def test_aligned(self):
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = self.seqs,
per_site = False,
aligned = True,
ignore_gaps = True)
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
self.assertEqual(
self.expected[seq1.id][seq2.id],
d)
self.assertEqual(i, 2)
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = self.seqs,
per_site = True,
aligned = True,
ignore_gaps = False)
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
self.assertAlmostEqual(
self.expected[seq1.id][seq2.id] / 6.0,
d)
self.assertEqual(i, 2)
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = self.seqs,
per_site = True,
aligned = True,
ignore_gaps = True)
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
self.assertAlmostEqual(
self.expected[seq1.id][seq2.id] / 4.0,
d)
self.assertEqual(i, 2)
def test_unaligned_muscle(self):
if not functions.which('muscle'):
_LOG.warning('muscle not found... skipping tests.')
return
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = self.seqs,
per_site = False,
aligned = False,
ignore_gaps = True,
aligner_tools = ['muscle'])
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
self.assertEqual(
self.expected[seq1.id][seq2.id],
d)
self.assertEqual(i, 2)
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = self.seqs,
per_site = True,
aligned = False,
ignore_gaps = True,
aligner_tools = ['muscle'])
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
self.assertAlmostEqual(
self.expected[seq1.id][seq2.id] / 4.0,
d)
self.assertEqual(i, 2)
def test_amino_acid_seqs(self):
seqs = [
SeqRecord(Seq('MILV*XQP*'), id='1'),
SeqRecord(Seq('MILV*XQQ*'), id='2'),
SeqRecord(Seq('MILV*XPP*'), id='3'),
]
expected = {}
expected['1'] = {'2': 1, '3': 1}
expected['2'] = {'1': 1, '3': 2}
expected['3'] = {'1': 1, '2': 2}
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = seqs,
alphabet = alphabets.ProteinAlphabet(),
per_site = False,
aligned = True,
ignore_gaps = True)
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
self.assertEqual(drc, None)
self.assertEqual(
expected[seq1.id][seq2.id],
d)
self.assertEqual(i, 2)
def summarize_distances(seq_iter,
sample_size=0,
per_site=True,
aligned=False,
ignore_gaps=True,
alphabet=None,
do_full_alignment=False,
full_alignment_out_path=None,
aligner_tools=['mafft', 'muscle'],
full_aligner_tools=None,
rng=None,
log_frequency=0):
if ((not aligned) and (do_full_alignment)):
if not full_aligner_tools:
full_aligner_tools = aligner_tools
seq_iter = align.align(seq_iter,
tools=full_aligner_tools,
out_path=full_alignment_out_path)
aligned = True
if sample_size > 0:
distance_iter = seqstats.sample_distance_iter(
seq_iter=seq_iter,
sample_size=sample_size,
aligned=aligned,
ignore_gaps=ignore_gaps,
per_site=per_site,
alphabet=alphabet,
aligner_tools=aligner_tools,
rng=rng)
else:
distance_iter = seqstats.pairwise_distance_iter(
seq_iter=seq_iter,
aligned=aligned,
ignore_gaps=ignore_gaps,
per_site=per_site,
alphabet=alphabet,
aligner_tools=aligner_tools)
distances = {}
rev_comp_errors = []
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
if (log_frequency > 0) and (((i + 1) % log_frequency) == 0):
_LOG.info('{0}: Calulating distance for comparison {1}...'.format(
datetime.datetime.now(), (i + 1)))
if (drc != None) and (drc < d):
_LOG.warning('reverse complement of {0} is more similar to '
'{1} ({2:.5f} vs {3:.5f})'.format(
seq1.id, seq2.id, drc, d))
rev_comp_errors.append((seq1.id, seq2.id, d, drc))
if sample_size > 0:
if not seq1.id in distances:
distances[seq1.id] = stats.SampleSummarizer(samples=[d])
continue
distances[seq1.id].add_sample(d)
else:
if not seq1.id in distances:
distances[seq1.id] = stats.SampleSummarizer(samples=[d])
else:
distances[seq1.id].add_sample(d)
if not seq2.id in distances:
distances[seq2.id] = stats.SampleSummarizer(samples=[d])
else:
distances[seq2.id].add_sample(d)
return distances, rev_comp_errors
def summarize_distances(seq_iter,
sample_size = 0,
per_site = True,
aligned = False,
ignore_gaps = True,
alphabet = None,
do_full_alignment = False,
full_alignment_out_path = None,
aligner_tools = ['mafft', 'muscle'],
full_aligner_tools = None,
rng = None,
log_frequency = 0):
if ((not aligned) and (do_full_alignment)):
if not full_aligner_tools:
full_aligner_tools = aligner_tools
seq_iter = align.align(seq_iter,
tools = full_aligner_tools,
out_path = full_alignment_out_path)
aligned = True
if sample_size > 0:
distance_iter = seqstats.sample_distance_iter(
seq_iter = seq_iter,
sample_size = sample_size,
aligned = aligned,
ignore_gaps = ignore_gaps,
per_site = per_site,
alphabet = alphabet,
aligner_tools = aligner_tools,
rng = rng)
else:
distance_iter = seqstats.pairwise_distance_iter(
seq_iter = seq_iter,
aligned = aligned,
ignore_gaps = ignore_gaps,
per_site = per_site,
alphabet = alphabet,
aligner_tools = aligner_tools)
distances = {}
rev_comp_errors = []
for i, (seq1, seq2, d, drc) in enumerate(distance_iter):
if (log_frequency > 0) and (((i + 1) % log_frequency) == 0):
_LOG.info('{0}: Calulating distance for comparison {1}...'.format(
datetime.datetime.now(),
(i + 1)))
if (drc != None) and (drc < d):
_LOG.warning('reverse complement of {0} is more similar to '
'{1} ({2:.5f} vs {3:.5f})'.format(seq1.id, seq2.id, drc, d))
rev_comp_errors.append((seq1.id, seq2.id, d, drc))
if sample_size > 0:
if not seq1.id in distances:
distances[seq1.id] = stats.SampleSummarizer(samples = [d])
continue
distances[seq1.id].add_sample(d)
else:
if not seq1.id in distances:
distances[seq1.id] = stats.SampleSummarizer(samples = [d])
else:
distances[seq1.id].add_sample(d)
if not seq2.id in distances:
distances[seq2.id] = stats.SampleSummarizer(samples = [d])
else:
distances[seq2.id].add_sample(d)
return distances, rev_comp_errors