def get_overlaping_clusters(self, region, overlap=1):
clusters = []
bam_tell, read_start = self.get_bam_tell(region)
print "TELL", bam_tell, read_start
if bam_tell or region.start < LINEAR_SIZE*4:
r = BamReader(self.bam_path, self.logger, bam_tell, read_start, self.chr_dict_inv)
for line in r:
c = Cluster(read=SAM, cached=False, read_half_open=self.read_half_open, rounding=self.rounding)
try:
c.read_line(line)
except InvalidLine:
print "Invalid line, .bam or .bai corrupt"
break
if c.overlap(region) >= overlap:
clusters.append(c)
elif c.start > region.end or c.name != region.name:
break
if len(clusters) > 0:
print "Num clusters", len(clusters)
print "first:", clusters[0].start, clusters[0].end
if len(clusters) > 1: print "end:", clusters[-1].start, clusters[-1].end
else:
print "No clusters found!"
print
return clusters
def test_overlap(self):
cluster1 = Cluster(read=BED)
cluster1.read_line('chr1 1 100 hola 666 +')
cluster2 = Cluster(read=BED)
cluster2.read_line('chr1 51 200 hola 666 +')
cluster3 = Cluster(read=PK)
cluster3.read_line('chr3 1 100 100:1')
cluster4 = Cluster(read=BED)
cluster4.read_line('chr5 1 1000 hola 666 +')
cluster5 = Cluster(read=PK)
cluster5.read_line('chr5 1 300 300:1')
cluster6 = Cluster(read=BED)
cluster6.read_line('chr5 100 900 hola 666 +')
cluster_discard = Cluster(read=BED, read_half_open=True)
cluster_discard.read_line('chrX 61836251 61836287 id:105282 1000 +')
cluster_satellite = Cluster(read=BED)
cluster_satellite.read_line('chrX 61836270 61837703 Satellite')
self.assertEqual(cluster_discard.overlap(cluster_satellite), 0.5)
self.assertEqual(cluster1.overlap(cluster3), 0)
self.assertEqual(cluster1.overlap(cluster1), 1)
self.assertEqual(cluster1.overlap(cluster2), 0.5)
def get_overlaping_clusters(self, region, overlap=1):
clusters = []
self.logger.debug('Launching Samtools for %s...'%region)
proc = subprocess.Popen("samtools view %s %s:%s-%s"%(self.bam_path, region.name, region.start, region.end), stdout=subprocess.PIPE, shell=True)
out, err = proc.communicate()
self.logger.debug('... done')
lines = filter(None, out.split("\n"))
self.logger.debug('Numlines in %s: %s'%(region, len(lines)))
for line in lines:
c = Cluster(read=SAM, cached=False, read_half_open=self.read_half_open, rounding=self.rounding)
try:
c.read_line(line)
except InvalidLine:
print "Invalid line, .bam or .bai corrupt"
break
if c.overlap(region) >= overlap:
clusters.append(c)
elif c.start > region.end or c.name != region.name:
break
return clusters
