def test_toProbs(self):
"""Counts toProbs should return valid prob matrix."""
c = Counts([1,2,3,4,2,2,2,2,0.2,0.4,0.6,0.8,1,0,0,0], RnaPairs)
p = c.toProbs()
assert isinstance(p, Probs)
self.assertEqual(p, Probs([0.1,0.2,0.3,0.4,0.25,0.25,0.25,0.25, \
0.1,0.2,0.3,0.4,1.0,0.0,0.0,0.0], RnaPairs))
self.assertEqual(p['U','U'], 0.1)
self.assertEqual(p['G','U'], 1.0)
self.assertEqual(p['G','G'], 0.0)
def test_toProbs(self):
"""Counts toProbs should return valid prob matrix."""
c = Counts([1, 2, 3, 4, 2, 2, 2, 2, 0.2, 0.4, 0.6, 0.8, 1, 0, 0, 0],
RnaPairs)
p = c.toProbs()
assert isinstance(p, Probs)
self.assertEqual(p, Probs([0.1,0.2,0.3,0.4,0.25,0.25,0.25,0.25, \
0.1,0.2,0.3,0.4,1.0,0.0,0.0,0.0], RnaPairs))
self.assertEqual(p['U', 'U'], 0.1)
self.assertEqual(p['G', 'U'], 1.0)
self.assertEqual(p['G', 'G'], 0.0)
def test_toCounts(self):
"""Probs toCounts should return counts object w/ right numbers"""
a = Alphabet('abc')**2
m = Probs([0.5, 0.25, 0.25, 0.1, 0.8, 0.1, 0.3, 0.6, 0.1], a)
obs = m.toCounts(30)
assert isinstance(obs, Counts)
exp = Counts([[5., 2.5, 2.5, 1, 8, 1, 3, 6, 1]], a)
self.assertEqual(obs, exp)
def test_fromPair(self):
"""Counts fromPair should return correct counts."""
s = Counts.fromPair( RnaSequence('UCCGAUCGAUUAUCGGGUACGUA'), \
RnaSequence('GUCGAGUAUAGCGUACGGCUACG'),
RnaPairs)
assert isinstance(s, Counts)
vals = [
('U','U',0),('U','C',2.5),('U','A',1),('U','G',2.5),
('C','U',2.5),('C','C',1),('C','A',1),('C','G',0.5),
('A','U',1),('A','C',1),('A','A',1),('A','G',2),
('G','U',2.5),('G','C',0.5),('G','A',2),('G','G',2),
]
for i, j, val in vals:
self.assertFloatEqual(s[i,j], val)
#check that it works for big seqs
s = Counts.fromPair( RnaSequence('UCAG'*1000), \
RnaSequence('UGAG'*1000),
RnaPairs)
assert isinstance(s, Counts)
vals = [
('U','U',1000),('U','C',0),('U','A',0),('U','G',0),
('C','U',0),('C','C',0),('C','A',0),('C','G',500),
('A','U',0),('A','C',0),('A','A',1000),('A','G',0),
('G','U',0),('G','C',500),('G','A',0),('G','G',1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i,j], val)
#check that it works for codon seqs
s1 = ModelRnaCodonSequence('UUCGCG')
s2 = ModelRnaCodonSequence('UUUGGG')
c = Counts.fromPair(s1, s2, RNA.Alphabet.Triples**2)
self.assertEqual(c._data.sum(), 2)
self.assertEqual(c._data[0,1], 0.5)
self.assertEqual(c._data[1,0], 0.5)
self.assertEqual(c._data[55,63], 0.5)
self.assertEqual(c._data[63,55], 0.5)
def test_counts_to_probs(self):
"""counts_to_probs should skip cases with zero rows"""
counts = {
(0,1): Counts(array([[0,1],[1,0]]), ABPairs),
(1,2): Counts(array([[0,0],[1,0]]), ABPairs), #bad row
(0,3): Counts(array([[0,0],[0,0]]), ABPairs), #bad row
(0,4): Counts(array([[0.0,0.0],[0.0,0.0]]), ABPairs), #bad row
(0,5): Counts(array([[0.1,0.3],[0.0,0.0]]), ABPairs), #bad row
(3,4): Counts(array([[0.1,0.3],[0.4,0.1]]), ABPairs),
(2,1): Counts(array([[0,5],[1,0]]), ABPairs),
}
result = counts_to_probs(counts)
self.assertEqual(len(result), 3)
self.assertFloatEqual(result[(0,1)]._data, array([[0,1],[1,0]]))
self.assertFloatEqual(result[(3,4)]._data, \
array([[0.25,0.75],[0.8,0.2]]))
self.assertFloatEqual(result[(2,1)]._data, array([[0,1],[1,0]]))
def test_fromTriple(self):
"""Counts fromTriple should return correct counts."""
cft = Counts.fromTriple
rs = RnaSequence
A, C, G, U = map(rs, 'ACGU')
#counts if different from both the other groups
s = cft(A, C, C, RnaPairs)
assert isinstance(s, Counts)
self.assertEqual(s['C','A'], 1)
self.assertEqual(s['A','C'], 0)
self.assertEqual(s['C','C'], 0)
#try it with longer sequences
AAA, CCC = map(rs, ['AAA', 'CCC'])
s = cft(AAA, CCC, CCC, RnaPairs)
self.assertEqual(s['C','A'], 3)
self.assertEqual(s['A','C'], 0)
#doesn't count if all three differ
ACG, CGA, GAC = map(rs, ['ACG','CGA','GAC'])
s = cft(ACG, CGA, GAC, RnaPairs)
self.assertEqual(s['C','A'], 0)
self.assertEqual(s['A','C'], 0)
self.assertEqual(s, Counts.empty(RnaPairs))
#counts as no change if same as other sequence...
s = cft(AAA, AAA, CCC, RnaPairs)
self.assertEqual(s['A','A'], 3)
self.assertEqual(s['A','C'], 0)
#...or same as the outgroup
s = cft(AAA, CCC, AAA, RnaPairs)
self.assertEqual(s['A','A'], 3)
self.assertEqual(s['A','C'], 0)
#spot-check a mixed example
s = cft( \
rs('AUCGCUAGCAUACGUCA'),
rs('AAGCUGCGUAGCGCAUA'),
rs('GCGCAUAUGACGAUAGC'),
RnaPairs
)
vals = [
('U','U',1),('U','C',0),('U','A',0),('U','G',0),
('C','U',0),('C','C',0),('C','A',0),('C','G',1),
('A','U',1),('A','C',0),('A','A',4),('A','G',0),
('G','U',0),('G','C',1),('G','A',0),('G','G',1),
]
for i, j, val in vals:
self.assertFloatEqual(s[i,j], val)
#check a long sequence
s = cft( \
rs('AUCGCUAGCAUACGUCA'*1000),
rs('AAGCUGCGUAGCGCAUA'*1000),
rs('GCGCAUAUGACGAUAGC'*1000),
RnaPairs
)
vals = [
('U','U',1000),('U','C',0),('U','A',0),('U','G',0),
('C','U',0),('C','C',0),('C','A',0),('C','G',1000),
('A','U',1000),('A','C',0),('A','A',4000),('A','G',0),
('G','U',0),('G','C',1000),('G','A',0),('G','G',1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i,j], val)
#check that it works when forced to use both variants of fromTriple
s = cft( \
rs('AUCGCUAGCAUACGUCA'*1000),
rs('AAGCUGCGUAGCGCAUA'*1000),
rs('GCGCAUAUGACGAUAGC'*1000),
RnaPairs,
threshold=0 #forces "large" method
)
vals = [
('U','U',1000),('U','C',0),('U','A',0),('U','G',0),
('C','U',0),('C','C',0),('C','A',0),('C','G',1000),
('A','U',1000),('A','C',0),('A','A',4000),('A','G',0),
('G','U',0),('G','C',1000),('G','A',0),('G','G',1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i,j], val)
s = cft( \
rs('AUCGCUAGCAUACGUCA'*1000),
rs('AAGCUGCGUAGCGCAUA'*1000),
rs('GCGCAUAUGACGAUAGC'*1000),
RnaPairs,
threshold=1e12 #forces "small" method
)
vals = [
('U','U',1000),('U','C',0),('U','A',0),('U','G',0),
('C','U',0),('C','C',0),('C','A',0),('C','G',1000),
('A','U',1000),('A','C',0),('A','A',4000),('A','G',0),
('G','U',0),('G','C',1000),('G','A',0),('G','G',1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i,j], val)
#check that it works for codon seqs
s1 = ModelRnaCodonSequence('UUCGCG')
s2 = ModelRnaCodonSequence('UUUGGG')
s3 = s2
c = Counts.fromTriple(s1, s2, s3, RNA.Alphabet.Triples**2)
self.assertEqual(c._data.sum(), 2)
self.assertEqual(c._data[0,1], 1)
self.assertEqual(c._data[63,55], 1)
def test_fromTriple(self):
"""Counts fromTriple should return correct counts."""
cft = Counts.fromTriple
rs = RnaSequence
A, C, G, U = map(rs, 'ACGU')
#counts if different from both the other groups
s = cft(A, C, C, RnaPairs)
assert isinstance(s, Counts)
self.assertEqual(s['C', 'A'], 1)
self.assertEqual(s['A', 'C'], 0)
self.assertEqual(s['C', 'C'], 0)
#try it with longer sequences
AAA, CCC = map(rs, ['AAA', 'CCC'])
s = cft(AAA, CCC, CCC, RnaPairs)
self.assertEqual(s['C', 'A'], 3)
self.assertEqual(s['A', 'C'], 0)
#doesn't count if all three differ
ACG, CGA, GAC = map(rs, ['ACG', 'CGA', 'GAC'])
s = cft(ACG, CGA, GAC, RnaPairs)
self.assertEqual(s['C', 'A'], 0)
self.assertEqual(s['A', 'C'], 0)
self.assertEqual(s, Counts.empty(RnaPairs))
#counts as no change if same as other sequence...
s = cft(AAA, AAA, CCC, RnaPairs)
self.assertEqual(s['A', 'A'], 3)
self.assertEqual(s['A', 'C'], 0)
#...or same as the outgroup
s = cft(AAA, CCC, AAA, RnaPairs)
self.assertEqual(s['A', 'A'], 3)
self.assertEqual(s['A', 'C'], 0)
#spot-check a mixed example
s = cft( \
rs('AUCGCUAGCAUACGUCA'),
rs('AAGCUGCGUAGCGCAUA'),
rs('GCGCAUAUGACGAUAGC'),
RnaPairs
)
vals = [
('U', 'U', 1),
('U', 'C', 0),
('U', 'A', 0),
('U', 'G', 0),
('C', 'U', 0),
('C', 'C', 0),
('C', 'A', 0),
('C', 'G', 1),
('A', 'U', 1),
('A', 'C', 0),
('A', 'A', 4),
('A', 'G', 0),
('G', 'U', 0),
('G', 'C', 1),
('G', 'A', 0),
('G', 'G', 1),
]
for i, j, val in vals:
self.assertFloatEqual(s[i, j], val)
#check a long sequence
s = cft( \
rs('AUCGCUAGCAUACGUCA'*1000),
rs('AAGCUGCGUAGCGCAUA'*1000),
rs('GCGCAUAUGACGAUAGC'*1000),
RnaPairs
)
vals = [
('U', 'U', 1000),
('U', 'C', 0),
('U', 'A', 0),
('U', 'G', 0),
('C', 'U', 0),
('C', 'C', 0),
('C', 'A', 0),
('C', 'G', 1000),
('A', 'U', 1000),
('A', 'C', 0),
('A', 'A', 4000),
('A', 'G', 0),
('G', 'U', 0),
('G', 'C', 1000),
('G', 'A', 0),
('G', 'G', 1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i, j], val)
#check that it works when forced to use both variants of fromTriple
s = cft( \
rs('AUCGCUAGCAUACGUCA'*1000),
rs('AAGCUGCGUAGCGCAUA'*1000),
rs('GCGCAUAUGACGAUAGC'*1000),
RnaPairs,
threshold=0 #forces "large" method
)
vals = [
('U', 'U', 1000),
('U', 'C', 0),
('U', 'A', 0),
('U', 'G', 0),
('C', 'U', 0),
('C', 'C', 0),
('C', 'A', 0),
('C', 'G', 1000),
('A', 'U', 1000),
('A', 'C', 0),
('A', 'A', 4000),
('A', 'G', 0),
('G', 'U', 0),
('G', 'C', 1000),
('G', 'A', 0),
('G', 'G', 1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i, j], val)
s = cft( \
rs('AUCGCUAGCAUACGUCA'*1000),
rs('AAGCUGCGUAGCGCAUA'*1000),
rs('GCGCAUAUGACGAUAGC'*1000),
RnaPairs,
threshold=1e12 #forces "small" method
)
vals = [
('U', 'U', 1000),
('U', 'C', 0),
('U', 'A', 0),
('U', 'G', 0),
('C', 'U', 0),
('C', 'C', 0),
('C', 'A', 0),
('C', 'G', 1000),
('A', 'U', 1000),
('A', 'C', 0),
('A', 'A', 4000),
('A', 'G', 0),
('G', 'U', 0),
('G', 'C', 1000),
('G', 'A', 0),
('G', 'G', 1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i, j], val)
#check that it works for codon seqs
s1 = ModelRnaCodonSequence('UUCGCG')
s2 = ModelRnaCodonSequence('UUUGGG')
s3 = s2
c = Counts.fromTriple(s1, s2, s3, RNA.Alphabet.Triples**2)
self.assertEqual(c._data.sum(), 2)
self.assertEqual(c._data[0, 1], 1)
self.assertEqual(c._data[63, 55], 1)
def test_fromPair(self):
"""Counts fromPair should return correct counts."""
s = Counts.fromPair( RnaSequence('UCCGAUCGAUUAUCGGGUACGUA'), \
RnaSequence('GUCGAGUAUAGCGUACGGCUACG'),
RnaPairs)
assert isinstance(s, Counts)
vals = [
('U', 'U', 0),
('U', 'C', 2.5),
('U', 'A', 1),
('U', 'G', 2.5),
('C', 'U', 2.5),
('C', 'C', 1),
('C', 'A', 1),
('C', 'G', 0.5),
('A', 'U', 1),
('A', 'C', 1),
('A', 'A', 1),
('A', 'G', 2),
('G', 'U', 2.5),
('G', 'C', 0.5),
('G', 'A', 2),
('G', 'G', 2),
]
for i, j, val in vals:
self.assertFloatEqual(s[i, j], val)
#check that it works for big seqs
s = Counts.fromPair( RnaSequence('UCAG'*1000), \
RnaSequence('UGAG'*1000),
RnaPairs)
assert isinstance(s, Counts)
vals = [
('U', 'U', 1000),
('U', 'C', 0),
('U', 'A', 0),
('U', 'G', 0),
('C', 'U', 0),
('C', 'C', 0),
('C', 'A', 0),
('C', 'G', 500),
('A', 'U', 0),
('A', 'C', 0),
('A', 'A', 1000),
('A', 'G', 0),
('G', 'U', 0),
('G', 'C', 500),
('G', 'A', 0),
('G', 'G', 1000),
]
for i, j, val in vals:
self.assertFloatEqual(s[i, j], val)
#check that it works for codon seqs
s1 = ModelRnaCodonSequence('UUCGCG')
s2 = ModelRnaCodonSequence('UUUGGG')
c = Counts.fromPair(s1, s2, RNA.Alphabet.Triples**2)
self.assertEqual(c._data.sum(), 2)
self.assertEqual(c._data[0, 1], 0.5)
self.assertEqual(c._data[1, 0], 0.5)
self.assertEqual(c._data[55, 63], 0.5)
self.assertEqual(c._data[63, 55], 0.5)
