def get_median_read_coverage(output_path, num_threads, overwrite_files):
""" Given the read alignments, use samtools stats to return an approximate median coverage value. """
log("Calculating global read coverage")
if os.path.isfile(output_path + "c_reads_against_query.s.bam.stats"):
if not overwrite_files:
log("retaining pre-existing file: " + output_path + "c_reads_against_query.s.bam.stats")
else:
log("overwriting pre-existing file: " + output_path + "c_reads_against_query.s.bam.stats")
st = pysam.stats("-@", str(num_threads), output_path + "c_reads_against_query.s.bam")
with open(output_path + "c_reads_against_query.s.bam.stats", "w") as f:
f.write(st)
else:
st = pysam.stats("-@", str(num_threads), output_path + "c_reads_against_query.s.bam")
with open(output_path + "c_reads_against_query.s.bam.stats", "w") as f:
f.write(st)
# Get the coverage histogram (for 1 to 1k)
covs = []
with open(output_path + "c_reads_against_query.s.bam.stats") as f:
for line in f:
if line.startswith("COV"):
covs.append(int(line.split("\t")[3]))
# Get the median from the histogram
covs = np.asarray(covs, dtype=np.int32)
# Remove the last value, which is a catch-all for coverages > 1k
covs = covs[:-1]
mid = sum(covs) // 2
cs = 0
for i in range(len(covs)):
cs += covs[i]
if cs >= mid:
return i
raise ValueError("Unable to calculate read coverage. Check SAM/BAM files and stats file.")
def insert_size(this, bamFile, threads=1):
result_str = pysam.stats("-@", str(threads), bamFile)
insert_size = float(
re.search('SN\s+insert size average:\s+(\S+)',
result_str).group(1))
insert_std = float(
re.search('SN\s+insert size standard deviation:\s+(\S+)',
result_str).group(1))
return (insert_size, insert_std)
def samtools_stats(x):
"""Extract insert size of PE BAM file
samtools stats {in.bam} | grep 'insert size average'
output of command: samtools stats
Summary Numbers
grep ^SN | cut -f 2- | grep 'insert size average'
"""
stat = pysam.stats('-@', '8', x)
d = {}
for line in stat.split('\n'):
# check
if not line.startswith('SN'):
continue
# stat
sn, group, num = line.strip().split('\t')[0:3]
group = group.strip(':')
d[group] = num
return d