def testConvertRegion(self):
'''
Convert region/junction coordinates
'''
ngsutils.bam.convertregion.bam_convertregion(infname,
outfname,
chromsizes,
quiet=True)
bam = ngsutils.bam.bam_open(outfname)
for read in ngsutils.bam.bam_iter(bam):
if read.qname == 'A':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos,
109) # 0-based in BAM, 1-based in SAM
self.assertEqual(read.cigar, ngsutils.bam.cigar_fromstr('50M'))
elif read.qname == 'B':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 174)
self.assertEqual(read.cigar,
ngsutils.bam.cigar_fromstr('26M100N24M'))
elif read.qname == 'C':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 197)
self.assertEqual(read.cigar,
ngsutils.bam.cigar_fromstr('3M100N47M'))
elif read.qname == 'Z':
self.assertTrue(read.is_unmapped)
def testConvertRegionOverlap(self):
'''
Convert region/junction coordinates, require overlap with junction
'''
ngsutils.bam.convertregion.bam_convertregion(infname,
outfname,
chromsizes,
enforce_overlap=True,
quiet=True)
foundA = False
foundB = False
foundC = False
foundZ = False
bam = ngsutils.bam.bam_open(outfname)
for read in ngsutils.bam.bam_iter(bam):
if read.qname == 'A':
foundA = True
elif read.qname == 'B':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 174)
self.assertEqual(read.cigar,
ngsutils.bam.cigar_fromstr('26M100N24M'))
foundB = True
elif read.qname == 'C':
foundC = True
elif read.qname == 'Z':
foundZ = True
self.assertFalse(foundA)
self.assertTrue(foundB)
self.assertFalse(foundC)
self.assertFalse(foundZ)
def testConvertRegionOverlap(self):
'''
Convert region/junction coordinates, require overlap with junction
'''
ngsutils.bam.convertregion.bam_convertregion(infname, outfname, chromsizes, enforce_overlap=True, quiet=True)
foundA = False
foundB = False
foundC = False
foundZ = False
bam = ngsutils.bam.bam_open(outfname)
for read in ngsutils.bam.bam_iter(bam):
if read.qname == 'A':
foundA = True
elif read.qname == 'B':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 174)
self.assertEqual(read.cigar, ngsutils.bam.cigar_fromstr('26M100N24M'))
foundB = True
elif read.qname == 'C':
foundC = True
elif read.qname == 'Z':
foundZ = True
self.assertFalse(foundA)
self.assertTrue(foundB)
self.assertFalse(foundC)
self.assertFalse(foundZ)
def testMerge(self):
'''
Merge test.bam and test2.bam to tmp.bam
'''
fname1 = os.path.join(os.path.dirname(__file__), 'test.bam')
fname2 = os.path.join(os.path.dirname(__file__), 'test2.bam')
outfname = os.path.join(os.path.dirname(__file__), 'tmp.bam')
ngsutils.bam.merge.bam_merge(outfname, [fname1, fname2], quiet=True)
bam = ngsutils.bam.bam_open(outfname)
for read in ngsutils.bam.bam_iter(bam):
if read.qname == 'A': # AS chr2
self.assertEqual(bam.getrname(read.tid), 'chr2')
elif read.qname == 'B': # AS tie
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'C': # AS tie
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'D': # AS tie
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'E': # AS chr2
self.assertEqual(bam.getrname(read.tid), 'chr2')
elif read.qname == 'F': # AS chr1
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'Z': # still unmapped
self.assertTrue(read.is_unmapped)
def testMerge(self):
'''
Merge test.bam and test2.bam to tmp.bam
'''
fname1 = os.path.join(os.path.dirname(__file__), 'test.bam')
fname2 = os.path.join(os.path.dirname(__file__), 'test2.bam')
outfname = os.path.join(os.path.dirname(__file__), 'tmp.bam')
ngsutils.bam.merge.bam_merge(outfname, [fname1, fname2], quiet=True)
bam = ngsutils.bam.bam_open(outfname)
for read in ngsutils.bam.bam_iter(bam):
if read.qname == 'A': # AS chr2
self.assertEqual(bam.getrname(read.tid), 'chr2')
elif read.qname == 'B': # AS tie
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'C': # AS tie
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'D': # AS tie
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'E': # AS chr2
self.assertEqual(bam.getrname(read.tid), 'chr2')
elif read.qname == 'F': # AS chr1
self.assertEqual(bam.getrname(read.tid), 'chr1')
elif read.qname == 'Z': # still unmapped
self.assertTrue(read.is_unmapped)
def testConvertRegion(self):
'''
Convert region/junction coordinates
'''
ngsutils.bam.convertregion.bam_convertregion(infname, outfname, chromsizes, quiet=True)
bam = ngsutils.bam.bam_open(outfname)
for read in ngsutils.bam.bam_iter(bam):
if read.qname == 'A':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 109) # 0-based in BAM, 1-based in SAM
self.assertEqual(read.cigar, ngsutils.bam.cigar_fromstr('50M'))
elif read.qname == 'B':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 174)
self.assertEqual(read.cigar, ngsutils.bam.cigar_fromstr('26M100N24M'))
elif read.qname == 'C':
self.assertEqual(bam.getrname(read.tid), 'chr1')
self.assertEqual(read.pos, 197)
self.assertEqual(read.cigar, ngsutils.bam.cigar_fromstr('3M100N47M'))
elif read.qname == 'Z':
self.assertTrue(read.is_unmapped)
def find_nearest(bam, bed, maxdist=100000, out=sys.stdout):
for read in ngsutils.bam.bam_iter(bam):
if read.is_unmapped:
continue
dists = [] # will be an list tuples: (abs_val of the distance, 'up/down') (respective to + strand)
chrom = bam.getrname(read.tid)
strand = '-' if read.is_reverse else '+'
start = max(0, read.pos - maxdist)
end = min(read.aend + maxdist, bam.lengths[read.tid])
for region in bed.fetch(chrom, start, end, strand):
if region.start <= read.pos <= region.end:
# start is w/in region
dists.append((0, ''))
continue
elif region.start <= read.aend <= region.end:
# end is w/in region
dists.append((0, ''))
continue
elif read.pos <= region.start <= region.end <= read.aend:
# read spans the entire region
dists.append((0,''))
continue
elif region.end < read.pos:
dists.append((read.pos - region.end, 'up'))
else:
dists.append((region.start - read.aend, 'down'))
if dists:
dists.sort()
distance = dists[0][0]
orient = dists[0][1]
if distance > 0:
if not read.is_reverse and orient == 'down':
distance = -distance
elif read.is_reverse and orient == 'up':
distance = -distance
out.write('%s\t%s\n' % (read.qname, distance))
else:
out.write('%s\t*\n' % (read.qname,))
def find_nearest(bam, bed, maxdist=100000, out=sys.stdout):
for read in ngsutils.bam.bam_iter(bam):
if read.is_unmapped:
continue
dists = [
] # will be an list tuples: (abs_val of the distance, 'up/down') (respective to + strand)
chrom = bam.getrname(read.tid)
strand = '-' if read.is_reverse else '+'
start = max(0, read.pos - maxdist)
end = min(read.aend + maxdist, bam.lengths[read.tid])
for region in bed.fetch(chrom, start, end, strand):
if region.start <= read.pos <= region.end:
# start is w/in region
dists.append((0, ''))
continue
elif region.start <= read.aend <= region.end:
# end is w/in region
dists.append((0, ''))
continue
elif read.pos <= region.start <= region.end <= read.aend:
# read spans the entire region
dists.append((0, ''))
continue
elif region.end < read.pos:
dists.append((read.pos - region.end, 'up'))
else:
dists.append((region.start - read.aend, 'down'))
if dists:
dists.sort()
distance = dists[0][0]
orient = dists[0][1]
if distance > 0:
if not read.is_reverse and orient == 'down':
distance = -distance
elif read.is_reverse and orient == 'up':
distance = -distance
out.write('%s\t%s\n' % (read.qname, distance))
else:
out.write('%s\t*\n' % (read.qname, ))