def test_reverse_transcribe_preserves_all_metadata(self):
im = IntervalMetadata(4)
im.add([(0, 2)], metadata={'gene': 'p53'})
seq = RNA('AGUU', metadata={'foo': 'bar'},
positional_metadata={'foo': range(4)},
interval_metadata=im)
exp = DNA('AGTT', metadata={'foo': 'bar'},
positional_metadata={'foo': range(4)},
interval_metadata=im)
self.assertEqual(seq.reverse_transcribe(), exp)
def test_translate_six_frames(self):
seq = RNA('AUGCUAACAUAAA') # rc = UUUAUGUUAGCAU
# test default behavior
exp = [
Protein('MLT*'),
Protein('C*HK'),
Protein('ANI'),
Protein('FMLA'),
Protein('LC*H'),
Protein('YVS')
]
obs = list(self.sgc.translate_six_frames(seq))
self.assertEqual(obs, exp)
# test that start/stop are respected
exp = [
Protein('MLT'),
Protein('C'),
Protein('ANI'),
Protein('MLA'),
Protein('LC'),
Protein('YVS')
]
obs = list(
self.sgc.translate_six_frames(seq,
start='optional',
stop='optional'))
self.assertEqual(obs, exp)
def test_translate_trim_to_cds(self):
seq = RNA('UAAUUGCCUCAUUAAUAACAAUGA')
# find first start codon, trim all before it, convert alternative start
# codon to M, finally trim to first stop codon following the start
# codon
exp = Protein('MPH')
for param in {'require', 'optional'}:
obs = self.sgc.translate(seq, start=param, stop=param)
self.assertEqual(obs, exp)
exp = Protein('*LPH**Q*')
obs = self.sgc.translate(seq, start='ignore', stop='ignore')
self.assertEqual(obs, exp)
# alternative reading frame disrupts cds:
# AAUUGCCUCAUUAAUAACAAUGA
# NCLINNN
with self.assertRaisesRegex(ValueError,
r'reading_frame=2.*start=\'require\''):
self.sgc.translate(seq, reading_frame=2, start='require')
with self.assertRaisesRegex(ValueError,
r'reading_frame=2.*stop=\'require\''):
self.sgc.translate(seq, reading_frame=2, stop='require')
exp = Protein('NCLINNN')
for param in {'ignore', 'optional'}:
obs = self.sgc.translate(seq,
reading_frame=2,
start=param,
stop=param)
self.assertEqual(obs, exp)
def test_translate_varied_genetic_codes(self):
# spot check using a few NCBI and custom genetic codes to translate
seq = RNA('AAUGAUGUGACUAUCAGAAGG')
# table_id=2
exp = Protein('NDVTI**')
obs = GeneticCode.from_ncbi(2).translate(seq)
self.assertEqual(obs, exp)
exp = Protein('MTI')
obs = GeneticCode.from_ncbi(2).translate(seq,
start='require',
stop='require')
self.assertEqual(obs, exp)
# table_id=22
exp = Protein('NDVTIRR')
obs = GeneticCode.from_ncbi(22).translate(seq)
self.assertEqual(obs, exp)
with self.assertRaisesRegex(ValueError,
r'reading_frame=1.*start=\'require\''):
GeneticCode.from_ncbi(22).translate(seq,
start='require',
stop='require')
# custom, no start codons
gc = GeneticCode('MWN*' * 16, '-' * 64)
exp = Protein('MM*MWN*')
obs = gc.translate(seq)
self.assertEqual(obs, exp)
with self.assertRaisesRegex(ValueError,
r'reading_frame=1.*start=\'require\''):
gc.translate(seq, start='require', stop='require')
def compute_distance_matrix(msa_file, csvfile="distance_mat.csv"):
"""
load up some aligned sequences, and compute a distance matrix
compute distances between the sequences using the hamming function
see also:
scipy.spatial.distance.hamming
@args msa_file: multiple sequence alignment in fasta format
@type msa_file: str
@args csvfile: output distance matrix file in csv format
@type csvfile: str
"""
records = []
for rec in SeqIO.parse(msa_file, "fasta"):
records.append(RNA(rec.seq, rec.id))
aln = Alignment(records)
master_dm = aln.distances()
## writing the result to a csv file
csv_header_row = [header for header in master_dm.ids]
## result as a list of list
with open(csvfile, "w") as output:
writer = csv.writer(output, lineterminator="\n")
writer.writerows([csv_header_row])
writer.writerows(master_dm)
output.close()
def test_translate_reading_frame_non_empty_translation(self):
seq = RNA('AUGGUGGAA') # rc = UUCCACCAU
for reading_frame, exp_str in ((1, 'MVE'), (2, 'WW'), (3, 'GG'),
(-1, 'FHH'), (-2, 'ST'), (-3, 'PP')):
exp = Protein(exp_str)
obs = self.sgc.translate(seq, reading_frame=reading_frame)
self.assertEqual(obs, exp)
def test_translate_preserves_metadata(self):
obs = self.sgc.translate(
RNA('AUG', metadata={'foo': 'bar', 'baz': 42},
positional_metadata={'foo': range(3)}))
# metadata retained, positional metadata dropped
self.assertEqual(obs, Protein('M',
metadata={'foo': 'bar', 'baz': 42}))
def test_update_ids_default_behavior(self):
# 3 seqs
exp_sc = SequenceCollection([
RNA('GAUUACA', id="1"),
RNA('UUG', id="2"),
RNA('U-----UGCC--', id="3")
])
exp_id_map = {'1': 'r1', '2': 'r2', '3': 'r3'}
obs_sc, obs_id_map = self.s2.update_ids()
self._assert_sequence_collections_equal(obs_sc, exp_sc)
self.assertEqual(obs_id_map, exp_id_map)
# empty
obs_sc, obs_id_map = self.empty.update_ids()
self._assert_sequence_collections_equal(obs_sc, self.empty)
self.assertEqual(obs_id_map, {})
def test_update_ids_ids_parameter(self):
# 3 seqs
exp_sc = SequenceCollection([
RNA('GAUUACA', id="abc"),
RNA('UUG', id="def"),
RNA('U-----UGCC--', id="ghi")
])
exp_id_map = {'abc': 'r1', 'def': 'r2', 'ghi': 'r3'}
obs_sc, obs_id_map = self.s2.update_ids(ids=('abc', 'def', 'ghi'))
self._assert_sequence_collections_equal(obs_sc, exp_sc)
self.assertEqual(obs_id_map, exp_id_map)
# empty
obs_sc, obs_id_map = self.empty.update_ids(ids=[])
self._assert_sequence_collections_equal(obs_sc, self.empty)
self.assertEqual(obs_id_map, {})
def test_constructor_not_monomorphic(self):
with six.assertRaisesRegex(self, TypeError, 'mixed types.*RNA.*DNA'):
TabularMSA([DNA(''), RNA('')])
with six.assertRaisesRegex(self, TypeError,
'mixed types.*float.*Protein'):
TabularMSA([Protein(''), Protein(''), 42.0, Protein('')])
def test_update_ids_prefix(self):
# 3 seqs
exp_sc = SequenceCollection([
RNA('GAUUACA', id="abc1"),
RNA('UUG', id="abc2"),
RNA('U-----UGCC--', id="abc3")
])
exp_id_map = {'abc1': 'r1', 'abc2': 'r2', 'abc3': 'r3'}
obs_sc, obs_id_map = self.s2.update_ids(prefix='abc')
self._assert_sequence_collections_equal(obs_sc, exp_sc)
self.assertEqual(obs_id_map, exp_id_map)
# empty
obs_sc, obs_id_map = self.empty.update_ids(prefix='abc')
self._assert_sequence_collections_equal(obs_sc, self.empty)
self.assertEqual(obs_id_map, {})
def test_translate_ncbi_table_id(self):
for seq in RNA('AAAUUUAUGCAU'), DNA('AAATTTATGCAT'):
# default
obs = seq.translate()
self.assertEqual(obs, Protein('KFMH'))
obs = seq.translate(9)
self.assertEqual(obs, Protein('NFMH'))
def test_translate_six_frames_preserves_metadata(self):
seq = RNA('AUG', metadata={'foo': 'bar', 'baz': 42},
positional_metadata={'foo': range(3)})
obs = list(self.sgc.translate_six_frames(seq))[:2]
# metadata retained, positional metadata dropped
self.assertEqual(
obs,
[Protein('M', metadata={'foo': 'bar', 'baz': 42}),
Protein('', metadata={'foo': 'bar', 'baz': 42})])
def setUp(self):
self.d1 = DNA('GATTACA', metadata={'id': "d1"})
self.d2 = DNA('TTG', metadata={'id': "d2"})
self.d3 = DNA('GTATACA', metadata={'id': "d3"})
self.r1 = RNA('GAUUACA', metadata={'id': "r1"})
self.r2 = RNA('UUG', metadata={'id': "r2"})
self.r3 = RNA('U-----UGCC--', metadata={'id': "r3"})
self.seqs1 = [self.d1, self.d2]
self.seqs2 = [self.r1, self.r2, self.r3]
self.seqs3 = self.seqs1 + self.seqs2
self.seqs4 = [self.d1, self.d3]
self.s1 = SequenceCollection(self.seqs1)
self.s2 = SequenceCollection(self.seqs2)
self.s3 = SequenceCollection(self.seqs3)
self.s4 = SequenceCollection(self.seqs4)
self.empty = SequenceCollection([])
def setUp(self):
# ids all same length, seqs longer than 10 chars
dna_3_seqs = Alignment([
DNA('..ACC-GTTGG..', metadata={'id': "d1"}),
DNA('TTACCGGT-GGCC', metadata={'id': "d2"}),
DNA('.-ACC-GTTGC--', metadata={'id': "d3"})])
# id lengths from 0 to 10, with mixes of numbers, characters, and
# spaces. sequence characters are a mix of cases and gap characters.
# sequences are shorter than 10 chars
variable_length_ids = Alignment([
RNA('.-ACGU', metadata={'id': ''}),
RNA('UGCA-.', metadata={'id': 'a'}),
RNA('.ACGU-', metadata={'id': 'bb'}),
RNA('ugca-.', metadata={'id': '1'}, validate=False),
RNA('AaAaAa', metadata={'id': 'abcdefghij'}, validate=False),
RNA('GGGGGG', metadata={'id': 'ab def42ij'})])
# sequences with 20 chars = exactly two chunks of size 10
two_chunks = Alignment([
DNA('..ACC-GTTGG..AATGC.C', metadata={'id': 'foo'}),
DNA('TTACCGGT-GGCCTA-GCAT', metadata={'id': 'bar'})])
# single sequence with more than two chunks
single_seq_long = Alignment([
DNA('..ACC-GTTGG..AATGC.C----', metadata={'id': 'foo'})])
# single sequence with only a single character (minimal writeable
# alignment)
single_seq_short = Alignment([DNA('-', metadata={'id': ''})])
# alignments that can be written in phylip format
self.objs = [dna_3_seqs, variable_length_ids, two_chunks,
single_seq_long, single_seq_short]
self.fps = map(get_data_path,
['phylip_dna_3_seqs', 'phylip_variable_length_ids',
'phylip_two_chunks', 'phylip_single_seq_long',
'phylip_single_seq_short'])
# alignments that cannot be written in phylip format, paired with their
# expected error message regexps
self.invalid_objs = [
# no seqs
(Alignment([]), 'one sequence'),
# no positions
(Alignment([DNA('', metadata={'id': "d1"}),
DNA('', metadata={'id': "d2"})]), 'one position'),
# ids too long
(Alignment([RNA('ACGU', metadata={'id': "foo"}),
RNA('UGCA', metadata={'id': "alongsequenceid"})]),
'10.*alongsequenceid')
]
def test_bool(self):
self.assertFalse(TabularMSA([]))
self.assertFalse(TabularMSA([RNA('')]))
self.assertFalse(
TabularMSA(
[RNA('', metadata={'id': 1}),
RNA('', metadata={'id': 2})],
key='id'))
self.assertTrue(TabularMSA([RNA('U')]))
self.assertTrue(TabularMSA([RNA('--'), RNA('..')]))
self.assertTrue(TabularMSA([RNA('AUC'), RNA('GCA')]))
def test_update_ids_fn_parameter(self):
def append_42(ids):
return [id_ + '-42' for id_ in ids]
# 3 seqs
exp_sc = SequenceCollection([
RNA('GAUUACA', id="r1-42"),
RNA('UUG', id="r2-42"),
RNA('U-----UGCC--', id="r3-42")
])
exp_id_map = {'r1-42': 'r1', 'r2-42': 'r2', 'r3-42': 'r3'}
obs_sc, obs_id_map = self.s2.update_ids(fn=append_42)
self._assert_sequence_collections_equal(obs_sc, exp_sc)
self.assertEqual(obs_id_map, exp_id_map)
# empty
obs_sc, obs_id_map = self.empty.update_ids(fn=append_42)
self._assert_sequence_collections_equal(obs_sc, self.empty)
self.assertEqual(obs_id_map, {})
def test_translate_stop_no_stop_codon(self):
seq = RNA('GAAUCU')
exp = Protein('ES')
for stop in {'ignore', 'optional'}:
obs = self.sgc.translate(seq, stop=stop)
self.assertEqual(obs, exp)
with six.assertRaisesRegex(self, ValueError,
'reading_frame=1.*stop=\'require\''):
self.sgc.translate(seq, stop='require')
def test_iter_positions(self):
actual = list(self.a2.iter_positions())
expected = [
[RNA('U', metadata={'id': 'r1'}),
RNA('A', metadata={'id': 'r2'})],
[RNA('U', metadata={'id': 'r1'}),
RNA('C', metadata={'id': 'r2'})],
[RNA('A', metadata={'id': 'r1'}),
RNA('G', metadata={'id': 'r2'})],
[RNA('U', metadata={'id': 'r1'}),
RNA('U', metadata={'id': 'r2'})],
[RNA('-', metadata={'id': 'r1'}),
RNA('U', metadata={'id': 'r2'})]
]
self.assertEqual(actual, expected)
actual = list(self.a2.iter_positions(constructor=str))
expected = [list('UA'), list('UC'), list('AG'), list('UU'), list('-U')]
self.assertEqual(actual, expected)
def test_multiple_msa_file(self):
fp = get_data_path('stockholm_multiple_msa')
msa = _stockholm_to_tabular_msa(fp, constructor=RNA)
exp = TabularMSA([
RNA('AAGGGUUAUUUAUAUACUUU'),
RNA('UGCUAAGAGUGGGGAUGAUU'),
RNA('GCCACAACCGAUUAGAUAGA'),
RNA('UUAGAAACCGAUGGACCGAA')
],
metadata={
'AC': 'G2134T23',
'ID': 'ARD'
},
positional_metadata=({
'AC_cons':
list('GGGACUGGACAUCUAUUCAG')
}),
index=['RTC2231', 'RTF2124', 'RTH3322', 'RTB1512'])
self.assertEqual(msa, exp)
def test_translate_six_frames_ncbi_table_id(self):
# rc = CAAUUU
for seq in RNA('AAAUUG'), DNA('AAATTG'):
# default
obs = list(seq.translate_six_frames())
self.assertEqual(obs, [Protein('KL'), Protein('N'), Protein('I'),
Protein('QF'), Protein('N'), Protein('I')])
obs = list(seq.translate_six_frames(9))
self.assertEqual(obs, [Protein('NL'), Protein('N'), Protein('I'),
Protein('QF'), Protein('N'), Protein('I')])
def test_translate_preserves_metadata(self):
metadata = {'foo': 'bar', 'baz': 42}
positional_metadata = {'foo': range(3)}
for seq in (RNA('AUG', metadata=metadata,
positional_metadata=positional_metadata),
DNA('ATG', metadata=metadata,
positional_metadata=positional_metadata)):
obs = seq.translate()
# metadata retained, positional metadata dropped
self.assertEqual(obs,
Protein('M', metadata={'foo': 'bar', 'baz': 42}))
def test_transcribe_preserves_all_metadata(self):
im = IntervalMetadata(4)
im.add([(0, 2)], metadata={'gene': 'p53'})
exp = RNA('AGUU', metadata={'foo': 'bar'},
positional_metadata={'foo': range(4)},
interval_metadata=im)
seq = DNA('AGTT', metadata={'foo': 'bar'},
positional_metadata={'foo': range(4)},
interval_metadata=im)
self.assertEqual(seq.transcribe(), exp)
def test_translate_six_frames_genetic_code_object(self):
gc = GeneticCode('M' * 64, '-' * 64)
for seq in RNA('AAAUUG'), DNA('AAATTG'):
obs = list(seq.translate_six_frames(gc))
self.assertEqual(obs, [
Protein('MM'),
Protein('M'),
Protein('M'),
Protein('MM'),
Protein('M'),
Protein('M')
])
def test_translate_six_frames_preserves_metadata(self):
metadata = {'foo': 'bar', 'baz': 42}
positional_metadata = {'foo': range(3)}
for seq in (RNA('AUG', metadata=metadata,
positional_metadata=positional_metadata),
DNA('ATG', metadata=metadata,
positional_metadata=positional_metadata)):
obs = list(seq.translate_six_frames())[:2]
# metadata retained, positional metadata dropped
self.assertEqual(
obs,
[Protein('M', metadata={'foo': 'bar', 'baz': 42}),
Protein('', metadata={'foo': 'bar', 'baz': 42})])
def test_translate_start_no_accidental_mutation(self):
# `start` mutates a vector in-place that is derived from
# GeneticCode._offset_table. the current code doesn't perform an
# explicit copy because numpy's advanced indexing is used, which always
# returns a copy. test this assumption here in case that behavior
# changes in the future
offset_table = self.sgc._offset_table.copy()
seq = RNA('CAUUUGCUGAAAUGA')
obs = self.sgc.translate(seq, start='require')
self.assertEqual(obs, Protein('MLK*'))
npt.assert_array_equal(self.sgc._offset_table, offset_table)
def test_transcribe(self):
# without changes
self.assertEqual(DNA('').transcribe(), RNA(''))
self.assertEqual(DNA('A').transcribe(), RNA('A'))
self.assertEqual(DNA('.ACGW-').transcribe(), RNA('.ACGW-'))
# with changes
self.assertEqual(DNA('T').transcribe(), RNA('U'))
self.assertEqual(DNA('TT').transcribe(), RNA('UU'))
self.assertEqual(DNA('ATCTG').transcribe(), RNA('AUCUG'))
self.assertEqual(DNA('TTTG').transcribe(), RNA('UUUG'))
def test_kmer_frequencies(self):
expected1 = Counter({'GAT': 1, 'TAC': 1})
expected2 = Counter({'TTG': 1})
self.assertEqual(
self.s1.kmer_frequencies(k=3, overlap=False, relative=False),
[expected1, expected2])
expected1 = defaultdict(float)
expected1['A'] = 3 / 7.
expected1['C'] = 1 / 7.
expected1['G'] = 1 / 7.
expected1['T'] = 2 / 7.
expected2 = defaultdict(float)
expected2['G'] = 1 / 3.
expected2['T'] = 2 / 3.
self.assertEqual(self.s1.kmer_frequencies(k=1, relative=True),
[expected1, expected2])
expected1 = defaultdict(float)
expected1['GAT'] = 1 / 2.
expected1['TAC'] = 1 / 2.
expected2 = defaultdict(float)
expected2['TTG'] = 1 / 1.
self.assertEqual(
self.s1.kmer_frequencies(k=3, overlap=False, relative=True),
[expected1, expected2])
self.assertEqual(self.empty.kmer_frequencies(k=1, relative=True), [])
# Test to ensure floating point precision bug isn't present. See the
# tests for Sequence.kmer_frequencies for more details.
sc = SequenceCollection([
RNA('C' * 10, metadata={'id': 's1'}),
RNA('G' * 10, metadata={'id': 's2'})
])
self.assertEqual(
sc.kmer_frequencies(1, relative=True),
[defaultdict(float, {'C': 1.0}),
defaultdict(float, {'G': 1.0})])
def test_motif_pyrimidine_run(self):
seq = RNA("")
self.assertEqual(list(seq.find_motifs("pyrimidine-run")), [])
seq = RNA("AARC--UCRG")
self.assertEqual(list(seq.find_motifs("pyrimidine-run")),
[slice(3, 4), slice(6, 8)])
seq = RNA("AA-RC--UCR-G")
self.assertEqual(list(seq.find_motifs("pyrimidine-run", min_length=3,
ignore=seq.gaps())),
[slice(4, 9)])
def test_motif_purine_run(self):
seq = RNA("")
self.assertEqual(list(seq.find_motifs("purine-run")), [])
seq = RNA("AARC--UCRG")
self.assertEqual(list(seq.find_motifs("purine-run")),
[slice(0, 3), slice(8, 10)])
seq = RNA("AA-RC--UCR-G")
self.assertEqual(list(seq.find_motifs("purine-run", min_length=3,
ignore=seq.gaps())),
[slice(0, 4)])
def test_translate_start_no_start_codon(self):
seq = RNA('CAACAACAGCAA')
exp = Protein('QQQQ')
for start in {'ignore', 'optional'}:
obs = self.sgc.translate(seq, start=start)
self.assertEqual(obs, exp)
with six.assertRaisesRegex(self, ValueError,
'reading_frame=1.*start=\'require\''):
self.sgc.translate(seq, start='require')
# non-start codon that translates to an AA that start codons also map
# to. should catch bug if code attempts to search and trim *after*
# translation -- this must happen *before* translation
seq = RNA('UUACAA')
exp = Protein('LQ')
for start in {'ignore', 'optional'}:
obs = self.sgc.translate(seq, start=start)
self.assertEqual(obs, exp)
with six.assertRaisesRegex(self, ValueError,
'reading_frame=1.*start=\'require\''):
self.sgc.translate(seq, start='require')
def rnaAlign(seq1, seq2, gap_open_penalty, gap_extend_penalty, local=False):
seq1 = seq1.upper()
seq2 = seq2.upper()
if local:
aln, score, _ = local_pairwise_align_nucleotide(
RNA(seq1), RNA(seq2), gap_open_penalty, gap_extend_penalty)
else:
aln, score, _ = global_pairwise_align_nucleotide(
RNA(seq1), RNA(seq2), gap_open_penalty, gap_extend_penalty)
response = {
'aln1':
str(aln[0]),
'aln2':
str(aln[1]),
'score':
score,
'similarity':
float('{:.2f}'.format(aln[0].match_frequency(aln[1], relative=True) *
100))
}
return response
def test_degenerate_map(self):
exp = {
'B': set(['C', 'U', 'G']),
'D': set(['A', 'U', 'G']),
'H': set(['A', 'C', 'U']),
'K': set(['U', 'G']),
'M': set(['A', 'C']),
'N': set(['A', 'C', 'U', 'G']),
'S': set(['C', 'G']),
'R': set(['A', 'G']),
'W': set(['A', 'U']),
'V': set(['A', 'C', 'G']),
'Y': set(['C', 'U'])
}
self.assertEqual(RNA('').degenerate_map, exp)
self.assertEqual(RNA.degenerate_map, exp)
def test_reverse_transcribe_does_not_modify_input(self):
seq = RNA('AUAU')
self.assertEqual(seq.reverse_transcribe(), DNA('ATAT'))
self.assertEqual(seq, RNA('AUAU'))
def test_reverse_transcribe_preserves_all_metadata(self):
seq = RNA('AGUU', metadata={'foo': 'bar'},
positional_metadata={'foo': range(4)})
exp = DNA('AGTT', metadata={'foo': 'bar'},
positional_metadata={'foo': range(4)})
self.assertEqual(seq.reverse_transcribe(), exp)