def get_coverage(output_dirpath, ref_fpath, ref_name, bam_fpath,
bam_sorted_fpath, log_path, err_path, cov_fpath,
physical_cov_fpath):
raw_cov_fpath = cov_fpath + '_raw'
if not is_non_empty_file(cov_fpath):
logger.info(' Calculating reads coverage...')
if not is_non_empty_file(raw_cov_fpath):
if not is_non_empty_file(bam_sorted_fpath):
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'sort', '-t',
str(qconfig.max_threads), '-o', bam_sorted_fpath, bam_fpath
],
stdout=open(log_path, 'a'),
stderr=open(err_path, 'a'))
chr_len_fpath = get_chr_len_fpath(ref_fpath)
qutils.call_subprocess([
bedtools_fpath('bedtools'), 'genomecov', '-bga', '-ibam',
bam_sorted_fpath, '-g', chr_len_fpath
],
stdout=open(raw_cov_fpath, 'w'),
stderr=open(err_path, 'a'))
qutils.assert_file_exists(raw_cov_fpath, 'coverage file')
proceed_cov_file(raw_cov_fpath, cov_fpath)
if not is_non_empty_file(physical_cov_fpath):
physical_cov_fpath = get_physical_coverage(output_dirpath, ref_fpath,
ref_name, bam_fpath,
log_path, err_path,
physical_cov_fpath)
return cov_fpath, physical_cov_fpath
def process_one_ref(cur_ref_fpath, output_dirpath, err_path, bed_fpath=None):
ref = qutils.name_from_fpath(cur_ref_fpath)
ref_sam_fpath = os.path.join(output_dirpath, ref + '.sam')
ref_bam_fpath = os.path.join(output_dirpath, ref + '.bam')
ref_bamsorted_fpath = os.path.join(output_dirpath, ref + '.sorted')
ref_bed_fpath = bed_fpath if bed_fpath else os.path.join(output_dirpath, ref + '.bed')
if os.path.getsize(ref_sam_fpath) < 1024 * 1024: # TODO: make it better (small files will cause Manta crush -- "not enough reads...")
logger.info(' SAM file is too small for Manta (%d Kb), skipping..' % (os.path.getsize(ref_sam_fpath) // 1024))
return None
if is_non_empty_file(ref_bed_fpath):
logger.info(' Using existing Manta BED-file: ' + ref_bed_fpath)
return ref_bed_fpath
if not os.path.exists(ref_bamsorted_fpath + '.bam'):
qutils.call_subprocess([samtools_fpath('samtools'), 'view', '-bS', ref_sam_fpath], stdout=open(ref_bam_fpath, 'w'),
stderr=open(err_path, 'a'), logger=logger)
qutils.call_subprocess([samtools_fpath('samtools'), 'sort', ref_bam_fpath, ref_bamsorted_fpath],
stderr=open(err_path, 'a'), logger=logger)
if not is_non_empty_file(ref_bamsorted_fpath + '.bam.bai'):
qutils.call_subprocess([samtools_fpath('samtools'), 'index', ref_bamsorted_fpath + '.bam'],
stderr=open(err_path, 'a'), logger=logger)
if not is_non_empty_file(cur_ref_fpath + '.fai'):
qutils.call_subprocess([samtools_fpath('samtools'), 'faidx', cur_ref_fpath],
stderr=open(err_path, 'a'), logger=logger)
vcfoutput_dirpath = os.path.join(output_dirpath, ref + '_manta')
found_SV_fpath = os.path.join(vcfoutput_dirpath, 'results/variants/diploidSV.vcf.gz')
unpacked_SV_fpath = found_SV_fpath + '.unpacked'
if not is_non_empty_file(found_SV_fpath):
if os.path.exists(vcfoutput_dirpath):
shutil.rmtree(vcfoutput_dirpath, ignore_errors=True)
os.makedirs(vcfoutput_dirpath)
qutils.call_subprocess([config_manta_fpath, '--normalBam', ref_bamsorted_fpath + '.bam',
'--referenceFasta', cur_ref_fpath, '--runDir', vcfoutput_dirpath],
stdout=open(err_path, 'a'), stderr=open(err_path, 'a'), logger=logger)
if not os.path.exists(os.path.join(vcfoutput_dirpath, 'runWorkflow.py')):
return None
qutils.call_subprocess([os.path.join(vcfoutput_dirpath, 'runWorkflow.py'), '-m', 'local', '-j', str(qconfig.max_threads)],
stderr=open(err_path, 'a'), logger=logger)
if not is_non_empty_file(unpacked_SV_fpath):
cmd = 'gunzip -c %s' % found_SV_fpath
qutils.call_subprocess(shlex.split(cmd), stdout=open(unpacked_SV_fpath, 'w'),
stderr=open(err_path, 'a'), logger=logger)
from manta import vcfToBedpe
vcfToBedpe.vcfToBedpe(open(unpacked_SV_fpath), open(ref_bed_fpath, 'w'))
return ref_bed_fpath
def get_physical_coverage(output_dirpath, ref_fpath, ref_name, bam_fpath,
log_path, err_path, cov_fpath):
if not os.path.exists(bedtools_fpath('bamToBed')):
logger.info(' Failed calculating physical coverage...')
return None
if not is_non_empty_file(cov_fpath):
logger.info(' Calculating physical coverage...')
## keep properly mapped, unique, and non-duplicate read pairs only
bam_filtered_fpath = os.path.join(output_dirpath,
ref_name + '.filtered.bam')
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'view', '-t',
str(qconfig.max_threads), '-h', '-f', 'bam', '-F',
'proper_pair and not supplementary and not duplicate', bam_fpath
],
stdout=open(bam_filtered_fpath, 'w'),
stderr=open(err_path, 'a'))
## sort by read names
bam_filtered_sorted_fpath = os.path.join(
output_dirpath, ref_name + '.filtered.sorted.bam')
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'sort', '-t',
str(qconfig.max_threads), '-n', '-o', bam_filtered_sorted_fpath,
bam_filtered_fpath
],
stdout=open(log_path, 'a'),
stderr=open(err_path, 'a'))
bedpe_fpath = os.path.join(output_dirpath, ref_name + '.bedpe')
qutils.call_subprocess([
bedtools_fpath('bamToBed'), '-i', bam_filtered_sorted_fpath,
'-bedpe'
],
stdout=open(bedpe_fpath, 'w'),
stderr=open(err_path, 'a'))
raw_bed_fpath = os.path.join(output_dirpath, ref_name + '.bed')
with open(bedpe_fpath, 'r') as bedpe:
with open(raw_bed_fpath, 'w') as bed_file:
for line in bedpe:
fs = line.split()
bed_file.write('\t'.join([fs[0], fs[1], fs[5] + '\n']))
sorted_bed_fpath = os.path.join(output_dirpath,
ref_name + '.sorted.bed')
qutils.call_subprocess(
[bedtools_fpath('bedtools'), 'sort', '-i', raw_bed_fpath],
stdout=open(sorted_bed_fpath, 'w'),
stderr=open(err_path, 'a'))
chr_len_fpath = get_chr_len_fpath(ref_fpath)
raw_cov_fpath = cov_fpath + '_raw'
qutils.call_subprocess([
bedtools_fpath('bedtools'), 'genomecov', '-bga', '-i',
sorted_bed_fpath, '-g', chr_len_fpath
],
stdout=open(raw_cov_fpath, 'w'),
stderr=open(err_path, 'a'))
proceed_cov_file(raw_cov_fpath, cov_fpath)
return cov_fpath
def run_processing_reads(main_ref_fpath, meta_ref_fpaths, ref_labels,
reads_fpaths, output_dirpath, res_path, log_path,
err_path):
ref_name = qutils.name_from_fpath(main_ref_fpath)
sam_fpath = os.path.join(output_dirpath, ref_name + '.sam')
bam_fpath = os.path.join(output_dirpath, ref_name + '.bam')
bam_sorted_fpath = os.path.join(output_dirpath, ref_name + '.sorted')
sam_sorted_fpath = os.path.join(output_dirpath, ref_name + '.sorted.sam')
bed_fpath = os.path.join(res_path, ref_name + '.bed')
if is_non_empty_file(bed_fpath):
logger.info(' Using existing BED-file: ' + bed_fpath)
return bed_fpath
logger.info(' ' + 'Pre-processing for searching structural variations...')
logger.info(' ' + 'Logging to %s...' % err_path)
if is_non_empty_file(sam_fpath):
logger.info(' Using existing SAM-file: ' + sam_fpath)
else:
logger.info(' Running Bowtie2...')
abs_reads_fpaths = [
] # use absolute paths because we will change workdir
for reads_fpath in reads_fpaths:
abs_reads_fpaths.append(os.path.abspath(reads_fpath))
prev_dir = os.getcwd()
os.chdir(output_dirpath)
cmd = [bin_fpath('bowtie2-build'), main_ref_fpath, ref_name]
qutils.call_subprocess(cmd,
stdout=open(log_path, 'a'),
stderr=open(err_path, 'a'),
logger=logger)
cmd = bin_fpath('bowtie2') + ' -x ' + ref_name + ' -1 ' + abs_reads_fpaths[0] + ' -2 ' + abs_reads_fpaths[1] + ' -S ' + \
sam_fpath + ' --no-unal -a -p %s' % str(qconfig.max_threads)
qutils.call_subprocess(shlex.split(cmd),
stdout=open(log_path, 'a'),
stderr=open(err_path, 'a'),
logger=logger)
logger.info(' Done.')
os.chdir(prev_dir)
if not os.path.exists(sam_fpath) or os.path.getsize(sam_fpath) == 0:
logger.error(' Failed running Bowtie2 for the reference. See ' +
log_path + ' for information.')
logger.info(' Failed searching structural variations.')
return None
logger.info(' Sorting SAM-file...')
if is_non_empty_file(sam_sorted_fpath):
logger.info(' Using existing sorted SAM-file: ' + sam_sorted_fpath)
else:
qutils.call_subprocess([
samtools_fpath('samtools'), 'view', '-@',
str(qconfig.max_threads), '-bS', sam_fpath
],
stdout=open(bam_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
qutils.call_subprocess([
samtools_fpath('samtools'), 'sort', '-@',
str(qconfig.max_threads), bam_fpath, bam_sorted_fpath
],
stderr=open(err_path, 'a'),
logger=logger)
qutils.call_subprocess([
samtools_fpath('samtools'), 'view', '-@',
str(qconfig.max_threads), bam_sorted_fpath + '.bam'
],
stdout=open(sam_sorted_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
if meta_ref_fpaths:
logger.info(' Splitting SAM-file by references...')
headers = []
seq_name_length = {}
with open(sam_fpath) as sam_file:
for line in sam_file:
if not line.startswith('@'):
break
if line.startswith('@SQ') and 'SN:' in line and 'LN:' in line:
seq_name = line.split('\tSN:')[1].split('\t')[0]
seq_length = int(line.split('\tLN:')[1].split('\t')[0])
seq_name_length[seq_name] = seq_length
headers.append(line.strip())
need_ref_splitting = False
if meta_ref_fpaths:
ref_files = {}
for cur_ref_fpath in meta_ref_fpaths:
ref = qutils.name_from_fpath(cur_ref_fpath)
new_ref_sam_fpath = os.path.join(output_dirpath, ref + '.sam')
if is_non_empty_file(new_ref_sam_fpath):
logger.info(' Using existing split SAM-file for %s: %s' %
(ref, new_ref_sam_fpath))
ref_files[ref] = None
else:
new_ref_sam_file = open(new_ref_sam_fpath, 'w')
new_ref_sam_file.write(headers[0] + '\n')
chrs = []
for h in (h for h in headers
if h.startswith('@SQ') and 'SN:' in h):
seq_name = h.split('\tSN:')[1].split('\t')[0]
if seq_name in ref_labels and ref_labels[seq_name] == ref:
new_ref_sam_file.write(h + '\n')
chrs.append(seq_name)
new_ref_sam_file.write(headers[-1] + '\n')
ref_files[ref] = new_ref_sam_file
need_ref_splitting = True
deletions = []
trivial_deletions_fpath = os.path.join(output_dirpath,
qconfig.trivial_deletions_fname)
logger.info(
' Looking for trivial deletions (long zero-covered fragments)...')
need_trivial_deletions = True
if os.path.exists(trivial_deletions_fpath):
need_trivial_deletions = False
logger.info(' Using existing file: ' + trivial_deletions_fpath)
if need_trivial_deletions or need_ref_splitting:
with open(sam_sorted_fpath) as sam_file:
cur_deletion = None
for line in sam_file:
mapping = Mapping.parse(line)
if mapping:
# common case: continue current deletion (potential) on the same reference
if cur_deletion and cur_deletion.ref == mapping.ref:
if cur_deletion.next_bad is None: # previous mapping was in region BEFORE 0-covered fragment
# just passed 0-covered fragment
if mapping.start - cur_deletion.prev_bad > QuastDeletion.MIN_GAP:
cur_deletion.set_next_bad(mapping)
if mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_next_good(mapping)
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_good(mapping)
# continue region BEFORE 0-covered fragment
elif mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_prev_good(mapping)
else:
cur_deletion.set_prev_bad(mapping)
else: # previous mapping was in region AFTER 0-covered fragment
# just passed another 0-cov fragment between end of cur_deletion BAD region and this mapping
if mapping.start - cur_deletion.next_bad_end > QuastDeletion.MIN_GAP:
if cur_deletion.is_valid(
): # add previous fragment's deletion if needed
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_bad(
position=cur_deletion.next_bad_end)
# continue region AFTER 0-covered fragment (old one or new/another one -- see "if" above)
if mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_next_good(mapping)
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_good(mapping)
else:
cur_deletion.set_next_bad_end(mapping)
# special case: just started or just switched to the next reference
else:
if cur_deletion and cur_deletion.ref in seq_name_length: # switched to the next ref
cur_deletion.set_next_good(
position=seq_name_length[cur_deletion.ref])
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_good(mapping)
if need_ref_splitting:
cur_ref = ref_labels[mapping.ref]
if mapping.ref_next.strip(
) == '=' or cur_ref == ref_labels[mapping.ref_next]:
if ref_files[cur_ref] is not None:
ref_files[cur_ref].write(line)
if cur_deletion and cur_deletion.ref in seq_name_length: # switched to the next ref
cur_deletion.set_next_good(
position=seq_name_length[cur_deletion.ref])
if cur_deletion.is_valid():
deletions.append(cur_deletion)
if need_ref_splitting:
for ref_handler in ref_files.values():
if ref_handler is not None:
ref_handler.close()
if need_trivial_deletions:
logger.info(' Trivial deletions: %d found' % len(deletions))
logger.info(' Saving to: ' + trivial_deletions_fpath)
with open(trivial_deletions_fpath, 'w') as f:
for deletion in deletions:
f.write(str(deletion) + '\n')
if os.path.exists(config_manta_fpath):
manta_sv_fpath = search_sv_with_manta(main_ref_fpath, meta_ref_fpaths,
output_dirpath, err_path)
qutils.cat_files([manta_sv_fpath, trivial_deletions_fpath], bed_fpath)
elif os.path.exists(trivial_deletions_fpath):
shutil.copy(trivial_deletions_fpath, bed_fpath)
if os.path.exists(bed_fpath):
logger.main_info(' Structural variations saved to ' + bed_fpath)
return bed_fpath
else:
logger.main_info(' Failed searching structural variations.')
return None
def process_one_ref(cur_ref_fpath, output_dirpath, err_path, bed_fpath=None):
ref = qutils.name_from_fpath(cur_ref_fpath)
ref_sam_fpath = os.path.join(output_dirpath, ref + '.sam')
ref_bam_fpath = os.path.join(output_dirpath, ref + '.bam')
ref_bamsorted_fpath = os.path.join(output_dirpath, ref + '.sorted')
ref_bed_fpath = bed_fpath if bed_fpath else os.path.join(
output_dirpath, ref + '.bed')
if os.path.getsize(
ref_sam_fpath
) < 1024 * 1024: # TODO: make it better (small files will cause Manta crush -- "not enough reads...")
logger.info(' SAM file is too small for Manta (%d Kb), skipping..' %
(os.path.getsize(ref_sam_fpath) // 1024))
return None
if is_non_empty_file(ref_bed_fpath):
logger.info(' Using existing Manta BED-file: ' + ref_bed_fpath)
return ref_bed_fpath
if not os.path.exists(ref_bamsorted_fpath + '.bam'):
qutils.call_subprocess(
[samtools_fpath('samtools'), 'view', '-bS', ref_sam_fpath],
stdout=open(ref_bam_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
qutils.call_subprocess([
samtools_fpath('samtools'), 'sort', ref_bam_fpath,
ref_bamsorted_fpath
],
stderr=open(err_path, 'a'),
logger=logger)
if not is_non_empty_file(ref_bamsorted_fpath + '.bam.bai'):
qutils.call_subprocess([
samtools_fpath('samtools'), 'index', ref_bamsorted_fpath + '.bam'
],
stderr=open(err_path, 'a'),
logger=logger)
if not is_non_empty_file(cur_ref_fpath + '.fai'):
qutils.call_subprocess(
[samtools_fpath('samtools'), 'faidx', cur_ref_fpath],
stderr=open(err_path, 'a'),
logger=logger)
vcfoutput_dirpath = os.path.join(output_dirpath, ref + '_manta')
found_SV_fpath = os.path.join(vcfoutput_dirpath,
'results/variants/diploidSV.vcf.gz')
unpacked_SV_fpath = found_SV_fpath + '.unpacked'
if not is_non_empty_file(found_SV_fpath):
if os.path.exists(vcfoutput_dirpath):
shutil.rmtree(vcfoutput_dirpath, ignore_errors=True)
os.makedirs(vcfoutput_dirpath)
qutils.call_subprocess([
config_manta_fpath, '--normalBam', ref_bamsorted_fpath + '.bam',
'--referenceFasta', cur_ref_fpath, '--runDir', vcfoutput_dirpath
],
stdout=open(err_path, 'a'),
stderr=open(err_path, 'a'),
logger=logger)
if not os.path.exists(os.path.join(vcfoutput_dirpath,
'runWorkflow.py')):
return None
qutils.call_subprocess([
os.path.join(vcfoutput_dirpath, 'runWorkflow.py'), '-m', 'local',
'-j',
str(qconfig.max_threads)
],
stderr=open(err_path, 'a'),
logger=logger)
if not is_non_empty_file(unpacked_SV_fpath):
cmd = 'gunzip -c %s' % found_SV_fpath
qutils.call_subprocess(shlex.split(cmd),
stdout=open(unpacked_SV_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
from manta import vcfToBedpe
vcfToBedpe.vcfToBedpe(open(unpacked_SV_fpath), open(ref_bed_fpath, 'w'))
return ref_bed_fpath
def run_processing_reads(main_ref_fpath, meta_ref_fpaths, ref_labels, reads_fpaths, output_dirpath, res_path, log_path, err_path):
ref_name = qutils.name_from_fpath(main_ref_fpath)
sam_fpath = os.path.join(output_dirpath, ref_name + '.sam')
bam_fpath = os.path.join(output_dirpath, ref_name + '.bam')
bam_sorted_fpath = os.path.join(output_dirpath, ref_name + '.sorted')
sam_sorted_fpath = os.path.join(output_dirpath, ref_name + '.sorted.sam')
bed_fpath = os.path.join(res_path, ref_name + '.bed')
if is_non_empty_file(bed_fpath):
logger.info(' Using existing BED-file: ' + bed_fpath)
return bed_fpath
logger.info(' ' + 'Pre-processing for searching structural variations...')
logger.info(' ' + 'Logging to %s...' % err_path)
if is_non_empty_file(sam_fpath):
logger.info(' Using existing SAM-file: ' + sam_fpath)
else:
logger.info(' Running Bowtie2...')
abs_reads_fpaths = [] # use absolute paths because we will change workdir
for reads_fpath in reads_fpaths:
abs_reads_fpaths.append(os.path.abspath(reads_fpath))
prev_dir = os.getcwd()
os.chdir(output_dirpath)
cmd = [bin_fpath('bowtie2-build'), main_ref_fpath, ref_name]
qutils.call_subprocess(cmd, stdout=open(log_path, 'a'), stderr=open(err_path, 'a'), logger=logger)
cmd = bin_fpath('bowtie2') + ' -x ' + ref_name + ' -1 ' + abs_reads_fpaths[0] + ' -2 ' + abs_reads_fpaths[1] + ' -S ' + \
sam_fpath + ' --no-unal -a -p %s' % str(qconfig.max_threads)
qutils.call_subprocess(shlex.split(cmd), stdout=open(log_path, 'a'), stderr=open(err_path, 'a'), logger=logger)
logger.info(' Done.')
os.chdir(prev_dir)
if not os.path.exists(sam_fpath) or os.path.getsize(sam_fpath) == 0:
logger.error(' Failed running Bowtie2 for the reference. See ' + log_path + ' for information.')
logger.info(' Failed searching structural variations.')
return None
logger.info(' Sorting SAM-file...')
if is_non_empty_file(sam_sorted_fpath):
logger.info(' Using existing sorted SAM-file: ' + sam_sorted_fpath)
else:
qutils.call_subprocess([samtools_fpath('samtools'), 'view', '-@', str(qconfig.max_threads), '-bS', sam_fpath], stdout=open(bam_fpath, 'w'),
stderr=open(err_path, 'a'), logger=logger)
qutils.call_subprocess([samtools_fpath('samtools'), 'sort', '-@', str(qconfig.max_threads), bam_fpath, bam_sorted_fpath],
stderr=open(err_path, 'a'), logger=logger)
qutils.call_subprocess([samtools_fpath('samtools'), 'view', '-@', str(qconfig.max_threads), bam_sorted_fpath + '.bam'], stdout=open(sam_sorted_fpath, 'w'),
stderr=open(err_path, 'a'), logger=logger)
if meta_ref_fpaths:
logger.info(' Splitting SAM-file by references...')
headers = []
seq_name_length = {}
with open(sam_fpath) as sam_file:
for line in sam_file:
if not line.startswith('@'):
break
if line.startswith('@SQ') and 'SN:' in line and 'LN:' in line:
seq_name = line.split('\tSN:')[1].split('\t')[0]
seq_length = int(line.split('\tLN:')[1].split('\t')[0])
seq_name_length[seq_name] = seq_length
headers.append(line.strip())
need_ref_splitting = False
if meta_ref_fpaths:
ref_files = {}
for cur_ref_fpath in meta_ref_fpaths:
ref = qutils.name_from_fpath(cur_ref_fpath)
new_ref_sam_fpath = os.path.join(output_dirpath, ref + '.sam')
if is_non_empty_file(new_ref_sam_fpath):
logger.info(' Using existing split SAM-file for %s: %s' % (ref, new_ref_sam_fpath))
ref_files[ref] = None
else:
new_ref_sam_file = open(new_ref_sam_fpath, 'w')
new_ref_sam_file.write(headers[0] + '\n')
chrs = []
for h in (h for h in headers if h.startswith('@SQ') and 'SN:' in h):
seq_name = h.split('\tSN:')[1].split('\t')[0]
if seq_name in ref_labels and ref_labels[seq_name] == ref:
new_ref_sam_file.write(h + '\n')
chrs.append(seq_name)
new_ref_sam_file.write(headers[-1] + '\n')
ref_files[ref] = new_ref_sam_file
need_ref_splitting = True
deletions = []
trivial_deletions_fpath = os.path.join(output_dirpath, qconfig.trivial_deletions_fname)
logger.info(' Looking for trivial deletions (long zero-covered fragments)...')
need_trivial_deletions = True
if os.path.exists(trivial_deletions_fpath):
need_trivial_deletions = False
logger.info(' Using existing file: ' + trivial_deletions_fpath)
if need_trivial_deletions or need_ref_splitting:
with open(sam_sorted_fpath) as sam_file:
cur_deletion = None
for line in sam_file:
mapping = Mapping.parse(line)
if mapping:
# common case: continue current deletion (potential) on the same reference
if cur_deletion and cur_deletion.ref == mapping.ref:
if cur_deletion.next_bad is None: # previous mapping was in region BEFORE 0-covered fragment
# just passed 0-covered fragment
if mapping.start - cur_deletion.prev_bad > QuastDeletion.MIN_GAP:
cur_deletion.set_next_bad(mapping)
if mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_next_good(mapping)
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(mapping.ref).set_prev_good(mapping)
# continue region BEFORE 0-covered fragment
elif mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_prev_good(mapping)
else:
cur_deletion.set_prev_bad(mapping)
else: # previous mapping was in region AFTER 0-covered fragment
# just passed another 0-cov fragment between end of cur_deletion BAD region and this mapping
if mapping.start - cur_deletion.next_bad_end > QuastDeletion.MIN_GAP:
if cur_deletion.is_valid(): # add previous fragment's deletion if needed
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(mapping.ref).set_prev_bad(position=cur_deletion.next_bad_end)
# continue region AFTER 0-covered fragment (old one or new/another one -- see "if" above)
if mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_next_good(mapping)
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(mapping.ref).set_prev_good(mapping)
else:
cur_deletion.set_next_bad_end(mapping)
# special case: just started or just switched to the next reference
else:
if cur_deletion and cur_deletion.ref in seq_name_length: # switched to the next ref
cur_deletion.set_next_good(position=seq_name_length[cur_deletion.ref])
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(mapping.ref).set_prev_good(mapping)
if need_ref_splitting:
cur_ref = ref_labels[mapping.ref]
if mapping.ref_next.strip() == '=' or cur_ref == ref_labels[mapping.ref_next]:
if ref_files[cur_ref] is not None:
ref_files[cur_ref].write(line)
if cur_deletion and cur_deletion.ref in seq_name_length: # switched to the next ref
cur_deletion.set_next_good(position=seq_name_length[cur_deletion.ref])
if cur_deletion.is_valid():
deletions.append(cur_deletion)
if need_ref_splitting:
for ref_handler in ref_files.values():
if ref_handler is not None:
ref_handler.close()
if need_trivial_deletions:
logger.info(' Trivial deletions: %d found' % len(deletions))
logger.info(' Saving to: ' + trivial_deletions_fpath)
with open(trivial_deletions_fpath, 'w') as f:
for deletion in deletions:
f.write(str(deletion) + '\n')
if os.path.exists(config_manta_fpath):
manta_sv_fpath = search_sv_with_manta(main_ref_fpath, meta_ref_fpaths, output_dirpath, err_path)
qutils.cat_files([manta_sv_fpath, trivial_deletions_fpath], bed_fpath)
elif os.path.exists(trivial_deletions_fpath):
shutil.copy(trivial_deletions_fpath, bed_fpath)
if os.path.exists(bed_fpath):
logger.main_info(' Structural variations saved to ' + bed_fpath)
return bed_fpath
else:
logger.main_info(' Failed searching structural variations.')
return None
def run_processing_reads(main_ref_fpath,
meta_ref_fpaths,
ref_labels,
reads_fpaths,
output_dirpath,
res_path,
log_path,
err_path,
sam_fpath=None,
bam_fpath=None,
bed_fpath=None):
ref_name = qutils.name_from_fpath(main_ref_fpath)
if not sam_fpath and bam_fpath:
sam_fpath = get_safe_fpath(output_dirpath, bam_fpath[:-4] + '.sam')
else:
sam_fpath = sam_fpath or os.path.join(output_dirpath,
ref_name + '.sam')
bam_fpath = bam_fpath or get_safe_fpath(output_dirpath,
sam_fpath[:-4] + '.bam')
sam_sorted_fpath = get_safe_fpath(output_dirpath,
add_suffix(sam_fpath, 'sorted'))
bam_sorted_fpath = get_safe_fpath(output_dirpath,
add_suffix(bam_fpath, 'sorted'))
bed_fpath = bed_fpath or os.path.join(res_path, ref_name + '.bed')
cov_fpath = os.path.join(res_path, ref_name + '.cov')
physical_cov_fpath = os.path.join(res_path, ref_name + '.physical.cov')
if qconfig.no_sv:
logger.info(
' Will not search Structural Variations (--fast or --no-sv is specified)'
)
bed_fpath = None
elif is_non_empty_file(bed_fpath):
logger.info(' Using existing BED-file: ' + bed_fpath)
if is_non_empty_file(cov_fpath):
is_correct_file = check_cov_file(cov_fpath)
if is_correct_file:
logger.info(' Using existing reads coverage file: ' + cov_fpath)
if is_non_empty_file(physical_cov_fpath):
logger.info(' Using existing physical coverage file: ' +
physical_cov_fpath)
if (is_non_empty_file(bed_fpath) or qconfig.no_sv) and is_non_empty_file(
cov_fpath) and is_non_empty_file(physical_cov_fpath):
return bed_fpath, cov_fpath, physical_cov_fpath
logger.info(' ' + 'Pre-processing reads...')
logger.info(' ' + 'Logging to %s...' % err_path)
if is_non_empty_file(sam_fpath):
logger.info(' Using existing SAM-file: ' + sam_fpath)
elif is_non_empty_file(bam_fpath):
logger.info(' Using existing BAM-file: ' + bam_fpath)
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'view', '-t',
str(qconfig.max_threads), '-h', bam_fpath
],
stdout=open(sam_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
else:
logger.info(' Running BWA...')
# use absolute paths because we will change workdir
sam_fpath = os.path.abspath(sam_fpath)
abs_reads_fpaths = []
for reads_fpath in reads_fpaths:
abs_reads_fpaths.append(os.path.abspath(reads_fpath))
if len(abs_reads_fpaths) != 2:
logger.error(
' You should specify files with forward and reverse reads.')
logger.info(' Failed searching structural variations.')
return None, None, None
if not qconfig.no_check:
if not paired_reads_names_are_equal(reads_fpaths, logger):
logger.info(
' Read names are discordant, skipping reads analysis!')
return None, None, None
prev_dir = os.getcwd()
os.chdir(output_dirpath)
cmd = [bwa_fpath('bwa'), 'index', '-p', ref_name, main_ref_fpath]
if os.path.getsize(
main_ref_fpath
) > 2 * 1024**3: # if reference size bigger than 2GB
cmd += ['-a', 'bwtsw']
qutils.call_subprocess(cmd,
stdout=open(log_path, 'a'),
stderr=open(err_path, 'a'),
logger=logger)
cmd = bwa_fpath('bwa') + ' mem -t ' + str(
qconfig.max_threads) + ' ' + ref_name + ' ' + abs_reads_fpaths[
0] + ' ' + abs_reads_fpaths[1]
qutils.call_subprocess(shlex.split(cmd),
stdout=open(sam_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
logger.info(' Done.')
os.chdir(prev_dir)
if not os.path.exists(sam_fpath) or os.path.getsize(sam_fpath) == 0:
logger.error(' Failed running BWA for the reference. See ' +
log_path + ' for information.')
logger.info(' Failed searching structural variations.')
return None, None, None
logger.info(' Sorting SAM-file...')
if (is_non_empty_file(sam_sorted_fpath)
and all_read_names_correct(sam_sorted_fpath)
) and is_non_empty_file(bam_fpath):
logger.info(' Using existing sorted SAM-file: ' + sam_sorted_fpath)
else:
correct_sam_fpath = os.path.join(output_dirpath, ref_name +
'.sam.correct') # write in output dir
clean_read_names(sam_fpath, correct_sam_fpath)
bam_fpath = os.path.join(output_dirpath, ref_name + '.bam')
bam_sorted_fpath = add_suffix(bam_fpath, 'sorted')
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'view', '-t',
str(qconfig.max_threads), '-h', '-f', 'bam', '-S',
correct_sam_fpath
],
stdout=open(bam_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'sort', '-t',
str(qconfig.max_threads), '-o', bam_sorted_fpath, bam_fpath
],
stderr=open(err_path, 'a'),
logger=logger)
qutils.call_subprocess([
sambamba_fpath('sambamba'), 'view', '-t',
str(qconfig.max_threads), '-h', bam_sorted_fpath
],
stdout=open(sam_sorted_fpath, 'w'),
stderr=open(err_path, 'a'),
logger=logger)
if not is_non_empty_file(cov_fpath) or not is_non_empty_file(
physical_cov_fpath):
cov_fpath, physical_cov_fpath = get_coverage(
output_dirpath, main_ref_fpath, ref_name, bam_fpath,
bam_sorted_fpath, log_path, err_path, cov_fpath,
physical_cov_fpath)
if not is_non_empty_file(bed_fpath) and not qconfig.no_sv:
if meta_ref_fpaths:
logger.info(' Splitting SAM-file by references...')
headers = []
seq_name_length = {}
with open(sam_fpath) as sam_file:
for line in sam_file:
if not line.startswith('@'):
break
if line.startswith('@SQ') and 'SN:' in line and 'LN:' in line:
seq_name = line.split('\tSN:')[1].split('\t')[0]
seq_length = int(line.split('\tLN:')[1].split('\t')[0])
seq_name_length[seq_name] = seq_length
headers.append(line.strip())
need_ref_splitting = False
if meta_ref_fpaths:
ref_files = {}
for cur_ref_fpath in meta_ref_fpaths:
ref = qutils.name_from_fpath(cur_ref_fpath)
new_ref_sam_fpath = os.path.join(output_dirpath, ref + '.sam')
if is_non_empty_file(new_ref_sam_fpath):
logger.info(
' Using existing split SAM-file for %s: %s' %
(ref, new_ref_sam_fpath))
ref_files[ref] = None
else:
new_ref_sam_file = open(new_ref_sam_fpath, 'w')
new_ref_sam_file.write(headers[0] + '\n')
chrs = []
for h in (h for h in headers
if h.startswith('@SQ') and 'SN:' in h):
seq_name = h.split('\tSN:')[1].split('\t')[0]
if seq_name in ref_labels and ref_labels[
seq_name] == ref:
new_ref_sam_file.write(h + '\n')
chrs.append(seq_name)
new_ref_sam_file.write(headers[-1] + '\n')
ref_files[ref] = new_ref_sam_file
need_ref_splitting = True
deletions = []
trivial_deletions_fpath = os.path.join(output_dirpath,
qconfig.trivial_deletions_fname)
logger.info(
' Looking for trivial deletions (long zero-covered fragments)...')
need_trivial_deletions = True
if os.path.exists(trivial_deletions_fpath):
need_trivial_deletions = False
logger.info(' Using existing file: ' + trivial_deletions_fpath)
if need_trivial_deletions or need_ref_splitting:
with open(sam_sorted_fpath) as sam_file:
cur_deletion = None
for line in sam_file:
mapping = Mapping.parse(line)
if mapping:
# common case: continue current deletion (potential) on the same reference
if cur_deletion and cur_deletion.ref == mapping.ref:
if cur_deletion.next_bad is None: # previous mapping was in region BEFORE 0-covered fragment
# just passed 0-covered fragment
if mapping.start - cur_deletion.prev_bad > QuastDeletion.MIN_GAP:
cur_deletion.set_next_bad(mapping)
if mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_next_good(mapping)
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_good(mapping)
# continue region BEFORE 0-covered fragment
elif mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_prev_good(mapping)
else:
cur_deletion.set_prev_bad(mapping)
else: # previous mapping was in region AFTER 0-covered fragment
# just passed another 0-cov fragment between end of cur_deletion BAD region and this mapping
if mapping.start - cur_deletion.next_bad_end > QuastDeletion.MIN_GAP:
if cur_deletion.is_valid(
): # add previous fragment's deletion if needed
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_bad(
position=cur_deletion.next_bad_end)
# continue region AFTER 0-covered fragment (old one or new/another one -- see "if" above)
if mapping.mapq >= Mapping.MIN_MAP_QUALITY:
cur_deletion.set_next_good(mapping)
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_good(mapping)
else:
cur_deletion.set_next_bad_end(mapping)
# special case: just started or just switched to the next reference
else:
if cur_deletion and cur_deletion.ref in seq_name_length: # switched to the next ref
cur_deletion.set_next_good(
position=seq_name_length[cur_deletion.ref])
if cur_deletion.is_valid():
deletions.append(cur_deletion)
cur_deletion = QuastDeletion(
mapping.ref).set_prev_good(mapping)
if need_ref_splitting:
cur_ref = ref_labels[mapping.ref]
if mapping.ref_next.strip(
) == '=' or cur_ref == ref_labels[
mapping.ref_next]:
if ref_files[cur_ref] is not None:
ref_files[cur_ref].write(line)
if cur_deletion and cur_deletion.ref in seq_name_length: # switched to the next ref
cur_deletion.set_next_good(
position=seq_name_length[cur_deletion.ref])
if cur_deletion.is_valid():
deletions.append(cur_deletion)
if need_ref_splitting:
for ref_handler in ref_files.values():
if ref_handler is not None:
ref_handler.close()
if need_trivial_deletions:
logger.info(' Trivial deletions: %d found' % len(deletions))
logger.info(' Saving to: ' + trivial_deletions_fpath)
with open(trivial_deletions_fpath, 'w') as f:
for deletion in deletions:
f.write(str(deletion) + '\n')
if isfile(config_manta_fpath):
try:
manta_sv_fpath = search_sv_with_manta(main_ref_fpath,
meta_ref_fpaths,
output_dirpath, err_path)
qutils.cat_files([manta_sv_fpath, trivial_deletions_fpath],
bed_fpath)
except:
pass
if os.path.exists(
trivial_deletions_fpath) and not is_non_empty_file(bed_fpath):
shutil.copy(trivial_deletions_fpath, bed_fpath)
if not qconfig.no_sv:
if is_non_empty_file(bed_fpath):
logger.main_info(' Structural variations are in ' + bed_fpath)
else:
if isfile(bed_fpath):
logger.main_info(' No structural variations were found.')
else:
logger.main_info(' Failed searching structural variations.')
bed_fpath = None
if is_non_empty_file(cov_fpath):
logger.main_info(
' Coverage distribution along the reference genome is in ' +
cov_fpath)
else:
logger.main_info(' Failed to calculate coverage distribution')
cov_fpath = None
return bed_fpath, cov_fpath, physical_cov_fpath
