def __init__(self, hdf_group, pe=False, strand=False, allele=False):
super(RnaseqTrack, self).__init__(hdf_group)
h5file = self._get_hdf_file()
if JUNCTION_GROUP not in self.hdf_group:
junc_group = h5file.createGroup(self.hdf_group, JUNCTION_GROUP)
else:
junc_group = self.hdf_group._f_getChild(JUNCTION_GROUP)
self.junc_track = IntervalTrack(junc_group, junction_dtype,
junction_expectedrows)
if COVERAGE_GROUP not in self.hdf_group:
cov_group = h5file.createGroup(self.hdf_group, COVERAGE_GROUP)
else:
cov_group = self.hdf_group._f_getChild(COVERAGE_GROUP)
self.cov_track = VectorTrack(cov_group, pe=pe, strand=strand,
allele=allele)
class RnaseqTrack(Track):
'''contains both genomic coverage data (array track) and splice
junction data (interval track)
'''
def __init__(self, hdf_group, pe=False, strand=False, allele=False):
super(RnaseqTrack, self).__init__(hdf_group)
h5file = self._get_hdf_file()
if JUNCTION_GROUP not in self.hdf_group:
junc_group = h5file.createGroup(self.hdf_group, JUNCTION_GROUP)
else:
junc_group = self.hdf_group._f_getChild(JUNCTION_GROUP)
self.junc_track = IntervalTrack(junc_group, junction_dtype,
junction_expectedrows)
if COVERAGE_GROUP not in self.hdf_group:
cov_group = h5file.createGroup(self.hdf_group, COVERAGE_GROUP)
else:
cov_group = self.hdf_group._f_getChild(COVERAGE_GROUP)
self.cov_track = VectorTrack(cov_group, pe=pe, strand=strand,
allele=allele)
def get_junction_track(self):
junc_group = self.hdf_group._f_getChild(JUNCTION_GROUP)
return IntervalTrack(junc_group)
def get_coverage_track(self):
cov_group = self.hdf_group._f_getChild(COVERAGE_GROUP)
return VectorTrack(cov_group)
def fromtophat(self, accepted_hits_bam, junctions_bed,
max_multimaps=None,
flip_read2_strand=True):
# insert splice junction track
track_name = self.hdf_group._v_name
rnames = set(self.get_rnames())
logging.info("[RnaseqTrack] adding junctions")
junc_iter = tophat_bed_to_juncs(track_name, open(junctions_bed))
for junc in junc_iter:
if junc[REF_COL_NAME] not in rnames:
logging.debug('Skipping junc %s' % str(junc))
else:
junc['id'] = self.junc_track.num_intervals
self.junc_track.add(junc)
self.junc_track.index(persist=True)
# insert coverage track
logging.info("creating coverage track")
bamfh = pysam.Samfile(accepted_hits_bam, "rb")
#cmdline = bamfh.header["PG"][0]["CL"]
#re.search(r'--max-multihits(?:\s+|=)(\d+)', cmdline)
intervalcoviter = BamCoverageIterator(bamfh,
norm_rlen=True,
num_hits_tag="NH",
hit_prob_tag=None,
max_multimaps=None,
keep_dup=True,
keep_qcfail=False,
flip_read2_strand=True)
self.cov_track.fromintervals(intervalcoviter)
# store coverage statistics to allow normalization calculations
stats = intervalcoviter.stats
logging.debug("\tProcessed '%d' valid reads" % (stats.num_reads))
logging.debug("\tTotal coverage '%f'" % (stats.total_cov))
bamfh.close()