def make_joined_read(mate, reads, tags=None):
if tags is None:
tags = []
# flip reverse strand reads
if not reads[0].is_unmapped and reads[0].is_reverse:
reads = sorted(reads, reverse=True)
# make new reads
a = pysam.AlignedRead()
# create paired-end reads but do not mark them
# as proper pairs and set all mate information
# to 'unmapped'
a.qname = reads[0].qname
a.seq = ''.join(r.seq for r in reads)
a.qual = ''.join(r.qual for r in reads)
a.is_paired = True
a.is_proper_pair = False
a.mate_is_unmapped = True
a.mrnm = -1
a.mpos = -1
if mate == 0:
a.is_read1 = True
a.is_read2 = False
else:
a.is_read1 = False
a.is_read2 = True
a.isize = 0
a.mapq = 255
a.is_unmapped = reads[0].is_unmapped
if a.is_unmapped:
a.rname = -1
a.pos = 0
# add the XM tag from bowtie saying whether unmapped
# due to multimapping or other reason
xm_tag = min(r.opt('XM') for r in reads)
tags.append(('XM', xm_tag))
else:
a.is_reverse = reads[0].is_reverse
a.rname = reads[0].rname
a.pos = reads[0].pos
a.cigar = ((0, len(a.seq)), )
# compute edit dist
edit_dist = 0
for r in reads:
edit_dist += r.opt('NM')
tags.append(('NM', edit_dist))
# compute mismatches to reference (MD)
tags.append(('MD', merge_MD_tags([r.opt('MD') for r in reads])))
a.tags = tags
return a
def translate_read(read, chrom, strand, intervals):
# skip unmapped reads
if read.is_unmapped:
return read
elif chrom == -1:
#logging.warning("discarded alignment %s that does not map to genomic references and cannot be translated" % (str(read)))
# throw away reads that cannot be translated by
# creating a dummy unmapped read
a = pysam.AlignedRead()
a.qname = read.qname
a.seq = read.seq
a.is_unmapped = True
a.rname = -1
a.pos = -1
a.mapq = 0
a.mrnm = -1
a.mpos = -1
a.isize = 0
a.qual = read.qual
a.tags = [("XM", 0)]
return a
elif (chrom >= 0) and (intervals is None):
# read maps directly to a genomic reference so simply
# alter the reference id to correctly refer to the new
# SAM header
read.rname = chrom
return read
genomic_intervals = translate_transcriptome_to_genomic_intervals(
read, chrom, strand, intervals)
spliced, cigar = get_cigar(genomic_intervals)
if spliced:
read.tags = read.tags + [("XS", "-" if strand else "+")]
# modify read
read.rname = chrom
read.pos = genomic_intervals[0][0]
read.cigar = cigar
# flip reads that aligned to negative strand genes
if strand == STRAND_REV:
rev_quals = read.qual[::-1]
read.is_reverse = not read.is_reverse
read.seq = DNA_reverse_complement(read.seq)
read.qual = rev_quals
new_tags = []
for name, val in read.tags:
if name == 'MD':
val = reverse_complement_MD_tag(val)
new_tags.append((name, val))
read.tags = new_tags
return read
def copy_read(r):
a = pysam.AlignedRead()
a.qname = r.qname
a.seq = r.seq
a.flag = r.flag
a.rname = r.rname
a.pos = r.pos
a.mapq = r.mapq
a.cigar = r.cigar
a.mrnm = r.mrnm
a.mpos = r.mpos
a.isize = r.isize
a.qual = r.qual
a.tags = r.tags
return a
def make_unmapped_copy(r):
a = pysam.AlignedRead()
a.qname = r.qname
a.seq = r.seq
a.qual = r.qual
a.is_unmapped = True
a.is_qcfail = False
a.is_paired = True
a.is_proper_pair = False
a.mate_is_unmapped = True
a.mrnm = -1
a.mpos = -1
a.is_read1 = r.is_read1
a.is_read2 = r.is_read2
a.isize = 0
a.mapq = 255
a.is_reverse = False
a.rname = -1
a.pos = 0
a.cigar = ()
a.tags = (('XM', 0), )
return a
def fastq_to_bam(fastq_files, qual_format, bam_file):
fqfhs = [parse_fastq(open(f)) for f in fastq_files]
qual_func = get_qual_conversion_func(qual_format)
header = {'HD': {'VN': '1.0', 'SO': 'unknown'}}
# 'SQ': [{'LN': 1, 'SN': 'dummy'}]}
bamfh = pysam.Samfile(bam_file, "wb", header=header)
try:
while True:
for i, fqiter in enumerate(fqfhs):
id, seq, qual = fqiter.next()
a = pysam.AlignedRead()
a.rname = -1
a.mrnm = -1
#a.pos = 0
#a.mpos = 0
a.qname = id
a.seq = seq
a.qual = qual_func(qual)
a.is_read1 = (i == 0)
a.is_read2 = (i == 1)
bamfh.write(a)
except StopIteration:
pass
bamfh.close()