def main():
parser = argparse.ArgumentParser(
description='Fix dbsnp VP calls and add OXOG filter')
parser.add_argument('inputvcf',
type=argparse.FileType('r'),
default=sys.stdin,
help="Merged and annotated VCF file")
parser.add_argument('-o',
'--output',
type=argparse.FileType('w'),
default=sys.stdout,
help="Specify output file (default:stdout)")
args = parser.parse_args()
reader = vcf.Reader(args.inputvcf)
reader.infos['dbsnp_somatic'] = vcf.parser._Info(
id='dbsnp_somatic',
num=None,
type='Flag',
desc='Known-somatic dbSNP variant',
source=None,
version=None)
reader.filters['OXOGFAIL'] = vcf.parser._Filter(
id='OXOGFAIL', desc="Failed OXOG oxidative artifact filter")
reader.metadata[
'reference'] = 'ftp://ftp.sanger.ac.uk/pub/project/PanCancer/genome.fa.gz'
writer = vcf.Writer(args.output, reader)
for record in reader:
new_info = {}
# copy some records over directly
for item in ['VAF', 't_alt_count', 't_ref_count']:
if item in record.INFO and record.INFO[item] > 0:
new_info[item] = record.INFO[item]
for item in [
'dbsnp', 'cosmic', 'Callers', 'NumCallers', 'repeat_masker',
'1000genomes_AF', '1000genomes_ID'
]:
if item in record.INFO:
new_info[item] = record.INFO[item]
if 'dbsnp_VP' in record.INFO:
qualbyte = int(record.INFO['dbsnp_VP'], 16) & 255
somatic = (qualbyte & 2**5) > 0
if somatic:
new_info['dbsnp_somatic'] = True
if 'OXOG_Fail' in record.INFO:
if record.INFO['OXOG_Fail'] == 'True':
if record.FILTER is None:
record.FILTER = ['OXOGFAIL']
else:
record.FILTER.append('OXOGFAIL')
record.INFO = new_info
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(description='Annotate merged vcf with VAF information where available')
parser.add_argument('inputvcf', type=argparse.FileType('r'), default=sys.stdin, help="Merged and annotated VCF file")
parser.add_argument('-o', '--output', type=argparse.FileType('w'), default=sys.stdout, help="Specify output file (default:stdout)")
args = parser.parse_args()
reader = vcf.Reader(args.inputvcf)
reader.metadata['reference']='ftp://ftp.sanger.ac.uk/pub/project/PanCancer/genome.fa.gz'
reader.infos['dbsnp_somatic'] = vcf.parser._Info(id='dbsnp_somatic', num=None, type='Flag', desc='Known-somatic dbSNP variant', source=None, version=None)
reader.infos['t_vaf'] = vcf.parser._Info(id='t_vaf', num=1, type='Float', desc='VAF in tumor from sga', source=None, version=None)
reader.infos['n_vaf'] = vcf.parser._Info(id='n_vaf', num=1, type='Float', desc='VAF in normal from sga', source=None, version=None)
reader.infos['t_alt_count'] = vcf.parser._Info(id='t_alt_count', num=1, type='Integer', desc='Tumor alt count from sga', source=None, version=None)
reader.infos['t_ref_count'] = vcf.parser._Info(id='t_ref_count', num=1, type='Integer', desc='Tumor ref count from sga if non-zero', source=None, version=None)
reader.infos['n_alt_count'] = vcf.parser._Info(id='n_alt_count', num=1, type='Integer', desc='Normal alt count from sga if non-zero', source=None, version=None)
reader.infos['n_ref_count'] = vcf.parser._Info(id='n_ref_count', num=1, type='Integer', desc='Normal ref count from sga if non-zero', source=None, version=None)
reader.infos['model_score'] = vcf.parser._Info(id='model_score', num=1, type='Float', desc='consensus model score, 0-1', source=None, version=None)
reader.filters['LOWSUPPORT'] = vcf.parser._Filter(id='LOWSUPPORT', desc='Insufficient support in consensus model')
writer = vcf.Writer(args.output, reader)
for record in reader:
new_info = {}
# skip broad private calls
if record.INFO['Callers'] == ['broad']:
continue
# skip calls that are defined to be SVs
varlen = abs(len(record.REF) - len(record.ALT[0]))
if varlen >= 100:
continue
# copy some records over directly
for item in ['dbsnp', 'cosmic', 'Callers', 'NumCallers', 'repeat_masker', '1000genomes_AF', '1000genomes_ID']:
if item in record.INFO:
new_info[item] = record.INFO[item]
if 'model_score' in record.INFO:
new_info['model_score'] = round_dig(float(record.INFO['model_score']),3)
if 'dbsnp_VP' in record.INFO:
qualbyte = int(record.INFO['dbsnp_VP'],16) & 255
somatic = (qualbyte & 2**5) > 0
if somatic:
new_info['dbsnp_somatic'] = True
if ('TumorVarDepth' in record.INFO) and (record.INFO['TumorVarDepth'] > 0):
new_info['t_vaf'] = record.INFO['TumorVAF']
new_info['t_alt_count'] = record.INFO['TumorVarDepth']
new_info['t_ref_count'] = record.INFO['TumorTotalDepth']-record.INFO['TumorVarDepth']
if ('NormalVarDepth' in record.INFO) and (record.INFO['NormalVarDepth'] > 0):
new_info['n_vaf'] = record.INFO['NormalVAF']
new_info['n_alt_count'] = record.INFO['NormalVarDepth']
new_info['n_ref_count'] = record.INFO['NormalTotalDepth']-record.INFO['NormalVarDepth']
record.INFO = new_info
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(description='Fix dbsnp VP calls and add OXOG filter')
parser.add_argument('inputvcf', type=argparse.FileType('r'), default=sys.stdin, help="Merged and annotated VCF file")
parser.add_argument('-o', '--output', type=argparse.FileType('w'), default=sys.stdout, help="Specify output file (default:stdout)")
args = parser.parse_args()
reader = vcf.Reader(args.inputvcf)
reader.infos['dbsnp_somatic'] = vcf.parser._Info(id='dbsnp_somatic', num=None, type='Flag', desc='Known-somatic dbSNP variant', source=None, version=None)
reader.filters['OXOGFAIL'] = vcf.parser._Filter(id='OXOGFAIL', desc="Failed OXOG oxidative artifact filter")
reader.metadata['reference']='ftp://ftp.sanger.ac.uk/pub/project/PanCancer/genome.fa.gz'
writer = vcf.Writer(args.output, reader)
for record in reader:
new_info = {}
# copy some records over directly
for item in ['VAF', 't_alt_count', 't_ref_count']:
if item in record.INFO and record.INFO[item] > 0:
new_info[item] = record.INFO[item]
for item in ['dbsnp', 'cosmic', 'Callers', 'NumCallers', 'repeat_masker', '1000genomes_AF', '1000genomes_ID']:
if item in record.INFO:
new_info[item] = record.INFO[item]
if 'dbsnp_VP' in record.INFO:
qualbyte = int(record.INFO['dbsnp_VP'],16) & 255
somatic = (qualbyte & 2**5) > 0
if somatic:
new_info['dbsnp_somatic'] = True
if 'OXOG_Fail' in record.INFO:
if record.INFO['OXOG_Fail'] == 'True':
if record.FILTER is None:
record.FILTER = ['OXOGFAIL']
else:
record.FILTER.append('OXOGFAIL')
record.INFO = new_info
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(
description='Filter Y-chromosome calls if sex is female')
parser.add_argument('-i',
'--input',
type=argparse.FileType('r'),
default=sys.stdin,
help="input VCF file")
parser.add_argument('-o',
'--output',
type=argparse.FileType('w'),
default=sys.stdout,
help="Specify output file (default:stdout)")
parser.add_argument(
'-s',
'--sex',
type=str,
default="male",
help="Apply filter if this flag is exactly 'female' (default:male)")
args = parser.parse_args()
apply_filter = args.sex == "female"
filtername = 'SEXF'
reader = vcf.Reader(args.input)
if apply_filter:
reader.filters[filtername] = vcf.parser._Filter(
id=filtername,
desc=
'Likely artifact or call in PAR region: Y-chromosome variant in female donor'
)
writer = vcf.Writer(args.output, reader)
for record in reader:
if (apply_filter and record.CHROM in ['Y', 'chrY']):
if not record.FILTER:
record.FILTER = [filtername]
else:
record.FILTER += [filtername]
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(description='Filter Y-chromosome calls if sex is female')
parser.add_argument('-i', '--input', type=argparse.FileType('r'), default=sys.stdin, help="input VCF file")
parser.add_argument('-o', '--output', type=argparse.FileType('w'), default=sys.stdout, help="Specify output file (default:stdout)")
parser.add_argument('-s', '--sex', type=str, default="male", help="Apply filter if this flag is exactly 'female' (default:male)")
args = parser.parse_args()
apply_filter = args.sex == "female"
filtername = 'SEXF'
reader = vcf.Reader(args.input)
if apply_filter:
reader.filters[filtername] = vcf.parser._Filter(id=filtername, desc='Likely artifact or call in PAR region: Y-chromosome variant in female donor')
writer = vcf.Writer(args.output, reader)
for record in reader:
if (apply_filter and record.CHROM in ['Y', 'chrY']):
if not record.FILTER:
record.FILTER = [filtername]
else:
record.FILTER += [filtername]
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(
description='Add info or filter tag based on presence in MAF')
parser.add_argument('MAF',
type=str,
help="MAF file for variant classification annotation")
parser.add_argument('name', help='Name of info field or filter')
parser.add_argument('-i',
'--input',
type=argparse.FileType('r'),
default=sys.stdin,
help="VCF file to be processed (default: stdin)")
parser.add_argument('-o',
'--output',
type=argparse.FileType('w'),
default=sys.stdout,
help="Output file (default:stdout)")
parser.add_argument('-n',
'--info',
action='store_true',
help="Add info flag rather than filter")
parser.add_argument('-c',
'--column',
type=str,
help='column in MAF to use for info, if present')
parser.add_argument('-d',
'--description',
default="",
type=str,
help='description of new info/filter field')
args = parser.parse_args()
reader = vcf.Reader(args.input)
if args.info:
reader.infos[args.name] = vcf.parser._Info(id=args.name,
num='1',
type='String',
desc=args.description,
source=None,
version=None)
else:
reader.filters[args.name] = vcf.parser._Filter(id=args.name,
desc=args.description)
writer = vcf.Writer(args.output, reader)
classification_dict = get_classification_dict_from_MAF(
args.MAF, args.column)
for record in reader:
assert len(record.ALT) == 1
variant = variant_tuple(record, record.ALT[0])
if variant in classification_dict:
if args.info:
record.INFO[args.name] = classification_dict[variant]
else:
if not record.FILTER:
record.FILTER = [args.name]
else:
record.FILTER += [args.name]
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(description='Add info or filter tag based on presence in MAF')
parser.add_argument('MAF', type=str, help="MAF file for variant classification annotation")
parser.add_argument('name', help='Name of info field or filter')
parser.add_argument('-i', '--input', type=argparse.FileType('r'), default=sys.stdin, help="VCF file to be processed (default: stdin)")
parser.add_argument('-o', '--output', type=argparse.FileType('w'), default=sys.stdout, help="Output file (default:stdout)")
parser.add_argument('-a', '--action', choices=['info','filter'], default='info', help='add tag (info) or filter based on presence in MAF (default:info)')
parser.add_argument('-c', '--column', type=str, help='column in MAF to use for info, if present')
parser.add_argument('-d', '--description', default="", type=str, help='description of new info/filter field')
args = parser.parse_args()
reader = vcf.Reader(args.input)
if args.action == "info":
reader.infos[args.name] = vcf.parser._Info(id=args.name, num='1', type='String', desc=args.description, source=None, version=None)
else:
reader.filters[args.name] = vcf.parser._Filter(id=args.name, desc=args.description)
writer = vcf.Writer(args.output, reader)
classification_dict = get_classification_dict_from_MAF(args.MAF, args.column)
for record in reader:
assert len(record.ALT) == 1
variant = variant_tuple(record, record.ALT[0])
if variant in classification_dict:
if args.action == "info":
record.INFO[args.name] = classification_dict[variant]
else:
if not record.FILTER:
record.FILTER = [args.name]
else:
record.FILTER += [args.name]
writer.write_record(record)
return 0
def filter_calls():
parser = argparse.ArgumentParser( description='Set genotypes based on DP4 scores')
parser.add_argument('-i', '--input', type=argparse.FileType('r'), default=sys.stdin)
parser.add_argument('-o', '--output', type=argparse.FileType('w'), default=sys.stdout)
parser.add_argument('-e', '--error', type=float, default=0.01, help='Error rate')
parser.add_argument('-t', '--callthreshold', type=float, default=0.02, help='Max prob to call')
parser.add_argument('-s', '--strandbias', type=float, default=0.10, help='minimum strand ratio')
parser.add_argument('-m', '--mindepth', type=int, default=10, help='minimum total depth')
parser.add_argument('-g', '--germlineprob', type=float, default=0.02, help='Maximum prob of germline')
args = parser.parse_args()
vcf_reader = vcf.Reader(args.input)
vcf_reader.infos['Validation_status'] = vcf.parser._Info(id='Validation_status', num='.', type='String',
desc='Status from validation data',
source=None, version=None)
vcf_writer = vcf.Writer(args.output, vcf_reader)
for record in vcf_reader:
normal_reads = int(record.INFO['NormalReads'][0])
tumour_reads = int(record.INFO['TumourReads'][0])
normal_evidence = [int(x) for x in record.INFO['NormalEvidenceReads']]
tumour_evidence = [int(x) for x in record.INFO['TumourEvidenceReads']]
if min(normal_reads, tumour_reads) < args.mindepth:
record.FILTER = ['LOWDEPTH']
record.INFO['Validation_status'] = 'LOWDEPTH'
vcf_writer.write_record(record)
continue
record.FILTER = []
if sum(tumour_evidence) < 7 or not call_from_depths(tumour_reads, tumour_evidence, args.error, args.callthreshold):
record.FILTER = ['NOTSEEN']
if (tumour_reads > 0) > 0 and reject_from_strandbias(tumour_reads, tumour_evidence, args.strandbias):
record.FILTER += ['STRANDBIAS']
if germline_hom_het(normal_reads, normal_evidence, tumour_reads, tumour_evidence, args.germlineprob):
record.FILTER += ['GERMLINE']
elif reject_from_normal_evidence_vs_noise(normal_reads, normal_evidence, tumour_reads, tumour_evidence, args.error):
record.FILTER += ['NORMALEVIDENCE']
if len(record.FILTER) == 0:
record.FILTER = ['PASS']
record.INFO['Validation_status'] = ','.join(record.FILTER)
vcf_writer.write_record(record)
def filter_calls():
parser = argparse.ArgumentParser(
description='Set genotypes based on DP4 scores')
parser.add_argument('-i',
'--input',
type=argparse.FileType('r'),
default=sys.stdin)
parser.add_argument('-o',
'--output',
type=argparse.FileType('w'),
default=sys.stdout)
parser.add_argument('-e',
'--error',
type=float,
default=0.01,
help='Error rate')
parser.add_argument('-t',
'--callthreshold',
type=float,
default=0.02,
help='Max prob to call')
parser.add_argument('-s',
'--strandbias',
type=float,
default=0.10,
help='minimum strand ratio')
parser.add_argument('-m',
'--mindepth',
type=int,
default=10,
help='minimum total depth')
parser.add_argument('-g',
'--germlineprob',
type=float,
default=0.02,
help='Maximum prob of germline')
args = parser.parse_args()
vcf_reader = vcf.Reader(args.input)
vcf_reader.infos['Validation_status'] = vcf.parser._Info(
id='Validation_status',
num='.',
type='String',
desc='Status from validation data',
source=None,
version=None)
vcf_writer = vcf.Writer(args.output, vcf_reader)
for record in vcf_reader:
normal_reads = int(record.INFO['NormalReads'][0])
tumour_reads = int(record.INFO['TumourReads'][0])
normal_evidence = [int(x) for x in record.INFO['NormalEvidenceReads']]
tumour_evidence = [int(x) for x in record.INFO['TumourEvidenceReads']]
if min(normal_reads, tumour_reads) < args.mindepth:
record.FILTER = ['LOWDEPTH']
record.INFO['Validation_status'] = 'LOWDEPTH'
vcf_writer.write_record(record)
continue
record.FILTER = []
if sum(tumour_evidence) < 7 or not call_from_depths(
tumour_reads, tumour_evidence, args.error, args.callthreshold):
record.FILTER = ['NOTSEEN']
if (tumour_reads > 0) > 0 and reject_from_strandbias(
tumour_reads, tumour_evidence, args.strandbias):
record.FILTER += ['STRANDBIAS']
if germline_hom_het(normal_reads, normal_evidence, tumour_reads,
tumour_evidence, args.germlineprob):
record.FILTER += ['GERMLINE']
elif reject_from_normal_evidence_vs_noise(normal_reads,
normal_evidence,
tumour_reads,
tumour_evidence, args.error):
record.FILTER += ['NORMALEVIDENCE']
if len(record.FILTER) == 0:
record.FILTER = ['PASS']
record.INFO['Validation_status'] = ','.join(record.FILTER)
vcf_writer.write_record(record)
def main():
parser = argparse.ArgumentParser(
description='Fix dbsnp VP calls and add OXOG filter')
parser.add_argument('MAF', type=str, help="MAF file for filtering")
parser.add_argument('sample', type=str, help="tumour aliquot id")
parser.add_argument('filtername', type=str, help="Filter name to apply")
parser.add_argument('-o',
'--output',
type=argparse.FileType('w'),
default=sys.stdout,
help="Specify output file (default:stdout)")
parser.add_argument('-i',
'--input',
type=argparse.FileType('r'),
default=sys.stdin,
help="Merged and annotated VCF file (default: stdin)")
parser.add_argument('-d',
'--desc',
type=str,
default="",
help="Description of filter")
parser.add_argument('-n',
'--info',
action='store_true',
help="Add info flag rather than filter")
args = parser.parse_args()
reader = vcf.Reader(args.input)
if args.info:
reader.infos[args.filtername] = vcf.parser._Info(id=args.filtername,
num=0,
type='Flag',
desc=args.desc,
source=None,
version=None)
else:
reader.filters[args.filtername] = vcf.parser._Filter(
id=args.filtername, desc=args.desc)
writer = vcf.Writer(args.output, reader)
entries = get_entries_from_MAF(args.MAF)
for record in reader:
variants = [variant_tuple(args.sample, record)]
if len(record.ALT[0]) != len(record.REF):
variants += [
(args.sample, record.CHROM,
record.POS + abs(len(record.ALT[0]) - len(record.REF)))
]
for variant in variants:
if variant in entries:
if not args.info:
if not record.FILTER:
record.FILTER = [args.filtername]
elif args.filtername not in record.FILTER:
record.FILTER = record.FILTER + [args.filtername]
else:
record.INFO[args.filtername] = True
writer.write_record(record)
return 0
def main():
parser = argparse.ArgumentParser(
description='Annotate merged vcf with VAF information where available')
parser.add_argument('inputvcf',
type=argparse.FileType('r'),
default=sys.stdin,
help="Merged and annotated VCF file")
parser.add_argument('-o',
'--output',
type=argparse.FileType('w'),
default=sys.stdout,
help="Specify output file (default:stdout)")
args = parser.parse_args()
reader = vcf.Reader(args.inputvcf)
reader.metadata[
'reference'] = 'ftp://ftp.sanger.ac.uk/pub/project/PanCancer/genome.fa.gz'
reader.infos['dbsnp_somatic'] = vcf.parser._Info(
id='dbsnp_somatic',
num=None,
type='Flag',
desc='Known-somatic dbSNP variant',
source=None,
version=None)
reader.infos['t_vaf'] = vcf.parser._Info(id='t_vaf',
num=1,
type='Float',
desc='VAF in tumor from sga',
source=None,
version=None)
reader.infos['n_vaf'] = vcf.parser._Info(id='n_vaf',
num=1,
type='Float',
desc='VAF in normal from sga',
source=None,
version=None)
reader.infos['t_alt_count'] = vcf.parser._Info(
id='t_alt_count',
num=1,
type='Integer',
desc='Tumor alt count from sga',
source=None,
version=None)
reader.infos['t_ref_count'] = vcf.parser._Info(
id='t_ref_count',
num=1,
type='Integer',
desc='Tumor ref count from sga if non-zero',
source=None,
version=None)
reader.infos['n_alt_count'] = vcf.parser._Info(
id='n_alt_count',
num=1,
type='Integer',
desc='Normal alt count from sga if non-zero',
source=None,
version=None)
reader.infos['n_ref_count'] = vcf.parser._Info(
id='n_ref_count',
num=1,
type='Integer',
desc='Normal ref count from sga if non-zero',
source=None,
version=None)
reader.infos['model_score'] = vcf.parser._Info(
id='model_score',
num=1,
type='Float',
desc='consensus model score, 0-1',
source=None,
version=None)
reader.filters['LOWSUPPORT'] = vcf.parser._Filter(
id='LOWSUPPORT', desc='Insufficient support in consensus model')
writer = vcf.Writer(args.output, reader)
for record in reader:
new_info = {}
# skip broad private calls
if record.INFO['Callers'] == ['broad']:
continue
# skip calls that are defined to be SVs
varlen = abs(len(record.REF) - len(record.ALT[0]))
if varlen >= 100:
continue
# copy some records over directly
for item in [
'dbsnp', 'cosmic', 'Callers', 'NumCallers', 'repeat_masker',
'1000genomes_AF', '1000genomes_ID'
]:
if item in record.INFO:
new_info[item] = record.INFO[item]
if 'model_score' in record.INFO:
new_info['model_score'] = round_dig(
float(record.INFO['model_score']), 3)
if 'dbsnp_VP' in record.INFO:
qualbyte = int(record.INFO['dbsnp_VP'], 16) & 255
somatic = (qualbyte & 2**5) > 0
if somatic:
new_info['dbsnp_somatic'] = True
if ('TumorVarDepth'
in record.INFO) and (record.INFO['TumorVarDepth'] > 0):
new_info['t_vaf'] = record.INFO['TumorVAF']
new_info['t_alt_count'] = record.INFO['TumorVarDepth']
new_info['t_ref_count'] = record.INFO[
'TumorTotalDepth'] - record.INFO['TumorVarDepth']
if ('NormalVarDepth'
in record.INFO) and (record.INFO['NormalVarDepth'] > 0):
new_info['n_vaf'] = record.INFO['NormalVAF']
new_info['n_alt_count'] = record.INFO['NormalVarDepth']
new_info['n_ref_count'] = record.INFO[
'NormalTotalDepth'] - record.INFO['NormalVarDepth']
record.INFO = new_info
writer.write_record(record)
return 0