def connect_with_matepairs(bam_fpath, output_dirpath, err_fpath):
bam_filtered_fpath = add_suffix(bam_fpath, 'filtered')
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_fpath, 'a'), logger=logger)
## sort by read names
bam_filtered_sorted_fpath = add_suffix(bam_filtered_fpath, 'sorted')
sort_bam(bam_filtered_fpath, bam_filtered_sorted_fpath, err_fpath, logger, sort_rule='-n')
bed_fpath = bam_to_bed(output_dirpath, 'matepairs', bam_filtered_sorted_fpath, err_fpath, logger, bedpe=True, only_intervals=True)
matepair_regions = defaultdict(list)
with open(bed_fpath) as bed:
for l in bed:
fs = l.split()
matepair_regions[fs[0]].append((int(fs[1]), int(fs[2])))
return matepair_regions
def get_physical_coverage(output_dirpath, ref_name, bam_fpath, log_path, err_fpath, cov_fpath, chr_len_fpath):
if not isfile(bedtools_fpath('bamToBed')):
logger.info(' Failed calculating physical coverage...')
return None
raw_cov_fpath = add_suffix(cov_fpath, 'raw')
if not is_non_empty_file(raw_cov_fpath):
logger.info(' Calculating physical coverage...')
## keep properly mapped, unique, and non-duplicate read pairs only
bam_filtered_fpath = join(output_dirpath, ref_name + '.filtered.bam')
sambamba_view(bam_fpath, bam_filtered_fpath, qconfig.max_threads, err_fpath, logger,
filter_rule='proper_pair and not supplementary and not duplicate')
## sort by read names
bam_filtered_sorted_fpath = join(output_dirpath, ref_name + '.filtered.sorted.bam')
sort_bam(bam_filtered_fpath, bam_filtered_sorted_fpath, err_fpath, logger, sort_rule='-n')
bed_fpath = bam_to_bed(output_dirpath, ref_name, bam_filtered_sorted_fpath, err_fpath, logger, bedpe=True)
calculate_genome_cov(bed_fpath, raw_cov_fpath, chr_len_fpath, err_fpath, logger)
return raw_cov_fpath
def get_physical_coverage(output_dirpath, ref_name, bam_fpath, log_path, err_fpath, cov_fpath, chr_len_fpath):
if not isfile(bedtools_fpath('bamToBed')):
logger.info(' Failed calculating physical coverage...')
return None
raw_cov_fpath = add_suffix(cov_fpath, 'raw')
if not is_non_empty_file(raw_cov_fpath):
logger.info(' Calculating physical coverage...')
## keep properly mapped, unique, and non-duplicate read pairs only
bam_filtered_fpath = join(output_dirpath, ref_name + '.filtered.bam')
sambamba_view(bam_fpath, bam_filtered_fpath, qconfig.max_threads, err_fpath, logger,
filter_rule='proper_pair and not supplementary and not duplicate')
## sort by read names
bam_filtered_sorted_fpath = join(output_dirpath, ref_name + '.filtered.sorted.bam')
sort_bam(bam_filtered_fpath, bam_filtered_sorted_fpath, err_fpath, logger, sort_rule='-n')
bed_fpath = bam_to_bed(output_dirpath, ref_name, bam_filtered_sorted_fpath, err_fpath, logger, bedpe=True)
calculate_genome_cov(bed_fpath, raw_cov_fpath, chr_len_fpath, err_fpath, logger)
return raw_cov_fpath
def get_joiners(ref_name, sam_fpath, bam_fpath, output_dirpath, err_fpath,
using_reads):
bam_filtered_fpath = add_suffix(bam_fpath, 'filtered')
if not is_non_empty_file(bam_filtered_fpath):
filter_rule = 'not unmapped and not supplementary and not secondary_alignment'
sambamba_view(bam_fpath,
bam_filtered_fpath,
qconfig.max_threads,
err_fpath,
logger,
filter_rule=filter_rule)
bam_sorted_fpath = add_suffix(bam_fpath, 'sorted')
if not is_non_empty_file(bam_sorted_fpath):
sort_bam(bam_filtered_fpath,
bam_sorted_fpath,
err_fpath,
logger,
sort_rule='-n')
bed_fpath = bam_to_bed(output_dirpath,
using_reads,
bam_sorted_fpath,
err_fpath,
logger,
bedpe=using_reads == 'mp')
intervals = defaultdict(list)
if using_reads == 'mp':
insert_size, std_dev = calculate_insert_size(sam_fpath,
output_dirpath,
ref_name,
reads_suffix='mp')
min_is = insert_size - std_dev
max_is = insert_size + std_dev
with open(bed_fpath) as bed:
for l in bed:
fs = l.split()
if using_reads == 'mp' and insert_size:
interval_len = int(fs[2]) - int(fs[1])
if min_is <= abs(interval_len) <= max_is:
intervals[fs[0]].append((int(fs[1]), int(fs[2])))
else:
intervals[fs[0]].append((int(fs[1]), int(fs[2])))
return intervals
def analyse_coverage(output_dirpath, fpath, chr_names, bam_fpath, stats_fpath, err_fpath, logger):
filename = qutils.name_from_fpath(fpath)
bed_fpath = bam_to_bed(output_dirpath, filename, bam_fpath, err_fpath, logger)
chr_len_fpath = get_chr_len_fpath(fpath, chr_names)
cov_fpath = join(output_dirpath, filename + '.genomecov')
calculate_genome_cov(bed_fpath, cov_fpath, chr_len_fpath, err_fpath, logger, print_all_positions=False)
avg_depth = 0
coverage_for_thresholds = [0 for threshold in qconfig.coverage_thresholds]
with open(cov_fpath) as f:
for line in f:
l = line.split() # genome; depth; number of bases; size of genome; fraction of bases with depth
depth, genome_fraction = int(l[1]), float(l[4])
if l[0] == 'genome':
avg_depth += depth * genome_fraction
for i, threshold in enumerate(qconfig.coverage_thresholds):
if depth >= threshold:
coverage_for_thresholds[i] += genome_fraction
with open(stats_fpath, 'a') as out_f:
out_f.write('%s depth\n' % int(avg_depth))
for i, threshold in enumerate(qconfig.coverage_thresholds):
out_f.write('%.2f coverage >= %sx\n' % (coverage_for_thresholds[i] * 100, threshold))
def analyse_coverage(output_dirpath, fpath, chr_names, bam_fpath, stats_fpath, err_fpath, logger):
filename = qutils.name_from_fpath(fpath)
bed_fpath = bam_to_bed(output_dirpath, filename, bam_fpath, err_fpath, logger)
chr_len_fpath = get_chr_len_fpath(fpath, chr_names)
cov_fpath = join(output_dirpath, filename + '.genomecov')
calculate_genome_cov(bed_fpath, cov_fpath, chr_len_fpath, err_fpath, logger, print_all_positions=False)
avg_depth = 0
coverage_for_thresholds = [0 for threshold in qconfig.coverage_thresholds]
with open(cov_fpath) as f:
for line in f:
l = line.split() # genome; depth; number of bases; size of genome; fraction of bases with depth
depth, genome_fraction = int(l[1]), float(l[4])
if l[0] == 'genome':
avg_depth += depth * genome_fraction
for i, threshold in enumerate(qconfig.coverage_thresholds):
if depth >= threshold:
coverage_for_thresholds[i] += genome_fraction
with open(stats_fpath, 'a') as out_f:
out_f.write('%s depth\n' % int(avg_depth))
for i, threshold in enumerate(qconfig.coverage_thresholds):
out_f.write('%.2f coverage >= %sx\n' % (coverage_for_thresholds[i] * 100, threshold))