class TestIntervalTrack(unittest.TestCase):
def setUp(self):
filename = mktemp(prefix="tmp", suffix=".h5")
self.filename = filename
self.refs = (('gene1', 1200), ('gene2', 200))
self.tf = TrackFactory(self.filename, 'w', refs=self.refs)
self.interval_dtype = get_base_dtype_fields() + [('id', 'u4')]
self.t = self.tf.create_track('intervals1', IntervalTrack,
dtype=self.interval_dtype)
def tearDown(self):
self.tf.close()
if os.path.exists(self.filename):
os.remove(self.filename)
def test_save_load_index(self):
"""create a table, add intervals, save the index, and
ensure the index persists when reloading the interval table
"""
mytrack = self.t
ref = 'gene1'
intervals = []
for i in xrange(100):
start = np.random.randint(0, 1150)
end = start + np.random.randint(1, 50)
interval = Interval(ref, start, end, i)
intervals.append(interval)
mytrack.add(interval)
# create and save index
mytrack.index(persist=True)
# save the index table
tblcopy = mytrack.indexes['gene1'].tree.tbl.copy()
tblroot = mytrack.indexes['gene1'].tree.root_id
# close and reopen the file
self.tf.close()
self.tf = TrackFactory(self.filename, 'r', refs=self.refs)
mytrack = self.tf.get_track('intervals1')
# compare indexes
self.assertTrue(np.all(tblcopy == mytrack.indexes['gene1'].tree.tbl))
self.assertEqual(tblroot, mytrack.indexes['gene1'].tree.root_id)
def testNonOverlapping(self):
mytrack = self.t
# insert some non-overlapping intervals
intervals = []
for i,start in enumerate(xrange(0, 900, 50)):
ref = 'gene1'
end = start + 10
interval = Interval(ref, start, end, i)
intervals.append(interval)
mytrack.add(interval)
# check intervals
for i,interval in enumerate(intervals):
hits = mytrack.intersect(interval.ref, interval.start, interval.end)
self.assertTrue([x['id'] for x in hits] == [i])
# create interval tree index
mytrack.index()
for i,interval in enumerate(intervals):
hits = mytrack.intersect(interval.ref, interval.start, interval.end)
self.assertTrue([x['id'] for x in hits] == [i])
def testOverlappingIntervals(self):
mytrack = self.t
# insert overlapping intervals
intervals = []
for i,start in enumerate(xrange(0, 900, 50)):
ref = 'gene1'
end = start + 60
interval = Interval(ref, start, end, i)
intervals.append(interval)
mytrack.add(interval)
# check intervals
self.assertTrue([x['id'] for x in mytrack.intersect(intervals[0].ref, intervals[0].start, intervals[0].end)] == [0,1])
for i in xrange(1, len(intervals) - 1):
interval = intervals[i]
self.assertTrue([x['id'] for x in mytrack.intersect(interval.ref, interval.start, interval.end)] == [i-1, i, i+1])
self.assertTrue([x['id'] for x in mytrack.intersect(intervals[-1].ref, intervals[-1].start, intervals[-1].end)] == [len(intervals) - 2, len(intervals)-1])
# create interval tree index
mytrack.index()
self.assertTrue([x['id'] for x in mytrack.intersect(intervals[0].ref, intervals[0].start, intervals[0].end)] == [0,1])
for i in xrange(1, len(intervals) - 1):
interval = intervals[i]
self.assertTrue([x['id'] for x in mytrack.intersect(interval.ref, interval.start, interval.end)] == [i-1, i, i+1])
self.assertTrue([x['id'] for x in mytrack.intersect(intervals[-1].ref, intervals[-1].start, intervals[-1].end)] == [len(intervals) - 2, len(intervals)-1])
def testBeforeAfter(self):
mytrack = self.t
# insert overlapping intervals
intervals = []
for i,start in enumerate(xrange(0, 1000, 50)):
ref = 'gene1'
end = start + 10
interval = Interval(ref, start, end, i)
intervals.append(interval)
mytrack.add(interval)
# check intervals
mytrack.index()
for i,interval in enumerate(intervals):
res = mytrack.before('gene1', interval.start, num_intervals=len(intervals), max_dist=1200)
if [x['id'] for x in res] != range(i-1, -1, -1):
print res, range(i-1, -1, -1)
self.assertTrue([x['id'] for x in res] == range(i-1, -1, -1))
res = mytrack.after('gene1', interval.end, num_intervals=len(intervals), max_dist=1200)
if [x['id'] for x in res] != range(i+1, len(intervals)):
print res, range(i-1, -1, -1)
self.assertTrue([x['id'] for x in res] == range(i+1, len(intervals)))
def _boundary_checks(self, mytrack):
# test left boundary
self.assertTrue(len(mytrack.before('gene1', 199, 1, 2000)) == 0)
self.assertTrue(len(mytrack.before('gene1', 200, 1, 2000)) == 1)
self.assertTrue(len(mytrack.before('gene1', 201, 1, 2000)) == 1)
self.assertTrue(len(mytrack.before('gene1', 202, 1, 2000)) == 1)
# test right boundary
self.assertTrue(len(mytrack.after('gene1', 101, 1, 2000)) == 0)
self.assertTrue(len(mytrack.after('gene1', 100, 1, 2000)) == 0)
self.assertTrue(len(mytrack.after('gene1', 99, 1, 2000)) == 1)
self.assertTrue(len(mytrack.after('gene1', 98, 1, 2000)) == 1)
# test left max dist
self.assertTrue(len(mytrack.before('gene1', 200, 1, 1)) == 1)
self.assertTrue(len(mytrack.before('gene1', 201, 1, 1)) == 0)
self.assertTrue(len(mytrack.before('gene1', 202, 1, 1)) == 0)
self.assertTrue(len(mytrack.before('gene1', 300, 1, 102)) == 1)
self.assertTrue(len(mytrack.before('gene1', 300, 1, 101)) == 1)
self.assertTrue(len(mytrack.before('gene1', 300, 1, 100)) == 0)
self.assertTrue(len(mytrack.before('gene1', 300, 1, 99)) == 0)
# test right max dist
self.assertTrue(len(mytrack.after('gene1', 99, 1, 1)) == 1)
self.assertTrue(len(mytrack.after('gene1', 98, 1, 1)) == 0)
self.assertTrue(len(mytrack.after('gene1', 98, 1, 2)) == 1)
self.assertTrue(len(mytrack.after('gene1', 0, 1, 100)) == 1)
self.assertTrue(len(mytrack.after('gene1', 0, 1, 99)) == 0)
def testBeforeAfterBoundaries(self):
mytrack = self.t
# insert overlapping intervals
mytrack.add(Interval('gene1', 100, 200, 0))
self._boundary_checks(mytrack)
mytrack.index()
self._boundary_checks(mytrack)
def _intersect_checks(self, mytrack):
self.assertTrue(len(mytrack.intersect('gene1', 0, 100)) == 0)
self.assertTrue(len(mytrack.intersect('gene1', 0, 101)) == 1)
self.assertTrue(len(mytrack.intersect('gene1', 200, 210)) == 0)
self.assertTrue(len(mytrack.intersect('gene1', 199, 210)) == 1)
def testIntersectBoundaries(self):
mytrack = self.t
# insert overlapping intervals
mytrack.add(Interval('gene1', 100, 200, 0))
self._intersect_checks(mytrack)
mytrack.index()
self._intersect_checks(mytrack)
def testMultipleReferences(self):
mytrack = self.t
# add intervals to different references
ref = 'gene1'
for i in xrange(0, 10):
mytrack.add(Interval(ref, i, i+10, i))
if ref == 'gene1':
ref = 'gene2'
else:
ref = 'gene1'
ids = [r['id'] for r in mytrack.intersect('gene1', 0, 20)]
self.assertTrue(set(ids) == set(range(0, 10, 2)))
ids = [r['id'] for r in mytrack.intersect('gene2', 0, 20)]
self.assertTrue(set(ids) == set(range(1, 10, 2)))
def testAdd(self):
mytrack = self.t
myrow = np.zeros(1, dtype=self.interval_dtype)[0]
intervals = []
for i,start in enumerate(xrange(0, 1000, 50)):
myrow['ref'] = 'gene1'
myrow['start'] = start
myrow['end'] = start + 10
myrow['id'] = i
ref = 'gene1'
end = start + 10
interval = Interval(ref, start, end, i)
mytrack.add(myrow)
intervals.append(interval)
for i,interval in enumerate(intervals):
res = mytrack[(interval.ref, interval.start, interval.end)]
self.assertTrue(len(res) == 1)
hit = res[0]
self.assertTrue((interval.ref, interval.start, interval.end, interval.id) ==
(hit['ref'], hit['start'], hit['end'], hit['id']))
class TestSequenceTrack(unittest.TestCase):
def setUp(self):
filename = mktemp(prefix="tmp", suffix=".h5")
self.filename = filename
self.refs = (('chr1', 100), ('chr2', 200))
self.tf = TrackFactory(self.filename, 'w', refs=self.refs)
def tearDown(self):
self.tf.close()
if os.path.exists(self.filename):
os.remove(self.filename)
def test_bpb(self):
for bpb in (2,3,4):
tname = 't%d' % (bpb)
t = self.tf.create_track(tname, SequenceTrack, bpb=bpb)
# write a cycle of bases one at a time and then
# check them
dna_iter = itertools.cycle('AATTTGCCTGC')
for i in xrange(100):
t[('chr1', i)] = dna_iter.next()
dna_iter = itertools.cycle('AATTTGCCTGC')
for i in xrange(100):
self.assertEqual(t[('chr1', i)], dna_iter.next())
# write chunks of bases and check them
dna_iter = itertools.repeat('CACATGTAGAGCT')
for i in xrange(0, 195, 13):
t[('chr2', i, i+13)] = dna_iter.next()
for i in xrange(0, 195, 13):
self.assertEqual(t[('chr2', i, i+13)], dna_iter.next())
# overwrite a range of bases without affecting others
t[('chr2', 99, 102)] = 'TTT'
self.assertEqual(t[('chr2', 94,107)], 'ATGTATTTCTCAC')
def test_bases(self):
t2 = self.tf.create_track("t2", SequenceTrack, bpb=2)
t3 = self.tf.create_track("t3", SequenceTrack, bpb=3)
t4 = self.tf.create_track("t4", SequenceTrack, bpb=4)
# test default bases
self.assertEqual(t2[('chr1', 35, 45)], 'AAAAAAAAAA')
self.assertEqual(t3[('chr1', 35, 45)], 'NNNNNNNNNN')
self.assertEqual(t4[('chr1', 35, 45)], 'NNNNNNNNNN')
# test storing lower case bases and 'N's
seq = 'NattgcgcNN'
t2[('chr1', 35, 45)] = seq
t3[('chr1', 35, 45)] = seq
t4[('chr1', 35, 45)] = seq
self.assertEqual(t2[('chr1', 35, 45)], seq.upper().replace('N', 'A'))
self.assertEqual(t3[('chr1', 35, 45)], seq.upper())
self.assertEqual(t4[('chr1', 35, 45)], seq)
def test_fromfasta(self):
# read/write to chromosome
seqlist = [">chr2",
"ATGCAGTGAC",
"GTGACGAGAG",
"TGTAGAGAGA",
"GTGATGTATG",
"GGGGGGGGGG",
"GGGCCCCCGC",
"CCCCCCCCCC"
"CCCCCCCCCC",
"ATGAAAAAAG",
"TTGCCAAAAA",
"AAAAAAAAAA",
"AAAAAAAAAA",
"AAAGTGATTT",
"TTTTTTGCCG",
"AGCGCGCGCG"]
fullseq = ''.join(seqlist[1:])
t2 = self.tf.create_track("t2", SequenceTrack, bpb=2)
t2.fromfasta(iter(seqlist))
self.assertEqual(fullseq, t2[('chr2', 0, len(fullseq))])
# insert at different places
for startpos in (0, 10, 20, 30, 40):
seqlist[0] = ">chr2:%d" % (startpos)
t2.fromfasta(iter(seqlist))
self.assertEqual(fullseq, t2[('chr2', startpos, startpos + len(fullseq))])