description="Info">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
'#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT Sample1\r\n',
'9 2 . A T . . HU=a,b GT 0/0'
]
file_content_2 = [
'##INFO=<ID=HU,Number=.,Type=Float,Description="Info">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
'#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT Sample2\r\n',
'19 2 . A T . . HU=a,b GT 0/1\n',
]
representative_header_lines = [
'##INFO=<ID=HU,Number=.,Type=String,Description="Info">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
]
variant_1 = Variant(
reference_name='9',
start=1,
end=2,
reference_bases='A',
alternate_bases=['T'],
info={'HU': VariantInfo(data=['a', 'b'], field_count=None)})
variant_1.calls.append(VariantCall(name='Sample1', genotype=[0, 0]))
variant_2 = Variant(
reference_name='19',
start=1,
end=2,
reference_bases='A',
alternate_bases=['T'],
info={'HU': VariantInfo(data=['a', 'b'], field_count=None)})
variant_2.calls.append(VariantCall(name='Sample2', genotype=[0, 1]))
read_data_1 = self._create_temp_file_and_read_records(
file_content_1, representative_header_lines)
self.assertEqual(1, len(read_data_1))
self._assert_variants_equal([variant_1], read_data_1)
read_data_2 = self._create_temp_file_and_read_records(
file_content_2, representative_header_lines)
self.assertEqual(1, len(read_data_2))
self._assert_variants_equal([variant_2], read_data_2)
def test_use_of_representative_header_two_files(self):
# Info field `HU` is defined as Float in file header while data is String.
# This results in parser failure. We test if parser completes successfully
# when a representative headers with String definition for field `HU` is
# given.
file_content_1 = [
'##INFO=<ID=HU,Number=.,Type=Float,Descri\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
'#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tSample1\r\n',
'9\t2\t.\tA\tT\t.\t.\tHU=a,b\tGT\t0/0'
]
file_content_2 = [
'##INFO=<ID=HU,Number=.,Type=Float,Descri\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
'#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tSample2\r\n',
'19\t2\t.\tA\tT\t.\t.\tHU=a,b\tGT\t0/1\n',
]
representative_header_lines = [
'##INFO=<ID=HU,Number=.,Type=String,Description="Info">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
]
variant_1 = Variant(reference_name='9',
start=1,
end=2,
reference_bases='A',
alternate_bases=['T'],
info={'HU': ['a', 'b']})
variant_1.calls.append(
VariantCall(sample_id=hash_name('Sample1'), genotype=[0, 0]))
variant_2 = Variant(reference_name='19',
start=1,
end=2,
reference_bases='A',
alternate_bases=['T'],
info={'HU': ['a', 'b']})
variant_2.calls.append(
VariantCall(sample_id=hash_name('Sample2'), genotype=[0, 1]))
read_data_1 = self._create_temp_file_and_read_records(
file_content_1, representative_header_lines)
self.assertEqual(1, len(read_data_1))
self._assert_variants_equal([variant_1], read_data_1)
read_data_2 = self._create_temp_file_and_read_records(
file_content_2, representative_header_lines)
self.assertEqual(1, len(read_data_2))
self._assert_variants_equal([variant_2], read_data_2)
def test_empty_sample_calls(self):
coder = self._get_coder()
variant = Variant()
variant.calls.append(
VariantCall(name='Sample2', genotype=-1))
expected = '. . . . . . . . GT .\n'
self._assert_variant_lines_equal(coder.encode(variant), expected)
def test_triploid_genotype(self):
coder = self._get_coder()
variant = Variant()
variant.calls.append(VariantCall(name='Sample', genotype=[1, 0, 1]))
expected = '. . . . . . . . GT 1/0/1\n'
self._assert_variant_lines_equal(coder.encode(variant), expected)
def test_format_numbers(self):
format_headers = [
'##FORMAT=<ID=FU,Number=.,Type=String,Description="Format_variable">\n',
'##FORMAT=<ID=F1,Number=1,Type=Integer,Description="Format_1">\n',
'##FORMAT=<ID=F2,Number=2,Type=Character,Description="Format_2">\n',
'##FORMAT=<ID=AO,Number=A,Type=Integer,Description="Format_3">\n',
'##FORMAT=<ID=AD,Number=G,Type=Integer,Description="Format_4">\n',]
record_lines = [
('19 2 . A T,C . . . '
'GT:FU:F1:F2:AO:AD 1/0:a1:3:a,b:1:3,4 '
'0/1:a2,a3:4:b,c:1,2:3')]
expected_variant = Variant(
reference_name='19', start=1, end=2, reference_bases='A',
alternate_bases=['T', 'C'])
expected_variant.calls.append(VariantCall(
sample_id=hash_name('Sample1'),
name='Sample1',
genotype=[1, 0],
info={'FU': ['a1'], 'F1': 3, 'F2': ['a', 'b'], 'AO': [1],
'AD': [3, 4]}))
expected_variant.calls.append(VariantCall(
sample_id=hash_name('Sample2'),
name='Sample2',
genotype=[0, 1],
info={'FU': ['a2', 'a3'], 'F1': 4, 'F2': ['b', 'c'], 'AO': [1, 2],
'AD':[3]}))
read_data = self._create_temp_file_and_read_records(
format_headers + _SAMPLE_HEADER_LINES[1:] + record_lines)
self.assertEqual(1, len(read_data))
self.assertEqual(expected_variant, read_data[0])
def test_use_of_representative_header(self):
# Info field `HU` is defined as Float in file header while data is String.
# This results in parser failure. We test if parser completes successfully
# when a representative headers with String definition for field `HU` is
# given.
file_content = [
'##INFO=<ID=HU,Number=.,Type=Float,Description="Info">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',
'#CHROM POS ID REF ALT QUAL FILTER INFO FORMAT Sample1 Sample2\r\n',
'19 2 . A T . . HU=a,b GT 0/0 0/1\n',]
representative_header_lines = [
'##INFO=<ID=HU,Number=.,Type=String,Description="Info">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\r\n',]
variant = Variant(
reference_name='19', start=1, end=2, reference_bases='A',
alternate_bases=['T'], info={'HU': ['a', 'b']})
variant.calls.append(VariantCall(sample_id=hash_name('Sample1'),
name='Sample1',
genotype=[0, 0]))
variant.calls.append(VariantCall(sample_id=hash_name('Sample2'),
name='Sample2',
genotype=[0, 1]))
# `file_headers` is used.
read_data = self._create_temp_file_and_read_records(file_content)
# Pysam expects Float value for HU, and returns Nones when list is given.
self.assertEqual([None, None], read_data[0].info['HU'])
# `representative_header` is used.
read_data = self._create_temp_file_and_read_records(
file_content, representative_header_lines)
self.assertEqual(1, len(read_data))
self._assert_variants_equal([variant], read_data)
def test_end_info_key_unknown_number_invalid(self):
end_info_header_line = (
'##INFO=<ID=END,Number=.,Type=Integer,Description="End of record.">\n')
# PySam should only take first END field.
variant = Variant(
reference_name='19', start=122, end=150, reference_bases='A',
alternate_bases=['T'])
variant.calls.append(VariantCall(sample_id=hash_name('Sample1'),
name='Sample1',
genotype=[1, 0]))
variant.calls.append(VariantCall(sample_id=hash_name('Sample2'),
name='Sample2',
genotype=[0, 1]))
read_data = self._create_temp_file_and_read_records(
[end_info_header_line] + _SAMPLE_HEADER_LINES[1:] +
['19 123 . A T . . END=150,160 GT 1/0 0/1\n'])
self.assertEqual(1, len(read_data))
self._assert_variants_equal([variant], read_data)
# END should be rounded down.
read_data = self._create_temp_file_and_read_records(
[end_info_header_line] + _SAMPLE_HEADER_LINES[1:] +
['19 123 . A T . . END=150.9 GT 1/0 0/1\n'])
self.assertEqual(1, len(read_data))
self._assert_variants_equal([variant], read_data)
# END should not be a string.
with self.assertRaises(ValueError):
self._create_temp_file_and_read_records(
[end_info_header_line] + _SAMPLE_HEADER_LINES[1:] +
['19 123 . A T . . END=text GT 1/0 0/1\n'])
def test_missing_genotype(self):
coder = self._get_coder()
variant = Variant()
variant.calls.append(VariantCall(
name='Sample', genotype=[1, vcfio.MISSING_GENOTYPE_VALUE]))
expected = '. . . . . . . . GT 1/.\n'
self._assert_variant_lines_equal(coder.encode(variant), expected)
def test_info_list(self):
coder = self._get_coder()
variant = Variant()
variant.calls.append(VariantCall(
name='Sample', genotype=[0, 1], info={'LI': [1, None, 3]}))
expected = '. . . . . . . . GT:LI 0/1:1,.,3\n'
self._assert_variant_lines_equal(coder.encode(variant), expected)
def test_end_info_key(self):
end_info_header_line = (
'##INFO=<ID=END,Number=1,Type=Integer,Description="End of record.">\n')
record_lines = ['19 123 . A . . . END=1111 GT 1/0 0/1\n',
'19 123 . A . . . . GT 0/1 1/1\n']
variant_1 = Variant(
reference_name='19', start=122, end=1111, reference_bases='A')
variant_1.calls.append(VariantCall(name='Sample1', genotype=[1, 0]))
variant_1.calls.append(VariantCall(name='Sample2', genotype=[0, 1]))
variant_2 = Variant(
reference_name='19', start=122, end=123, reference_bases='A')
variant_2.calls.append(VariantCall(name='Sample1', genotype=[0, 1]))
variant_2.calls.append(VariantCall(name='Sample2', genotype=[1, 1]))
read_data = self._create_temp_file_and_read_records(
[end_info_header_line] + _SAMPLE_HEADER_LINES[1:] + record_lines)
self.assertEqual(2, len(read_data))
self._assert_variants_equal([variant_1, variant_2], read_data)
def test_variant_equality(self):
base_variant = Variant(reference_name='a',
start=20,
end=22,
reference_bases='a',
alternate_bases=['g', 't'],
names=['variant'],
quality=9,
filters=['q10'],
info={'key': 'value'},
calls=[VariantCall(genotype=[0, 0])])
equal_variant = Variant(reference_name='a',
start=20,
end=22,
reference_bases='a',
alternate_bases=['g', 't'],
names=['variant'],
quality=9,
filters=['q10'],
info={'key': 'value'},
calls=[VariantCall(genotype=[0, 0])])
different_calls = Variant(reference_name='a',
start=20,
end=22,
reference_bases='a',
alternate_bases=['g', 't'],
names=['variant'],
quality=9,
filters=['q10'],
info={'key': 'value'},
calls=[VariantCall(genotype=[1, 0])])
missing_field = Variant(reference_name='a',
start=20,
end=22,
reference_bases='a',
alternate_bases=['g', 't'],
names=['variant'],
quality=9,
filters=['q10'],
info={'key': 'value'})
self.assertEqual(base_variant, equal_variant)
self.assertNotEqual(base_variant, different_calls)
self.assertNotEqual(base_variant, missing_field)
def test_no_samples(self):
header_line = '#CHROM POS ID REF ALT QUAL FILTER INFO\n'
record_line = '19 123 . G A . PASS AF=0.2'
expected_variant = Variant(
reference_name='19', start=122, end=123, reference_bases='G',
alternate_bases=['A'], filters=['PASS'], info={'AF': [0.2]})
read_data = self._create_temp_file_and_read_records(
_SAMPLE_HEADER_LINES[:-1] + [header_line, record_line])
self.assertEqual(1, len(read_data))
self.assertEqual(expected_variant, read_data[0])
def test_no_info(self):
record_line = 'chr19 123 . . . . . . GT . .'
expected_variant = Variant(reference_name='chr19', start=122, end=123)
expected_variant.calls.append(
VariantCall(name='Sample1', genotype=[vcfio.MISSING_GENOTYPE_VALUE]))
expected_variant.calls.append(
VariantCall(name='Sample2', genotype=[vcfio.MISSING_GENOTYPE_VALUE]))
read_data = self._create_temp_file_and_read_records(
_SAMPLE_HEADER_LINES + [record_line])
self.assertEqual(1, len(read_data))
self.assertEqual(expected_variant, read_data[0])
def test_missing_info_key(self):
coder = self._get_coder()
variant = Variant()
variant.calls.append(VariantCall(
name='Sample1', genotype=[0, 1], info={'GQ': 10, 'AF': 20}))
variant.calls.append(VariantCall(
name='Sample2', genotype=[0, 1], info={'AF': 20}))
expected = ('. . . . . . . . GT:AF:GQ 0/1:20:10 '
'0/1:20:.\n')
self._assert_variant_lines_equal(coder.encode(variant), expected)
def get_merged_variants(self, variants, unused_key=None):
# type: (List[Variant], str) -> List[Variant]
if not variants:
return []
merged_variant = None
for variant in variants:
if not merged_variant:
merged_variant = Variant(
reference_name=variant.reference_name,
start=variant.start,
end=variant.end,
reference_bases=variant.reference_bases,
alternate_bases=variant.alternate_bases)
# Since we use hash function in generating the merge key, there is
# a chance (extremely low though) to have variants with different
# `reference_bases` or `alternate_base` here due to a collision in
# the hash function.
assert variant.reference_bases == merged_variant.reference_bases, (
'Cannot merge variants with different reference bases. {} vs {}'
.format(variant.reference_bases,
merged_variant.reference_bases))
assert variant.alternate_bases == merged_variant.alternate_bases, (
'Cannot merge variants with different alternate bases. {} vs {}'
.format(variant.alternate_bases,
merged_variant.alternate_bases))
merged_variant.names.extend(variant.names)
merged_variant.filters.extend(variant.filters)
merged_variant.quality = max(merged_variant.quality,
variant.quality)
self.move_data_to_calls(variant)
self.move_data_to_merged(variant, merged_variant)
merged_variant.calls.extend(variant.calls)
# Deduplicate names and filters.
merged_variant.names = sorted(set(merged_variant.names))
merged_variant.filters = sorted(set(merged_variant.filters))
return [merged_variant]
def test_custom_phaseset(self):
phaseset_header_line = (
'##FORMAT=<ID=PS,Number=1,Type=Integer,Description="Phaseset">\n')
record_lines = ['19 123 . A T . . . GT:PS 1|0:1111 0/1:.\n',
'19 121 . A T . . . GT:PS 1|0:2222 0/1:2222\n']
variant_1 = Variant(
reference_name='19', start=122, end=123, reference_bases='A',
alternate_bases=['T'])
variant_1.calls.append(
VariantCall(name='Sample1', genotype=[1, 0], phaseset='1111'))
variant_1.calls.append(VariantCall(name='Sample2', genotype=[0, 1]))
variant_2 = Variant(
reference_name='19', start=120, end=121, reference_bases='A',
alternate_bases=['T'])
variant_2.calls.append(
VariantCall(name='Sample1', genotype=[1, 0], phaseset='2222'))
variant_2.calls.append(
VariantCall(name='Sample2', genotype=[0, 1], phaseset='2222'))
read_data = self._create_temp_file_and_read_records(
[phaseset_header_line] + _SAMPLE_HEADER_LINES[1:] + record_lines)
self.assertEqual(2, len(read_data))
self._assert_variants_equal([variant_1, variant_2], read_data)
def test_end_info_key_unknown_number(self):
end_info_header_line = (
'##INFO=<ID=END,Number=.,Type=Integer,Description="End of record.">\n')
record_lines = ['19 123 . A T . . END=1111 GT 1/0 0/1\n']
variant_1 = Variant(
reference_name='19', start=122, end=1111, reference_bases='A',
alternate_bases=['T'])
variant_1.calls.append(VariantCall(sample_id=hash_name('Sample1'),
name='Sample1',
genotype=[1, 0]))
variant_1.calls.append(VariantCall(sample_id=hash_name('Sample2'),
name='Sample2',
genotype=[0, 1]))
read_data = self._create_temp_file_and_read_records(
[end_info_header_line] + _SAMPLE_HEADER_LINES[1:] + record_lines)
self.assertEqual(1, len(read_data))
self._assert_variants_equal([variant_1], read_data)
def test_info_field_count(self):
coder = self._get_coder()
variant = Variant()
variant.info['NS'] = 3
variant.info['AF'] = [0.333, 0.667]
variant.info['DB'] = True
variant.info['CSQ'] = ['G|upstream_gene_variant||MODIFIER',
'T|||MODIFIER']
expected = ('. . . . . . . NS=3;AF=0.333,0.667;DB;'
'CSQ=G|upstream_gene_variant||MODIFIER,T|||MODIFIER .\n')
self._assert_variant_lines_equal(coder.encode(variant), expected)