def test_subtract_with_gaps(self):
"""
238c238
< chr17 147132 147296 5:1.00|22:2.00|37:1.00|37:2.00|64:1.00 2.0 . 147180 224.0
---
> chr17 147132 147260 5:1.00|22:2.00|37:1.00|37:2.00|28:1.00 2.0 . 147180 188.0
experiment
chr17 146970 147070
chr17 147058 147158
chr17 147132 147232
chr17 147196 147296
chr17 147303 147403
chr17 147445 147545
chr17 147461 147561
control
chr17 146904 147004
chr17 147036 147136
chr17 147261 147361
chr17 147437 147537
chr17 147472 147572
chr17 147472 147572
"""
exp_cached = Cluster(cached=True)
control0 = Cluster(read=BED, cached=True)
control1 = Cluster(read=BED, cached=True)
control2 = Cluster(read=BED, cached=True)
control3 = Cluster(read=BED, cached=True)
control4 = Cluster(read=BED, cached=True)
#experiment.bed
exp_cached.read_line("chr17 146970 147296 88:1.00|13:2.00|61:1.00|27:2.00|37:1.00|37:2.00|64:1.00 2.0 . 147139 404.0")
#print exp_cached.write_line()
#control.bed
r = Region("chr17", 1, 2000)
control0.read_line("chr17 146904 147004")
control1.read_line("chr17 147036 147136")
control2.read_line("chr17 147261 147361")
control3.read_line("chr17 147437 147537")
control4.read_line("chr17 147472 147572")
r.add_tags([control1, control2])
meta = r.get_metacluster()
#print meta._profile
exp_cached -= meta
def test_get_profile(self):
r = Region(start=1, end=1999)
c = Cluster(read=BED)
c.read_line('chr4 1 40')
r.add_tags(c, True)
c = Cluster(read=BED, read_half_open=True)
c.read_line('chr4 400 500')
r.add_tags(c, True)
meta = r.get_metacluster()
self.assertEqual(meta._levels, [[40, 1.0], [360, 0.0], [100, 1.0]])
def test_FDR(self):
r = Region('', 1, 1999)
tags = []
for i in range(0, 50):
c = Cluster()
c.read_line('chr4 %s %s 20:1'%(i, i+50))
tags.append(c)
r.add_tags(tags, True)
c = Cluster()
c.read_line('chr4 55555 55558 7:1')
r.add_tags(c)
result = r.get_FDR_clusters()
self.assertEqual(len(result), 1)
def test_region_swap_rpkm(self):
total_reads = 400000000
total_reads_b = 500000000
r = Region("chr1", 1, 1000, name2="bla")
c = Cluster(read=BED)
c.read_line('chr1 1 40')
r.add_tags(c)
c = Cluster(read=BED)
c.read_line('chr1 2 40')
r.add_tags(c)
c = Cluster(read=BED)
c.read_line('chr1 3 40')
r.add_tags(c)
r2 = Region("chr1", 1, 1000, name2="bla")
c = Cluster(read=BED)
c.read_line('chr1 100 140')
r2.add_tags(c)
c = Cluster(read=BED)
c.read_line('chr1 101 140')
r2.add_tags(c)
c = Cluster(read=BED)
c.read_line('chr1 102 140')
r2.add_tags(c)
swap1, swap2 = r.swap(r2)
len(r.tags), len(r2.tags), len(swap1.tags), len(swap2.tags), r.rpkm(total_reads), r2.rpkm(total_reads_b), swap1.rpkm((total_reads+total_reads_b)/2), swap2.rpkm((total_reads+total_reads_b)/2) #TODO dont know how to test this
def test_normalized_counts(self):
total_number_reads = 1e7
region = Region("chr1", 1, 300)
region_bed12 = Region("chr1", 1, 300, exome_size = 200)
c = Cluster(read=BED)
for i in range(0, 5):
c.read_line("chr1 1 100")
region.add_tags(c, True)
region_bed12.add_tags(c, True)
c.clear()
self.assertEqual(region.normalized_counts(), 5.) #simple-counts
self.assertEqual(region.normalized_counts(region_norm=True, total_n_norm=True, total_reads = total_number_reads), 1.666666666666667) #rpkm
self.assertEqual(region_bed12.normalized_counts(region_norm=True, total_n_norm=True, total_reads = total_number_reads), 2.5) #rpkm with exon_size
self.assertEqual(region.normalized_counts(pseudocount=True), 6.) #with pseudocounts
self.assertEqual(region.normalized_counts(region_norm=True, total_n_norm=True, total_reads = total_number_reads, regions_analyzed=10000, pseudocount=True), 1.998001998001998)