def test_beta_phylogenetic_empty_table(self): t = self.get_data_path('empty.biom') tree = self.get_data_path('three_feature.tree') with self.assertRaisesRegex(ValueError, 'empty'): beta_phylogenetic(table=t, phylogeny=tree, metric='unweighted_unifrac')
def test_beta_phylogenetic_unknown_metric(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read( io.StringIO('((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) with self.assertRaises(ValueError): beta_phylogenetic(table=t, phylogeny=tree, metric='not-a-metric')
def test_beta_phylogenetic_unknown_metric(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read(io.StringIO( '((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) with self.assertRaises(ValueError): beta_phylogenetic(table=t, phylogeny=tree, metric='not-a-metric')
def test_beta_phylogenetic_empty_table(self): t = Table(np.array([]), [], []) tree = skbio.TreeNode.read(io.StringIO( '((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) with self.assertRaisesRegex(ValueError, 'empty'): beta_phylogenetic(table=t, phylogeny=tree, metric='unweighted_unifrac')
def test_beta_phylogenetic_empty_table(self): t = Table(np.array([]), [], []) tree = skbio.TreeNode.read( io.StringIO('((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) with self.assertRaisesRegex(ValueError, 'empty'): beta_phylogenetic(table=t, phylogeny=tree, metric='unweighted_unifrac')
def test_beta_phylogenetic_too_many_jobs(self): bt_fp = self.get_data_path('crawford.biom') tree_fp = self.get_data_path('tree.nwk') with self.assertRaises(ValueError): # cannot guarantee that this will always be true, but it would be # odd to see a machine with these many CPUs beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='unweighted_unifrac', n_jobs=11117)
def test_beta_phylogenetic_skbio_error_rewriting(self): t = self.get_data_path('two_feature_table.biom') tree = self.get_data_path('vaw.nwk') # Verify through regex that there is a ``feature_ids`` substring # followed by a ``phylogeny`` with self.assertRaisesRegex(ValueError, 'represented by the phylogeny'): beta_phylogenetic(table=t, phylogeny=tree, metric='weighted_unifrac')
def test_beta_phylogenetic_alpha_on_non_generalized(self): bt_fp = self.get_data_path('crawford.biom') tree_fp = self.get_data_path('tree.nwk') with self.assertRaisesRegex(ValueError, 'The alpha parameter is only ' 'allowed when the choice of metric is ' 'generalized_unifrac'): beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='unweighted_unifrac', alpha=0.11)
def test_beta_phylogenetic_weighted_unifrac_threads_error(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read(io.StringIO( '((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) with self.assertRaisesRegex(ValueError, 'parallelizable'): beta_phylogenetic(table=t, phylogeny=tree, metric='weighted_unifrac', n_jobs=-1)
def test_beta_phylogenetic_weighted_unifrac_threads_error(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read( io.StringIO('((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) with self.assertRaisesRegex(ValueError, 'parallelizable'): beta_phylogenetic(table=t, phylogeny=tree, metric='weighted_unifrac', n_jobs=-1)
def test_beta_phylogenetic_skbio_error_rewriting(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read( io.StringIO('((O1:0.25):0.25, O3:0.75)root;')) # Verify through regex that there is a ``feature_ids`` substring # followed by a ``phylogeny`` with self.assertRaisesRegex(skbio.tree.MissingNodeError, 'feature_ids.*phylogeny'): beta_phylogenetic(table=t, phylogeny=tree, metric='weighted_unifrac')
def test_beta_phylogenetic_skbio_error_rewriting(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read(io.StringIO( '((O1:0.25):0.25, O3:0.75)root;')) # Verify through regex that there is a ``feature_ids`` substring # followed by a ``phylogeny`` with self.assertRaisesRegex(skbio.tree.MissingNodeError, 'feature_ids.*phylogeny'): beta_phylogenetic(table=t, phylogeny=tree, metric='weighted_unifrac')
def test_generalized_unifrac_no_alpha(self): bt_fp = self.get_data_path('crawford.biom') tree_fp = self.get_data_path('crawford.nwk') actual = beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='generalized_unifrac', alpha=None) # alpha=1 should be equal to weighted normalized UniFrac data = np.array([0.2821874, 0.16148405, 0.20186143, 0.1634832, 0.40351108, 0.29135056, 0.24790944, 0.41967404, 0.24642185, 0.22218489, 0.34007547, 0.27722011, 0.20963881, 0.16897221, 0.3217958, 0.15237816, 0.16899207, 0.36445044, 0.25408941, 0.23358681, 0.4069374, 0.24615927, 0.28573888, 0.20578184, 0.20742006, 0.31249151, 0.46169893, 0.35294595, 0.32522355, 0.48437103, 0.21534558, 0.30558908, 0.12091004, 0.19817777, 0.24792853, 0.34293674]) ids = ('10084.PC.481', '10084.PC.593', '10084.PC.356', '10084.PC.355', '10084.PC.354', '10084.PC.636', '10084.PC.635', '10084.PC.607', '10084.PC.634') expected = skbio.DistanceMatrix(data, ids=ids) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def test_generalized_unifrac(self): bt_fp = self.get_data_path('vaw.biom') tree_fp = self.get_data_path('vaw.nwk') actual = beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='generalized_unifrac', alpha=0.5) data = np.array([[0.0000000, 0.4040518, 0.6285560, 0.5869439, 0.4082483, 0.2995673], [0.4040518, 0.0000000, 0.4160597, 0.7071068, 0.7302479, 0.4860856], [0.6285560, 0.4160597, 0.0000000, 0.8005220, 0.9073159, 0.5218198], [0.5869439, 0.7071068, 0.8005220, 0.0000000, 0.4117216, 0.3485667], [0.4082483, 0.7302479, 0.9073159, 0.4117216, 0.0000000, 0.6188282], [0.2995673, 0.4860856, 0.5218198, 0.3485667, 0.6188282, 0.0000000]]) ids = ('Sample1', 'Sample2', 'Sample3', 'Sample4', 'Sample5', 'Sample6') expected = skbio.DistanceMatrix(data, ids=ids) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def test_variance_adjusted_normalized(self): bt_fp = self.get_data_path('vaw.biom') tree_fp = self.get_data_path('vaw.nwk') actual = beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='weighted_normalized_unifrac', variance_adjusted=True) data = np.array([[0.0000000, 0.4086040, 0.6240185, 0.4639481, 0.2857143, 0.2766318], [0.4086040, 0.0000000, 0.3798594, 0.6884992, 0.6807616, 0.4735781], [0.6240185, 0.3798594, 0.0000000, 0.7713254, 0.8812897, 0.5047114], [0.4639481, 0.6884992, 0.7713254, 0.0000000, 0.6666667, 0.2709298], [0.2857143, 0.6807616, 0.8812897, 0.6666667, 0.0000000, 0.4735991], [0.2766318, 0.4735781, 0.5047114, 0.2709298, 0.4735991, 0.0000000]]) ids = ('Sample1', 'Sample2', 'Sample3', 'Sample4', 'Sample5', 'Sample6') expected = skbio.DistanceMatrix(data, ids=ids) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def test_beta_weighted(self): bt_fp = self.get_data_path('crawford.biom') tree_fp = self.get_data_path('crawford.nwk') actual = beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='weighted_unifrac') # computed with beta-phylogenetic (weighted_unifrac) data = np.array([0.44656238, 0.23771096, 0.30489123, 0.23446002, 0.65723575, 0.44911772, 0.381904, 0.69144829, 0.39611776, 0.36568012, 0.53377975, 0.48908025, 0.35155196, 0.28318669, 0.57376916, 0.23395746, 0.24658122, 0.60271637, 0.39802552, 0.36567394, 0.68062701, 0.36862049, 0.48350632, 0.33024631, 0.33266697, 0.53464744, 0.74605075, 0.53951035, 0.49680733, 0.79178838, 0.37109012, 0.52629343, 0.22118218, 0.32400805, 0.43189708, 0.59705893]) ids = ('10084.PC.481', '10084.PC.593', '10084.PC.356', '10084.PC.355', '10084.PC.354', '10084.PC.636', '10084.PC.635', '10084.PC.607', '10084.PC.634') expected = skbio.DistanceMatrix(data, ids=ids) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def test_beta_unweighted_parallel(self): bt_fp = self.get_data_path('crawford.biom') tree_fp = self.get_data_path('crawford.nwk') actual = beta_phylogenetic(table=bt_fp, phylogeny=tree_fp, metric='unweighted_unifrac', n_jobs=2) # computed with beta-phylogenetic data = np.array([0.71836067, 0.71317361, 0.69746044, 0.62587207, 0.72826674, 0.72065895, 0.72640581, 0.73606053, 0.70302967, 0.73407301, 0.6548042, 0.71547381, 0.78397813, 0.72318399, 0.76138933, 0.61041275, 0.62331299, 0.71848305, 0.70416337, 0.75258475, 0.79249029, 0.64392779, 0.70052733, 0.69832716, 0.77818938, 0.72959894, 0.75782689, 0.71005144, 0.75065046, 0.78944369, 0.63593642, 0.71283615, 0.58314638, 0.69200762, 0.68972056, 0.71514083]) ids = ('10084.PC.481', '10084.PC.593', '10084.PC.356', '10084.PC.355', '10084.PC.354', '10084.PC.636', '10084.PC.635', '10084.PC.607', '10084.PC.634') expected = skbio.DistanceMatrix(data, ids=ids) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def core_metrics(table: biom.Table, phylogeny: skbio.TreeNode, sampling_depth: int) -> (pd.Series, pd.Series, pd.Series, pd.Series, skbio.DistanceMatrix, skbio.DistanceMatrix, skbio.DistanceMatrix, skbio.DistanceMatrix, skbio.OrdinationResults, skbio.OrdinationResults, skbio.OrdinationResults, skbio.OrdinationResults): rarefied_table = rarefy(table=table, sampling_depth=sampling_depth) faith_pd_vector = alpha_phylogenetic( table=rarefied_table, phylogeny=phylogeny, metric='faith_pd') observed_otus_vector = alpha(table=rarefied_table, metric='observed_otus') shannon_vector = alpha(table=rarefied_table, metric='shannon') evenness_vector = alpha(table=rarefied_table, metric='pielou_e') unweighted_unifrac_distance_matrix = beta_phylogenetic( table=rarefied_table, phylogeny=phylogeny, metric='unweighted_unifrac') weighted_unifrac_distance_matrix = beta_phylogenetic( table=rarefied_table, phylogeny=phylogeny, metric='weighted_unifrac') jaccard_distance_matrix = beta(table=rarefied_table, metric='jaccard') bray_curtis_distance_matrix = beta( table=rarefied_table, metric='braycurtis') unweighted_unifrac_pcoa_results = pcoa( distance_matrix=unweighted_unifrac_distance_matrix) weighted_unifrac_pcoa_results = pcoa( distance_matrix=weighted_unifrac_distance_matrix) jaccard_pcoa_results = pcoa(distance_matrix=jaccard_distance_matrix) bray_curtis_pcoa_results = pcoa( distance_matrix=bray_curtis_distance_matrix) return ( faith_pd_vector, observed_otus_vector, shannon_vector, evenness_vector, unweighted_unifrac_distance_matrix, weighted_unifrac_distance_matrix, jaccard_distance_matrix, bray_curtis_distance_matrix, unweighted_unifrac_pcoa_results, weighted_unifrac_pcoa_results, jaccard_pcoa_results, bray_curtis_pcoa_results )
def core_metrics( table: biom.Table, phylogeny: skbio.TreeNode, sampling_depth: int ) -> (pd.Series, pd.Series, pd.Series, pd.Series, skbio.DistanceMatrix, skbio.DistanceMatrix, skbio.DistanceMatrix, skbio.DistanceMatrix, skbio.OrdinationResults, skbio.OrdinationResults, skbio.OrdinationResults, skbio.OrdinationResults): rarefied_table = rarefy(table=table, sampling_depth=sampling_depth) faith_pd_vector = alpha_phylogenetic(table=rarefied_table, phylogeny=phylogeny, metric='faith_pd') observed_otus_vector = alpha(table=rarefied_table, metric='observed_otus') shannon_vector = alpha(table=rarefied_table, metric='shannon') evenness_vector = alpha(table=rarefied_table, metric='pielou_e') unweighted_unifrac_distance_matrix = beta_phylogenetic( table=rarefied_table, phylogeny=phylogeny, metric='unweighted_unifrac') weighted_unifrac_distance_matrix = beta_phylogenetic( table=rarefied_table, phylogeny=phylogeny, metric='weighted_unifrac') jaccard_distance_matrix = beta(table=rarefied_table, metric='jaccard') bray_curtis_distance_matrix = beta(table=rarefied_table, metric='braycurtis') unweighted_unifrac_pcoa_results = pcoa( distance_matrix=unweighted_unifrac_distance_matrix) weighted_unifrac_pcoa_results = pcoa( distance_matrix=weighted_unifrac_distance_matrix) jaccard_pcoa_results = pcoa(distance_matrix=jaccard_distance_matrix) bray_curtis_pcoa_results = pcoa( distance_matrix=bray_curtis_distance_matrix) return (faith_pd_vector, observed_otus_vector, shannon_vector, evenness_vector, unweighted_unifrac_distance_matrix, weighted_unifrac_distance_matrix, jaccard_distance_matrix, bray_curtis_distance_matrix, unweighted_unifrac_pcoa_results, weighted_unifrac_pcoa_results, jaccard_pcoa_results, bray_curtis_pcoa_results)
def test_beta_phylogenetic(self): t = self.get_data_path('two_feature_table.biom') tree = self.get_data_path('three_feature.tree') actual = beta_phylogenetic( table=t, phylogeny=tree, metric='unweighted_unifrac') # expected computed with skbio.diversity.beta_diversity expected = skbio.DistanceMatrix([[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]], ids=['S1', 'S2', 'S3']) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def test_beta_phylogenetic(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read( io.StringIO('((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) actual = beta_phylogenetic(table=t, phylogeny=tree, metric='unweighted_unifrac') # expected computed with skbio.diversity.beta_diversity expected = skbio.DistanceMatrix( [[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]], ids=['S1', 'S2', 'S3']) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])
def test_beta_phylogenetic(self): t = Table(np.array([[0, 1, 3], [1, 1, 2]]), ['O1', 'O2'], ['S1', 'S2', 'S3']) tree = skbio.TreeNode.read(io.StringIO( '((O1:0.25, O2:0.50):0.25, O3:0.75)root;')) actual = beta_phylogenetic( table=t, phylogeny=tree, metric='unweighted_unifrac') # expected computed with skbio.diversity.beta_diversity expected = skbio.DistanceMatrix([[0.00, 0.25, 0.25], [0.25, 0.00, 0.00], [0.25, 0.00, 0.00]], ids=['S1', 'S2', 'S3']) self.assertEqual(actual.ids, expected.ids) for id1 in actual.ids: for id2 in actual.ids: npt.assert_almost_equal(actual[id1, id2], expected[id1, id2])