def test_round_trip_vcf(self, test_datum_name):
# Round-trip variants through writing and reading:
# 1. Read variants v1 from VcfReader;
# 2. Write v1 to vcf using our VcfWriter;
# 3. Read back in using VcfReader -- v2;
# 4. compare v1 and v2.
in_file = test_utils.genomics_core_testdata(test_datum_name)
out_file = test_utils.test_tmpfile('output_' + test_datum_name)
v1_reader = genomics_io.make_vcf_reader(in_file, use_index=False)
v1_records = list(v1_reader.iterate())
self.assertTrue(v1_records, 'Reader failed to find records')
writer_options = core_pb2.VcfWriterOptions(
contigs=v1_reader.contigs,
sample_names=v1_reader.samples,
filters=v1_reader.filters)
with vcf_writer.VcfWriter.to_file(out_file, writer_options) as writer:
for record in v1_records:
writer.write(record)
v2_reader = genomics_io.make_vcf_reader(out_file, use_index=False)
v2_records = list(v2_reader.iterate())
self.assertEqual(v1_records, v2_records,
'Round-tripped variants not as expected')
def write_variant_to_tempfile(self, variant):
path = test_utils.test_tmpfile('test.vcf')
writer = genomics_io.make_vcf_writer(
outfile=path,
contigs=[core_pb2.ContigInfo(name='20')],
samples=[call.call_set_name for call in variant.calls],
filters=[])
with writer:
writer.write(variant)
return path
def _parse_read_with_aux_tags(self, tag_string):
# Minimal header line to create a valid SAM file.
header_lines = '@HD VN:1.3 SO:coordinate\n@SQ SN:chr1 LN:248956422\n'
# A single stock read we'll add our AUX fields to.
read = 'read_name 0 chr1 1 0 3M * 0 0 CCC AAA ' + tag_string
path = test_utils.test_tmpfile('aux_tags.bam')
with tf.gfile.FastGFile(path, 'w') as fout:
fout.write(header_lines)
fout.write(read + '\n')
with genomics_io.make_sam_reader(
path, use_index=False, parse_aux_fields=True) as reader:
return list(reader.iterate())
def setUp(self):
self.out_fname = test_utils.test_tmpfile('output.vcf')
self.options = core_pb2.VcfWriterOptions(
contigs=[
core_pb2.ContigInfo(name='Chr1', n_bases=50, pos_in_fasta=0),
core_pb2.ContigInfo(name='Chr2', n_bases=25, pos_in_fasta=1),
],
sample_names=['Fido', 'Spot'],
filters=[])
self.writer = vcf_writer.VcfWriter.to_file(self.out_fname, self.options)
self.variant = test_utils.make_variant(
chrom='Chr1', start=10, alleles=['A', 'C'])
self.variant.calls.add(genotype=[0, 0], call_set_name='Fido')
self.variant.calls.add(genotype=[0, 1], call_set_name='Spot')
def test_make_read_writer_tfrecords(self):
outfile = test_utils.test_tmpfile('test.tfrecord')
writer = genomics_io.make_read_writer(outfile=outfile)
# Test that the writer is a context manager and that we can write a read to
# it.
with writer:
writer.write(self.read1)
writer.write(self.read2)
# Our output should have exactly one read in it.
self.assertEqual([self.read1, self.read2],
list(
io_utils.read_tfrecords(outfile,
proto=reads_pb2.Read)))
def test_writing_canned_variants(self):
"""Tests writing all the variants that are 'canned' in our tfrecord file."""
# This file is in the TF record format
tfrecord_file = test_utils.genomics_core_testdata(
'test_samples.vcf.golden.tfrecord')
writer_options = core_pb2.VcfWriterOptions(
contigs=[
core_pb2.ContigInfo(name='chr1', n_bases=248956422),
core_pb2.ContigInfo(name='chr2', n_bases=242193529),
core_pb2.ContigInfo(name='chr3', n_bases=198295559),
core_pb2.ContigInfo(name='chrX', n_bases=156040895)
],
sample_names=['NA12878_18_99'],
filters=[
core_pb2.VcfFilterInfo(id='LowQual'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL95.00to96.00'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL96.00to97.00'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL97.00to99.00'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL99.00to99.50'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL99.50to99.90'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL99.90to99.95'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL99.95to100.00+'),
core_pb2.VcfFilterInfo(id='VQSRTrancheINDEL99.95to100.00'),
core_pb2.VcfFilterInfo(id='VQSRTrancheSNP99.50to99.60'),
core_pb2.VcfFilterInfo(id='VQSRTrancheSNP99.60to99.80'),
core_pb2.VcfFilterInfo(id='VQSRTrancheSNP99.80to99.90'),
core_pb2.VcfFilterInfo(id='VQSRTrancheSNP99.90to99.95'),
core_pb2.VcfFilterInfo(id='VQSRTrancheSNP99.95to100.00+'),
core_pb2.VcfFilterInfo(id='VQSRTrancheSNP99.95to100.00'),
])
variant_records = list(
io_utils.read_tfrecords(tfrecord_file, proto=variants_pb2.Variant))
out_fname = test_utils.test_tmpfile('output.vcf')
with vcf_writer.VcfWriter.to_file(out_fname, writer_options) as writer:
for record in variant_records[:5]:
writer.write(record)
# Check: are the variants written as expected?
# pylint: disable=line-too-long
expected_vcf_content = [
'##fileformat=VCFv4.2\n',
'##FILTER=<ID=PASS,Description="All filters passed">\n',
'##FILTER=<ID=LowQual,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL95.00to96.00,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL96.00to97.00,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL97.00to99.00,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL99.00to99.50,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL99.50to99.90,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL99.90to99.95,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL99.95to100.00+,Description="">\n',
'##FILTER=<ID=VQSRTrancheINDEL99.95to100.00,Description="">\n',
'##FILTER=<ID=VQSRTrancheSNP99.50to99.60,Description="">\n',
'##FILTER=<ID=VQSRTrancheSNP99.60to99.80,Description="">\n',
'##FILTER=<ID=VQSRTrancheSNP99.80to99.90,Description="">\n',
'##FILTER=<ID=VQSRTrancheSNP99.90to99.95,Description="">\n',
'##FILTER=<ID=VQSRTrancheSNP99.95to100.00+,Description="">\n',
'##FILTER=<ID=VQSRTrancheSNP99.95to100.00,Description="">\n',
'##FORMAT=<ID=GT,Number=1,Type=String,Description="Genotype">\n',
'##FORMAT=<ID=GQ,Number=1,Type=Integer,Description="Genotype Quality">\n',
'##FORMAT=<ID=DP,Number=1,Type=Integer,Description="Read depth of all '
'passing filters reads.">\n',
'##FORMAT=<ID=AD,Number=R,Type=Integer,Description="Read depth of all '
'passing filters reads for each allele.">\n',
'##FORMAT=<ID=VAF,Number=A,Type=Float,Description=\"Variant allele '
'fractions.">\n',
'##FORMAT=<ID=GL,Number=G,Type=Float,Description="Genotype '
'likelihoods, log10 encoded">\n',
'##FORMAT=<ID=PL,Number=G,Type=Integer,Description="Genotype '
'likelihoods, Phred encoded">\n',
'##INFO=<ID=END,Number=1,Type=Integer,Description="Stop position of '
'the interval">\n', '##contig=<ID=chr1,length=248956422>\n',
'##contig=<ID=chr2,length=242193529>\n',
'##contig=<ID=chr3,length=198295559>\n',
'##contig=<ID=chrX,length=156040895>\n',
'#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tNA12878_18_99\n',
'chr1\t13613\t.\tT\tA\t39.88\tVQSRTrancheSNP99.90to99.95\t.\tGT:GQ:DP:AD:PL\t0/1:16:4:1,3:68,0,16\n',
'chr1\t13813\t.\tT\tG\t90.28\tPASS\t.\tGT:GQ:DP:AD:PL\t1/1:9:3:0,3:118,9,0\n',
'chr1\t13838\trs28428499\tC\tT\t62.74\tPASS\t.\tGT:GQ:DP:AD:PL\t1/1:6:2:0,2:90,6,0\n',
'chr1\t14397\trs756427959\tCTGT\tC\t37.73\tPASS\t.\tGT:GQ:DP:AD:PL\t0/1:75:5:3,2:75,0,152\n',
'chr1\t14522\t.\tG\tA\t49.77\tVQSRTrancheSNP99.60to99.80\t.\tGT:GQ:DP:AD:PL\t0/1:78:10:6,4:78,0,118\n'
]
# pylint: enable=line-too-long
with tf.gfile.GFile(out_fname, 'r') as f:
self.assertEqual(f.readlines(), expected_vcf_content)
def write_test_protos(self, filename):
protos = [core_pb2.ContigInfo(name=str(i)) for i in range(10)]
path = test_utils.test_tmpfile(filename)
io.write_tfrecords(protos, path)
return protos, path