def main():
global options, args
# Open and parse each line of the vcf file
input_vcf = vcf.Reader(open(options.input_vcf, 'r'))
# If an FREQ field already exists in FORMAT or INFO, it has to be stored and be used when importing from input
former_vcfformat_freq = input_vcf.formats[
'FREQ'] if 'FREQ' in input_vcf.formats else None
former_vcfinfo_sfreq = input_vcf.infos[
'FREQ'] if 'FREQ' in input_vcf.infos else None
former_vcfinfo_sdp = input_vcf.infos[
'DPS'] if 'DPS' in input_vcf.infos else None
input_vcf.formats['FREQ'] = VcfFormat('FREQ', None, 'String',
'Variant allele frequency')
input_vcf.infos['SFREQ'] = VcfInfo(
'SFREQ', 1, 'Float', 'Maximum variant allele frequency of all samples')
input_vcf.infos['SDP'] = VcfInfo(
'SDP', 1, 'Integer', 'Maximum sequencing depth of all samples')
output_vcf = vcf.Writer(open(options.output_vcf, 'w'),
input_vcf,
lineterminator='\n')
if former_vcfformat_freq is not None:
input_vcf.formats['FREQ'] = former_vcfformat_freq
if former_vcfinfo_sfreq is not None:
input_vcf.infos['SFREQ'] = former_vcfinfo_sfreq
if former_vcfinfo_sdp is not None:
input_vcf.infos['SDP'] = former_vcfinfo_sdp
for record in input_vcf:
if not 'FREQ' in record.FORMAT.split(':'):
record.add_format('FREQ')
# Default values for added INFO fields
site_freq = None
site_depth = 0
# iterate over all call objects of record
for call in record.samples:
# Allele frequency and Depth evaluation among samples
try:
site_freq = max(site_freq, max(
call.aaf)) if call.aaf is not None else site_freq
site_depth = max(
call.depth,
site_depth) if call.depth is not None else site_depth
except Exception:
print "ERROR: unforeseen exception when normalizing record:", record
raise
call.add_format('FREQ', norm_freq(call.aaf))
# TODO: unfortunately GATK filtering doesn't yet deal correctly with "None" (.) values
if site_freq is None or site_freq == '.':
site_freq = 0
record.add_info('SFREQ', site_freq)
record.add_info('SDP', site_depth)
output_vcf.write_record(record)
def main():
global options, args
# Open and parse each line of the vcf file
input_vcf = vcf.Reader(open(options.input_vcf, 'r'))
# If an FREQ field already exists in FORMAT or INFO, it has to be stored and be used when importing from input
former_vcfformat_freq = input_vcf.formats['FREQ'] if 'FREQ' in input_vcf.formats else None
former_vcfinfo_sfreq = input_vcf.infos['FREQ'] if 'FREQ' in input_vcf.infos else None
former_vcfinfo_sdp = input_vcf.infos['DPS'] if 'DPS' in input_vcf.infos else None
input_vcf.formats['FREQ'] = VcfFormat('FREQ', None, 'String', 'Variant allele frequency')
input_vcf.infos['SFREQ'] = VcfInfo('SFREQ', 1, 'Float', 'Maximum variant allele frequency of all samples')
input_vcf.infos['SDP'] = VcfInfo('SDP', 1, 'Integer', 'Maximum sequencing depth of all samples')
output_vcf = vcf.Writer(open(options.output_vcf, 'w'), input_vcf, lineterminator='\n')
if former_vcfformat_freq is not None:
input_vcf.formats['FREQ'] = former_vcfformat_freq
if former_vcfinfo_sfreq is not None:
input_vcf.infos['SFREQ'] = former_vcfinfo_sfreq
if former_vcfinfo_sdp is not None:
input_vcf.infos['SDP'] = former_vcfinfo_sdp
for record in input_vcf:
if not 'FREQ' in record.FORMAT.split(':'):
record.add_format('FREQ')
# Default values for added INFO fields
site_freq = None
site_depth = 0
# iterate over all call objects of record
for call in record.samples:
# Allele frequency and Depth evaluation among samples
try:
site_freq = max(site_freq, max(call.aaf)) if call.aaf is not None else site_freq
site_depth = max(call.depth, site_depth) if call.depth is not None else site_depth
except Exception:
print "ERROR: unforeseen exception when normalizing record:", record
raise
call.add_format('FREQ', norm_freq(call.aaf))
# TODO: unfortunately GATK filtering doesn't yet deal correctly with "None" (.) values
if site_freq is None or site_freq == '.':
site_freq = 0
record.add_info('SFREQ', site_freq)
record.add_info('SDP', site_depth)
output_vcf.write_record(record)
def main():
global options, args
# Create template for outputting from input file
template_vcf = vcf.Reader(open(options.input_vcf, 'r'))
template_vcf.formats['FREQ'] = VcfFormat('FREQ', None, 'String',
'Variant allele frequency')
# Open and parse each line of the vcf file
input_vcf = vcf.Reader(open(options.input_vcf, 'r'))
repaired_vcf = vcf.Writer(open(options.output_vcf, 'w'),
template_vcf,
lineterminator='\n')
for record in input_vcf:
if len(record.ALT) > 1:
# Problem 1 in VarScan files: alternative allele repeated. Output unique alternative alleles
new_alt = []
for index, alternative in enumerate(record.ALT):
if not alternative in new_alt:
new_alt.append(alternative)
for call in record.samples:
if call.gt_nums is not None:
gt_alleles = []
for allele in call.gt_bases.split(call.gt_phase_char()):
try:
gt_alleles.append(str(new_alt.index(allele) + 1))
except ValueError:
gt_alleles.append('0')
call.add_format('GT',
call.gt_phase_char().join(gt_alleles))
call.gt_nums = call.gt_phase_char().join(gt_alleles)
record.ALT = new_alt
record.alleles = [record.REF] + new_alt
# Problem 2 in VarScan files: only a FREQ value is stored, even in case of multiple alternative alleles
for call in record.samples:
# Assign allele frequency to those alleles predicted by variant caller
freqs = []
for index in range(1, len(record.ALT) + 1):
if call.gt_nums == '0/0' or call.gt_nums == '0|0':
freqs.append(norm_freq(call.aaf[0]))
elif call.gt_alleles is None:
freqs.append(None)
elif str(index) in call.gt_alleles and (
'0' in call.gt_alleles or call.gt_type == 2):
freqs.append(norm_freq(call.aaf[0]))
elif str(index) not in call.gt_alleles:
freqs.append(None)
else:
print "ERROR: unforeseen case when obtaining freqs!!"
print "record:", record
raise
call.add_format('FREQ', freqs)
else:
for call in record.samples:
call.add_format('FREQ', norm_freq(call.aaf))
# Problem 3: hybrid records (both SNP and Indel in a single record)
snv_record = None
indel_record = None
if record.is_indel and len(record.ALT) > 1:
# Creation of lists containing indels and snvs
indel_alternatives = []
snv_alternatives = []
for index, alt in enumerate(record.ALT):
if len(alt) == len(
record.REF) and alt.sequence[1:] == record.REF[1:]:
if not alt in snv_alternatives:
snv_alternatives.append(alt)
else:
if not alt in indel_alternatives:
indel_alternatives.append(alt)
for index, alt in enumerate(record.ALT):
if len(alt) == len(record.REF):
# Check if alternative allele could be expressed as SNV
if alt.sequence[1:] == record.REF[1:]:
print "INFO: splitting hybrid record (containing both SNP and indel)", record
my_alternative = Substitution(alt.sequence[0])
my_samples = []
for call in record.samples:
my_samples.append(
Call(call.site, call.sample, call.data.copy()))
my_record = Record(record.CHROM, record.POS, record.ID,
record.REF[0], [my_alternative],
record.QUAL, record.FILTER,
record.INFO, record.FORMAT,
record._sample_indexes, my_samples)
for call in my_record.samples:
call.add_format('FREQ',
call.get_format('FREQ')[index])
if call.gt_nums is not None:
gt_alleles = []
for allele in call.gt_bases.split(
call.gt_phase_char()):
try:
gt_alleles.append(
str(
snv_alternatives.index(allele)
+ 1))
except ValueError:
gt_alleles.append('0')
call.add_format(
'GT',
call.gt_phase_char().join(gt_alleles))
call.gt_nums = call.gt_phase_char().join(
gt_alleles)
if snv_record is None:
snv_record = my_record
else:
if not snv_record.merge(my_record):
print "ERROR: Impossible to split record", record
raise
else:
# Indel
my_samples = []
for call in record.samples:
my_samples.append(
Call(call.site, call.sample, call.data.copy()))
my_record = Record(record.CHROM, record.POS, record.ID,
record.REF, [alt], record.QUAL,
record.FILTER, record.INFO,
record.FORMAT, record._sample_indexes,
my_samples)
for call in my_record.samples:
call.add_format('FREQ', call.get_format('FREQ')[index])
if call.gt_nums is not None:
gt_alleles = []
for allele in call.gt_bases.split(
call.gt_phase_char()):
try:
gt_alleles.append(
str(
indel_alternatives.index(allele) +
1))
except ValueError:
gt_alleles.append('0')
call.add_format(
'GT',
call.gt_phase_char().join(gt_alleles))
call.gt_nums = call.gt_phase_char().join(
gt_alleles)
if indel_record is None:
indel_record = my_record
else:
if not indel_record.merge(my_record):
print "ERROR: Impossible to split record", record
raise
if snv_record is not None:
repaired_vcf.write_record(snv_record)
repaired_vcf.write_record(indel_record)
else:
repaired_vcf.write_record(record)
def main():
global options, args
# Create template for outputting from input file
template_vcf = vcf.Reader(open(options.input_vcf, 'r'))
template_vcf.formats['FREQ'] = VcfFormat('FREQ', None, 'String', 'Variant allele frequency')
# Open and parse each line of the vcf file
input_vcf = vcf.Reader(open(options.input_vcf, 'r'))
repaired_vcf = vcf.Writer(open(options.output_vcf, 'w'), template_vcf, lineterminator='\n')
for record in input_vcf:
if len(record.ALT) > 1:
# Problem 1 in VarScan files: alternative allele repeated. Output unique alternative alleles
new_alt = []
for index, alternative in enumerate(record.ALT):
if not alternative in new_alt:
new_alt.append(alternative)
for call in record.samples:
if call.gt_nums is not None:
gt_alleles = []
for allele in call.gt_bases.split(call.gt_phase_char()):
try:
gt_alleles.append(str(new_alt.index(allele)+1))
except ValueError:
gt_alleles.append('0')
call.add_format('GT', call.gt_phase_char().join(gt_alleles))
call.gt_nums = call.gt_phase_char().join(gt_alleles)
record.ALT = new_alt
record.alleles = [record.REF] + new_alt
# Problem 2 in VarScan files: only a FREQ value is stored, even in case of multiple alternative alleles
for call in record.samples:
# Assign allele frequency to those alleles predicted by variant caller
freqs = []
for index in range(1, len(record.ALT) + 1):
if call.gt_nums == '0/0' or call.gt_nums == '0|0':
freqs.append(norm_freq(call.aaf[0]))
elif call.gt_alleles is None:
freqs.append(None)
elif str(index) in call.gt_alleles and ('0' in call.gt_alleles or call.gt_type == 2):
freqs.append(norm_freq(call.aaf[0]))
elif str(index) not in call.gt_alleles:
freqs.append(None)
else:
print "ERROR: unforeseen case when obtaining freqs!!"
print "record:", record
raise
call.add_format('FREQ', freqs)
else:
for call in record.samples:
call.add_format('FREQ', norm_freq(call.aaf))
# Problem 3: hybrid records (both SNP and Indel in a single record)
snv_record = None
indel_record = None
if record.is_indel and len(record.ALT) > 1:
# Creation of lists containing indels and snvs
indel_alternatives = []
snv_alternatives = []
for index, alt in enumerate(record.ALT):
if len(alt) == len(record.REF) and alt.sequence[1:] == record.REF[1:]:
if not alt in snv_alternatives:
snv_alternatives.append(alt)
else:
if not alt in indel_alternatives:
indel_alternatives.append(alt)
for index, alt in enumerate(record.ALT):
if len(alt) == len(record.REF):
# Check if alternative allele could be expressed as SNV
if alt.sequence[1:] == record.REF[1:]:
print "INFO: splitting hybrid record (containing both SNP and indel)", record
my_alternative = Substitution(alt.sequence[0])
my_samples = []
for call in record.samples:
my_samples.append(Call(call.site, call.sample, call.data.copy()))
my_record = Record(record.CHROM, record.POS, record.ID, record.REF[0], [my_alternative],
record.QUAL, record.FILTER, record.INFO, record.FORMAT,
record._sample_indexes, my_samples)
for call in my_record.samples:
call.add_format('FREQ', call.get_format('FREQ')[index])
if call.gt_nums is not None:
gt_alleles = []
for allele in call.gt_bases.split(call.gt_phase_char()):
try:
gt_alleles.append(str(snv_alternatives.index(allele)+1))
except ValueError:
gt_alleles.append('0')
call.add_format('GT', call.gt_phase_char().join(gt_alleles))
call.gt_nums = call.gt_phase_char().join(gt_alleles)
if snv_record is None:
snv_record = my_record
else:
if not snv_record.merge(my_record):
print "ERROR: Impossible to split record", record
raise
else:
# Indel
my_samples = []
for call in record.samples:
my_samples.append(Call(call.site, call.sample, call.data.copy()))
my_record = Record(record.CHROM, record.POS, record.ID, record.REF, [alt],
record.QUAL, record.FILTER, record.INFO, record.FORMAT,
record._sample_indexes, my_samples)
for call in my_record.samples:
call.add_format('FREQ', call.get_format('FREQ')[index])
if call.gt_nums is not None:
gt_alleles = []
for allele in call.gt_bases.split(call.gt_phase_char()):
try:
gt_alleles.append(str(indel_alternatives.index(allele)+1))
except ValueError:
gt_alleles.append('0')
call.add_format('GT', call.gt_phase_char().join(gt_alleles))
call.gt_nums = call.gt_phase_char().join(gt_alleles)
if indel_record is None:
indel_record = my_record
else:
if not indel_record.merge(my_record):
print "ERROR: Impossible to split record", record
raise
if snv_record is not None:
repaired_vcf.write_record(snv_record)
repaired_vcf.write_record(indel_record)
else:
repaired_vcf.write_record(record)
