def align(self, inBam, refFasta, outBam, min_qual=0, nodisk=True, JVMmemory=None, **kwargs):
with tools.samtools.SamtoolsTool().bam2fq_tmp(inBam) as (in1, in2), \
util.file.tmp_dir('_bbmap_align') as t_dir:
tmp_bam = os.path.join(t_dir, 'bbmap_out.bam')
self.execute(tool='bbmap.sh', in1=in1, in2=in2, ref=refFasta, out=tmp_bam, nodisk=nodisk, **kwargs)
# Samtools filter (optional)
if min_qual:
tmp_bam2 = os.path.join(tdir, 'bbmap.filtered.bam')
cmd = [samtools.install_and_get_path(), 'view', '-b', '-S', '-1', '-q', str(min_qual), tmp_bam]
_log.debug('%s > %s', ' '.join(cmd), tmp_bam2)
with open(tmp_bam2, 'wb') as outf:
util.misc.run_and_save(cmd, outf=outf)
os.unlink(tmp_bam)
tmp_bam = tmp_bam2
# Picard SortSam
sorter = tools.picard.SortSamTool()
sorter.execute(
tmp_bam,
outBam,
sort_order='coordinate',
picardOptions=['CREATE_INDEX=true', 'VALIDATION_STRINGENCY=SILENT'],
JVMmemory=JVMmemory
)
def align_one_rg_bam(self, inBam, refFasta, outBam, rgid=None, rgs=None, options=None, min_qual=0, JVMmemory=None):
''' Execute Novoalign on BAM inputs and outputs.
Requires that only one RG exists (will error otherwise).
Use Picard to sort and index the output BAM.
If min_qual>0, use Samtools to filter on mapping quality.
'''
options = options or ["-r", "Random"]
samtools = tools.samtools.SamtoolsTool()
# Require exactly one RG
rgs = rgs if rgs is not None else samtools.getReadGroups(inBam)
if len(rgs) == 0:
raise InvalidBamHeaderError("{} lacks read groups".format(inBam))
elif len(rgs) == 1:
if not rgid:
rgid = list(rgs.keys())[0]
elif not rgid:
raise InvalidBamHeaderError("{} has {} read groups, but we require exactly one".format(inBam, len(rgs)))
if rgid not in rgs:
raise InvalidBamHeaderError("{} has read groups, but not {}".format(inBam, rgid))
#rg = rgs[rgid]
# Strip inBam to just one RG (if necessary)
if len(rgs) == 1:
one_rg_inBam = inBam
else:
# strip inBam to one read group
tmp_bam = util.file.mkstempfname('.onebam.bam')
samtools.view(['-b', '-r', rgid], inBam, tmp_bam)
# special exit if this file is empty
if samtools.count(tmp_bam) == 0:
return
# simplify BAM header otherwise Novoalign gets confused
one_rg_inBam = util.file.mkstempfname('.{}.in.bam'.format(rgid))
headerFile = util.file.mkstempfname('.{}.header.txt'.format(rgid))
with open(headerFile, 'wt') as outf:
for row in samtools.getHeader(inBam):
if len(row) > 0 and row[0] == '@RG':
if rgid != list(x[3:] for x in row if x.startswith('ID:'))[0]:
# skip all read groups that are not rgid
continue
outf.write('\t'.join(row) + '\n')
samtools.reheader(tmp_bam, headerFile, one_rg_inBam)
os.unlink(tmp_bam)
os.unlink(headerFile)
# Novoalign
tmp_sam = util.file.mkstempfname('.novoalign.sam')
tmp_sam_err = util.file.mkstempfname('.novoalign.sam.err')
cmd = [self.install_and_get_path(), '-f', one_rg_inBam] + list(map(str, options))
cmd = cmd + ['-F', 'BAM', '-d', self._fasta_to_idx_name(refFasta), '-o', 'SAM']
_log.debug(' '.join(cmd))
with open(tmp_sam, 'wt') as outf:
util.misc.run_and_save(cmd, outf=outf)
# Samtools filter (optional)
if min_qual:
tmp_bam2 = util.file.mkstempfname('.filtered.bam')
cmd = [samtools.install_and_get_path(), 'view', '-b', '-S', '-1', '-q', str(min_qual), tmp_sam]
_log.debug('%s > %s', ' '.join(cmd), tmp_bam2)
with open(tmp_bam2, 'wb') as outf:
util.misc.run_and_save(cmd, outf=outf)
os.unlink(tmp_sam)
tmp_sam = tmp_bam2
# Picard SortSam
sorter = tools.picard.SortSamTool()
sorter.execute(
tmp_sam,
outBam,
sort_order='coordinate',
picardOptions=['CREATE_INDEX=true', 'VALIDATION_STRINGENCY=SILENT'],
JVMmemory=JVMmemory
)