def process_vcf(archive, vcf, vcf_index, output_prefix):
"""
Extracts and processes the caveman vcf file.
"""
out_raw_vcf = '{0}.tmp.vcf.gz'.format(output_prefix)
logger.info("Extracting raw vcf to tmp file {0}".format(out_raw_vcf))
extract_file(archive, vcf, out_raw_vcf)
out_raw_vcf_index = '{0}.tmp.vcf.gz.tbi'.format(output_prefix)
logger.info(
"Extracting raw vcf index to tmp file {0}".format(out_raw_vcf_index))
extract_file(archive, vcf_index, out_raw_vcf_index)
# Update the sample name using BGZFile which doesn't assert any VCF format
logger.info("Processing raw VCF to change TUMOUR -> TUMOR...")
out_formatted_vcf = '{0}.vcf.gz'.format(output_prefix)
logger.info("Creating final vcf {0}".format(out_formatted_vcf))
writer = pysam.BGZFile(out_formatted_vcf, mode='wb')
reader = pysam.BGZFile(out_raw_vcf, mode='rb')
try:
for line in reader:
line = line.decode('utf-8')
if line.startswith('##'):
if line.startswith('##SAMPLE=<ID=TUMOUR'):
new_line = line.replace('ID=TUMOUR', 'ID=TUMOR') + '\n'
writer.write(new_line.encode('utf-8'))
else:
new_line = line + '\n'
writer.write(new_line.encode('utf-8'))
elif line.startswith('#CHROM'):
new_line = line.replace('TUMOUR', 'TUMOR') + '\n'
writer.write(new_line.encode('utf-8'))
else:
# BINF-306: fix rare case of alt == ref in caveman vcf.
cols = line.split('\t')
if cols[3] == cols[4]:
logger.warn(
"Removing loci {0}:{1} where ref and alt alleles are same: {2} - {3}"
.format(cols[0], cols[1], cols[3], cols[4]))
continue
new_line = line + '\n'
writer.write(new_line.encode('utf-8'))
finally:
writer.close()
reader.close()
# tabix index
logger.info("Creating final vcf index {0}".format(out_formatted_vcf +
'.tbi'))
pysam.tabix_index(out_formatted_vcf, preset='vcf', force=True)
# clean up
logger.info("Cleaning up tmp files...")
os.remove(out_raw_vcf)
os.remove(out_raw_vcf_index)
def process_bedpe(archive, bedpe, bedpe_index, output_prefix):
"""
Extracts and processes the brass bedpe file.
"""
out_raw_bedpe = '{0}.tmp.bedpe.gz'.format(output_prefix)
logger.info("Extracting raw bedpe to tmp file {0}".format(out_raw_bedpe))
extract_file(archive, bedpe, out_raw_bedpe)
out_raw_bedpe_index = '{0}.tmp.bedpe.gz.tbi'.format(output_prefix)
logger.info("Extracting raw bedpe index to tmp file {0}".format(out_raw_bedpe_index))
extract_file(archive, bedpe_index, out_raw_bedpe_index)
out_formatted_bedpe = '{0}.bedpe.gz'.format(output_prefix)
logger.info("Creating final bedpe {0}".format(out_formatted_bedpe))
writer = pysam.BGZFile(out_formatted_bedpe, mode='wb')
reader = pysam.BGZFile(out_raw_bedpe, mode='rb')
try:
meta_line = None
process_header = False
hdr = []
for line in reader:
line = line.decode('utf-8')
if line.startswith('#'):
meta_line = line
else:
if not process_header:
hdr = format_header(meta_line)
assert len(hdr) == len(set(hdr)), \
"Duplicate header keys {0}".format(','.join(hdr))
writer.write(('#' + '\t'.join(hdr) + '\n').encode('utf-8'))
process_header = True
dat = dict(zip(hdr, line.rstrip('\r\n').split('\t')))
dat['brass_notation'] = dat['brass_notation'].replace('Chr.chr', 'chr')
new_line = "\t".join([dat[i] for i in hdr]) + '\n'
writer.write(new_line.encode('utf-8'))
finally:
writer.close()
reader.close()
# tabix index
logger.info("Creating final bedpe index {0}".format(out_formatted_bedpe + '.tbi'))
pysam.tabix_index( out_formatted_bedpe, preset='bed', force=True )
# clean up
logger.info("Cleaning up tmp files...")
os.remove(out_raw_bedpe)
os.remove(out_raw_bedpe_index)
def process_vcf(archive, vcf, vcf_index, output_prefix):
"""
Extracts and processes the pindel vcf file.
"""
out_raw_vcf = '{0}.tmp.vcf.gz'.format(output_prefix)
logger.info("Extracting raw vcf to tmp file {0}".format(out_raw_vcf))
extract_file(archive, vcf, out_raw_vcf)
out_raw_vcf_index = '{0}.tmp.vcf.gz.tbi'.format(output_prefix)
logger.info(
"Extracting raw vcf index to tmp file {0}".format(out_raw_vcf_index))
extract_file(archive, vcf_index, out_raw_vcf_index)
# Update the sample name using BGZFile which doesn't assert any VCF format
logger.info("Processing raw VCF to change TUMOUR -> TUMOR...")
out_formatted_vcf = '{0}.vcf.gz'.format(output_prefix)
logger.info("Creating final vcf {0}".format(out_formatted_vcf))
writer = pysam.BGZFile(out_formatted_vcf, mode='wb')
reader = pysam.BGZFile(out_raw_vcf, mode='rb')
try:
for line in reader:
line = line.decode('utf-8')
if line.startswith('##'):
if line.startswith('##SAMPLE=<ID=TUMOUR'):
new_line = line.replace('ID=TUMOUR', 'ID=TUMOR') + '\n'
writer.write(new_line.encode('utf-8'))
else:
new_line = line + '\n'
writer.write(new_line.encode('utf-8'))
elif line.startswith('#CHROM'):
new_line = line.replace('TUMOUR', 'TUMOR') + '\n'
writer.write(new_line.encode('utf-8'))
else:
new_line = line + '\n'
writer.write(new_line.encode('utf-8'))
finally:
writer.close()
reader.close()
# tabix index
logger.info("Creating final vcf index {0}".format(out_formatted_vcf +
'.tbi'))
pysam.tabix_index(out_formatted_vcf, preset='vcf', force=True)
# clean up
logger.info("Cleaning up tmp files...")
os.remove(out_raw_vcf)
os.remove(out_raw_vcf_index)
def addCADD(in_VCF, in_CADD_files, out_VCF):
cadds = CADD(in_CADD_files)
with pysam.VariantFile(in_VCF, 'r') as ifile, pysam.BGZFile(out_VCF,
'w') as ofile:
for x in ['CADD_RAW', 'CADD_PHRED']:
if x in ifile.header.info:
raise Exception(
'{} already exists in input VCF/BCF.'.format(x))
ifile.header.add_line(
'##INFO=<ID=CADD_RAW,Number=A,Type=Float,Description="Raw CADD scores">'
)
ifile.header.add_line(
'##INFO=<ID=CADD_PHRED,Number=A,Type=Float,Desctiption="Phred-scaled CADD scores">'
)
ofile.write('{}'.format(ifile.header).encode())
for record in ifile:
raw_scores = []
phred_scores = []
for alt in record.alts:
cadd_raw, cadd_phred = cadds.get(record.chrom, record.pos,
record.ref, alt)
raw_scores.append(cadd_raw)
phred_scores.append(cadd_phred)
if any(x for x in raw_scores) and any(x for x in phred_scores):
record.info['CADD_RAW'] = raw_scores
record.info['CADD_PHRED'] = phred_scores
ofile.write('{}'.format(record).encode())
def merge_coverage_files(coverage_files, out_coverage_file):
with pysam.BGZFile(out_coverage_file, 'w') as oz:
overlap_head = deque([])
overlap_tail = deque([])
for coverage_file in coverage_files:
last_position = None
with gzip.GzipFile(coverage_file['name']) as iz:
for line in iz:
fields = line.split('\t')
chrom = fields[0]
if chrom != coverage_file['chrom']:
raise Exception(
'Multiple chromosomes detected within {} coverage file!'
.format(coverage_file['name']))
position = long(fields[1])
if last_position is None or last_position < position:
last_position = position
else:
raise Exception(
'Positions within {} coverage file are not in ascending order or not unique!'
.format(coverage_file['name']))
while overlap_head:
(overlap_position, overlap_line) = overlap_head[0]
if overlap_position >= coverage_file[
'next_leftmost_position']:
raise Exception(
"Overlapping regions are present in more than two coverage files!"
)
if overlap_position < position:
oz.write(overlap_line)
overlap_head.popleft()
elif overlap_position == position:
overlap_head.popleft()
overlap_data = json.loads(
overlap_line.split('\t')[2])
data = json.loads(fields[2])
if overlap_data['mean'] > data['mean']:
line = overlap_line
else:
break
else:
break
if position < coverage_file['next_leftmost_position']:
oz.write(line)
else:
overlap_tail.append((position, line))
if overlap_head:
raise Exception(
"Nested regions detected in two coverage files!")
overlap_head = overlap_tail
overlap_tail = deque([])
if overlap_tail:
raise Exception("Error while merging coverage files!")
def add_percentiles(in_VCF, in_pctl_files, out_VCF):
percentiles = [Percentiles(x) for x in in_pctl_files]
with pysam.VariantFile(in_VCF, 'r') as ifile, pysam.BGZFile(out_VCF, 'w') as ofile:
for p in percentiles:
for desc in p.descriptions():
ifile.header.add_line(desc)
ofile.write('{}'.format(ifile.header))
for record in ifile:
for p in percentiles:
for key, values in p.get(record.chrom, record.pos, record.ref, record.alts):
record.info[key] = values
ofile.write('{}'.format(record))
def prune(in_coverage_file, out_coverage_file, fluctuation_limit=0.25):
with gzip.GzipFile(in_coverage_file,
'r') as iz, pysam.BGZFile(out_coverage_file, 'w') as oz:
fields = iz.readline().split('\t')
bin_data = json.loads(fields[2])
for line in iz:
fields = line.split('\t')
data = json.loads(fields[2])
if ((abs(bin_data['mean'] - data['mean']) > fluctuation_limit)
or (abs(bin_data['median'] - data['median']) >
fluctuation_limit)):
write_data(oz, bin_data)
bin_data = data
bin_data['end'] = data['start']
write_data(oz, bin_data)
def write_species_specific_vcf_gzip(self):
vcf = VCF()
samples = vcf.get_sample_names(self.gvcf_file)
vcf_header = vcf.get_vcf_header(self.gvcf_file)
total_nb = 0
specific_snv = 0
last_chr = None
with gzip.open(self.gvcf_file, 'rt') as vcf_reader, pysam.BGZFile(
self.output_vcf,
'wb') as output_writer, open(self.log_file, 'w') as log_writer:
output_writer.write(vcf_header.encode('utf-8'))
for line in vcf_reader:
line = line.rstrip()
if not line.startswith('#'):
snp = SNP(line, samples)
specific_snv, total_nb = self.__log(
snp.chrom, specific_snv, total_nb, log_writer, False)
if last_chr != snp.chrom:
if last_chr is not None:
specific_snv, total_nb = self.__log(
snp.chrom, specific_snv, total_nb, log_writer,
True)
last_chr = snp.chrom
if self.__is_not_INDEL(snp):
if not self.__is_variant_filtered(
snp): # general variant quality filter
if self.__is_variant_specific(
snp, self.samples_interest_names
): # only species specific ones are kept
output_writer.write(
(snp.vcf_line + "\n").encode('utf-8'))
specific_snv += 1
last_chr = snp.chrom
total_nb += 1
log_writer.write(
"Chromosome:\t{}\t\tSpecific/Total:\t{}/{}\n".format(
last_chr, specific_snv, total_nb))
log_writer.close()
output_writer.close()
vcf_reader.close()
dest='out_coverage_file',
required=True,
help='Output JSON coverage (compressed with bgzip) file')
def write(output_file, data):
output_file.write('{}\t{:d}\t{:d}\t'.format(data['chrom'], data['start'],
data['end']).encode())
rapidjson.dump(data, output_file)
output_file.write('\n'.encode())
if __name__ == '__main__':
args = argparser.parse_args()
with gzip.open(args.in_coverage_file,
'rt') as ifile, pysam.BGZFile(args.out_coverage_file,
'w') as ofile:
line = ifile.readline()
if not line:
sys.exit(0)
chrom, start, stop, data = line.rstrip().split('\t')
bin_data = rapidjson.loads(data)
for line in ifile:
chrom, start, stop, data = line.rstrip().split('\t')
data = rapidjson.loads(data)
if (abs(bin_data['mean'] - data['mean']) > args.fluctuation_limit
) or (abs(bin_data['median'] - data['median']) >
args.fluctuation_limit):
write(ofile, bin_data)
bin_data = data
else:
bin_data['end'] = data['end']
'Output file of depth information compressed with bgzip. In addition to this file, the tabix index will be produced.'
)
if __name__ == '__main__':
args = argparser.parse_args()
file_names = []
with open(args.in_files_list, 'r') as ifile:
for line in ifile:
line = line.strip()
if line:
file_names.append(line)
chromosomes = set()
positions = dict()
n_indv = len(file_names)
breaks = [1, 5, 10, 15, 20, 25, 30, 50, 100]
with ExitStack() as stack, pysam.BGZFile(args.out_file_name, 'w') as ofile:
ifiles = [
stack.enter_context(gzip.open(file_name, 'rt'))
for file_name in file_names
]
while True:
for i, ifile in enumerate(ifiles):
line = ifile.readline()
if line:
chromosome, position, dp = line.rstrip().split()
chromosomes.add(chromosome)
if len(chromosomes) > 1:
raise Exception(
f'Multiple chromosomes detected in input files, but only one is allowed.'
)
positions.setdefault(int(position), []).append(int(dp))
