def test_getSimilar(self):
"""Alignment getSimilar should get all sequences close to target seq"""
aln = self.many
x = Rna('gggggggggg')
y = Rna('----------')
#test min and max similarity ranges
result = aln.getSimilar(aln['a'],min_similarity=0.4,max_similarity=0.7)
for seq in 'cefg':
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 4)
result = aln.getSimilar(aln['a'],min_similarity=0.95,max_similarity=1)
for seq in 'a':
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 1)
result = aln.getSimilar(aln['a'], min_similarity=0.75, \
max_similarity=0.85)
for seq in 'bd':
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 2)
result = aln.getSimilar(aln['a'],min_similarity=0,max_similarity=0.2)
self.assertEqual(len(result), 0)
#test some sequence transformations
transform = lambda s: s[1:4]
result = aln.getSimilar(aln['a'], min_similarity=0.5, \
transform=transform)
for seq in 'abdfg':
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 5)
transform = lambda s: s[-3:]
result = aln.getSimilar(aln['a'], min_similarity=0.5, \
transform=transform)
for seq in 'abcde':
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 5)
#test a different distance metric
metric = lambda x, y: x.count('g') + y.count('g')
result = aln.getSimilar(aln['a'], min_similarity=5, max_similarity=10, \
metric=metric)
for seq in 'ef':
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 2)
#test the combination of a transform and a distance metric
aln = Alignment(dict(enumerate(map(Rna, ['aA-ac','A-aAC','aa-aa']))))
transform = lambda s: Rna(s.upper())
metric = RnaSequence.fracSameNonGaps
#first, do it without the transformation
result = aln.getSimilar(aln[0], min_similarity=0.5, metric=metric)
for seq in [0,2]:
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 2)
result = aln.getSimilar(aln[0], min_similarity=0.8, metric=metric)
for seq in [0]:
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 1)
#then, verify that the transform changes the results
result = aln.getSimilar(aln[0], min_similarity=0.5, metric=metric, \
transform=transform)
for seq in [0,1,2]:
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 3)
result = aln.getSimilar(aln[0], min_similarity=0.8, metric=metric, \
transform=transform)
for seq in [0,1]:
assert seq in result
assert result[seq] == aln[seq]
self.assertEqual(len(result), 2)
