Python Alignment.kmer_frequencies方法代码示例

# 需要导入模块: from skbio import Alignment [as 别名]
# 或者: from skbio.Alignment import kmer_frequencies [as 别名]

#.........这里部分代码省略.........
    def test_omit_gap_sequences(self):
        expected = self.a2
        self.assertEqual(self.a2.omit_gap_sequences(1.0), expected)
        self.assertEqual(self.a2.omit_gap_sequences(0.20), expected)

        expected = Alignment([self.r2])
        self.assertEqual(self.a2.omit_gap_sequences(0.19), expected)

        self.assertEqual(self.empty.omit_gap_sequences(0.0), self.empty)
        self.assertEqual(self.empty.omit_gap_sequences(0.2), self.empty)
        self.assertEqual(self.empty.omit_gap_sequences(1.0), self.empty)

        # Test to ensure floating point precision bug isn't present. See the
        # tests for Alignment.position_frequencies for more details.
        aln = Alignment([DNA('.' * 33, metadata={'id': 'abc'}),
                         DNA('-' * 33, metadata={'id': 'def'})])
        self.assertEqual(aln.omit_gap_sequences(1 - np.finfo(float).eps),
                         Alignment([]))

    def test_position_counters(self):
        self.assertEqual(self.empty.position_counters(), [])

        self.assertEqual(self.no_positions.position_counters(), [])

        expected = [Counter({'U': 1, 'A': 1}),
                    Counter({'U': 1, 'C': 1}),
                    Counter({'A': 1, 'G': 1}),
                    Counter({'U': 2}),
                    Counter({'-': 1, 'U': 1})]
        self.assertEqual(self.a2.position_counters(), expected)

    def test_position_frequencies(self):
        self.assertEqual(self.empty.position_frequencies(), [])

        self.assertEqual(self.no_positions.position_frequencies(), [])

        expected = [defaultdict(float, {'U': 0.5, 'A': 0.5}),
                    defaultdict(float, {'U': 0.5, 'C': 0.5}),
                    defaultdict(float, {'A': 0.5, 'G': 0.5}),
                    defaultdict(float, {'U': 1.0}),
                    defaultdict(float, {'-': 0.5, 'U': 0.5})]
        self.assertEqual(self.a2.position_frequencies(), expected)

    def test_position_frequencies_floating_point_precision(self):
        # Test that a position with no variation yields a frequency of exactly
        # 1.0. Note that it is important to use self.assertEqual here instead
        # of self.assertAlmostEqual because we want to test for exactly 1.0. A
        # previous implementation of Alignment.position_frequencies added
        # (1 / sequence_count) for each occurrence of a character in a position
        # to compute the frequencies (see
        # https://github.com/biocore/scikit-bio/issues/801). In certain cases,
        # this yielded a frequency slightly less than 1.0 due to roundoff
        # error. The test case here uses an alignment of 10 sequences with no
        # variation at a position. This test case exposes the roundoff error
        # present in the previous implementation because 1/10 added 10 times
        # yields a number slightly less than 1.0. This occurs because 1/10
        # cannot be represented exactly as a floating point number.
        seqs = []
        for i in range(10):
            seqs.append(DNA('A', metadata={'id': str(i)}))
        aln = Alignment(seqs)
        self.assertEqual(aln.position_frequencies(),
                         [defaultdict(float, {'A': 1.0})])

    def test_position_entropies(self):
        # tested by calculating values as described in this post:
        #  http://stackoverflow.com/a/15476958/3424666
        expected = [0.69314, 0.69314, 0.69314, 0.0, np.nan]
        np.testing.assert_almost_equal(self.a2.position_entropies(),
                                       expected, 5)

        expected = [1.0, 1.0, 1.0, 0.0, np.nan]
        np.testing.assert_almost_equal(self.a2.position_entropies(base=2),
                                       expected, 5)

        np.testing.assert_almost_equal(self.empty.position_entropies(base=2),
                                       [])

    def test_kmer_frequencies(self):
        expected = [defaultdict(float, {'U': 3 / 5, 'A': 1 / 5, '-': 1 / 5}),
                    defaultdict(float, {'A': 1 / 5, 'C': 1 / 5, 'G': 1 / 5,
                                        'U': 2 / 5})]
        actual = self.a2.kmer_frequencies(k=1, relative=True)
        for a, e in zip(actual, expected):
            self.assertEqual(sorted(a), sorted(e), 5)
            np.testing.assert_almost_equal(sorted(a.values()),
                                           sorted(e.values()), 5)

    def test_sequence_length(self):
        self.assertEqual(self.a1.sequence_length(), 13)
        self.assertEqual(self.a2.sequence_length(), 5)
        self.assertEqual(self.empty.sequence_length(), 0)

    def test_validate_lengths(self):
        self.assertTrue(self.a1._validate_lengths())
        self.assertTrue(self.a2._validate_lengths())
        self.assertTrue(self.empty._validate_lengths())

        self.assertTrue(Alignment([
            DNA('TTT', metadata={'id': "d1"})])._validate_lengths())