Loading deepchem/data/tests/test_fasta_loader.py +5 −3 Original line number Diff line number Diff line Loading @@ -9,6 +9,7 @@ __license__ = "MIT" import os import unittest import deepchem as dc Loading @@ -26,7 +27,8 @@ class TestFASTALoader(unittest.TestCase): "../../data/tests/example.fasta") loader = dc.data.FASTALoader() sequences = loader.featurize(input_file) # example.fasta contains 3 sequences each of length 58. # The one-hot encoding turns base-pairs into vectors of length 4. # There is one "image channel") assert sequences.X.shape == (3, 4, 58, 1) # The one-hot encoding turns base-pairs into vectors of length 5 (ATCGN). # There is one "image channel". assert sequences.X.shape == (3, 5, 58, 1) deepchem/utils/save.py +29 −7 Original line number Diff line number Diff line Loading @@ -110,14 +110,15 @@ def load_csv_files(filenames, shard_size=None, verbose=True): def seq_one_hot_encode(sequences, letters='ATCGN'): """One hot encodes list of genomic sequences. Sequences encoded have shape (N_sequences, 4, sequence_length, 1). Here 4 is for the 4 basepairs (ACGT) present in genomic sequences. Sequences encoded have shape (N_sequences, N_letters, sequence_length, 1). These sequences will be processed as images with one color channel. Parameters ---------- sequences: np.ndarray Array of genetic sequences letters: str String with the set of possible letters in the sequences. Raises ------ Loading @@ -126,7 +127,7 @@ def seq_one_hot_encode(sequences, letters='ATCGN'): Returns ------- np.ndarray: Shape (N_sequences, 4, sequence_length, 1). np.ndarray: Shape (N_sequences, N_letters, sequence_length, 1). """ # The label encoder is given characters for ACGTN Loading Loading @@ -168,9 +169,11 @@ def _seq_to_encoded(seq, letter_encoder, alphabet_length, sequence_length): return b # This could just be ambiguous_dna_letters, but that would be much higher dim. class IUPACUnambiguousDNAWithN(Alphabet): letters = unambiguous_dna_letters + "N" """ A deepchem specific DNA Alphabet with the 4 main letters and N. """ letters = "ACGTN" def encode_fasta_sequence(fname): Loading @@ -181,6 +184,10 @@ def encode_fasta_sequence(fname): ---------- fname: str Filename of fasta file. Returns ------- np.ndarray: Shape (N_sequences, 4, sequence_length, 1). """ dna_alphabet = IUPACUnambiguousDNAWithN() Loading @@ -188,12 +195,27 @@ def encode_fasta_sequence(fname): def encode_bio_sequence(fname, file_type="fasta", alphabet=None): # np.ndarray: Shape (N_sequences, 4, sequence_length, 1). """ Loads a sequence file and returns an array of one-hot sequences. Parameters ---------- fname: str Filename of fasta file. file_type: str The type of file encoding to process, e.g. fasta or fastq, this is passed to Biopython.SeqIO.parse. alphabet: Bio.Alphabet.Alphabet A Biopython Alphabet, which should have a letter class property. Returns ------- np.ndarray: Shape (N_sequences, N_letters, sequence_length, 1). """ if alphabet is None: alphabet = IUPACUnambiguousDNAWithN() # TODO: if None, then get from the filename sequences = SeqIO.parse(fname, file_type, alphabet) return seq_one_hot_encode(sequences, alphabet.letters) Loading deepchem/utils/test/data/example.fasta +2 −2 Original line number Diff line number Diff line >seq0 AC XY >seq1 GA ZX deepchem/utils/test/data/example.fastq 0 → 100644 +8 −0 Original line number Diff line number Diff line @seq0 XY + hh @seq1 ZX + hh deepchem/utils/test/test_seq.py +33 −5 Original line number Diff line number Diff line Loading @@ -7,16 +7,28 @@ from __future__ import unicode_literals __author__ = "Bharath Ramsundar" __license__ = "MIT" import numpy as np import unittest import os import numpy as np from Bio.Alphabet import Alphabet import deepchem as dc class MockAlphabet(Alphabet): letters = "XYZ" class TestSeq(unittest.TestCase): """ Tests sequence handling utilities. """ def setUp(self): super(TestSeq, self).setUp() self.current_dir = os.path.dirname(os.path.abspath(__file__)) def test_one_hot_simple(self): sequences = np.array(["ACGT", "GATA", "CGCG"]) sequences = dc.utils.save.seq_one_hot_encode(sequences) Loading @@ -31,13 +43,29 @@ class TestSeq(unittest.TestCase): sequences = dc.utils.save.seq_one_hot_encode(sequences) def test_encode_fasta_sequence(self): fname = "./data/example.fasta" # Test it's possible to load a sequence with an aribrary alphabet from a fasta file. fname = os.path.join(self.current_dir, "./data/example.fasta") encoded_seqs = dc.utils.save.encode_bio_sequence( fname, alphabet=MockAlphabet()) expected = np.expand_dims( np.array([ [[1, 0], [0, 1], [0, 0]], [[0, 1], [0, 0], [1, 0]], ]), -1) np.testing.assert_array_equal(expected, encoded_seqs) def test_encode_fastq_sequence(self): fname = os.path.join(self.current_dir, "./data/example.fastq") encoded_seqs = dc.utils.save.encode_bio_sequence( fname, file_type="fastq", alphabet=MockAlphabet()) encoded_seqs = dc.utils.save.encode_fasta_sequence(fname) expected = np.expand_dims( np.array([ [[0, 0], [1, 0], [0, 0], [0, 1], [0, 0]], [[1, 0], [0, 1], [0, 0], [0, 0], [0, 0]], [[1, 0], [0, 1], [0, 0]], [[0, 1], [0, 0], [1, 0]], ]), -1) np.testing.assert_array_equal(expected, encoded_seqs) Loading
deepchem/data/tests/test_fasta_loader.py +5 −3 Original line number Diff line number Diff line Loading @@ -9,6 +9,7 @@ __license__ = "MIT" import os import unittest import deepchem as dc Loading @@ -26,7 +27,8 @@ class TestFASTALoader(unittest.TestCase): "../../data/tests/example.fasta") loader = dc.data.FASTALoader() sequences = loader.featurize(input_file) # example.fasta contains 3 sequences each of length 58. # The one-hot encoding turns base-pairs into vectors of length 4. # There is one "image channel") assert sequences.X.shape == (3, 4, 58, 1) # The one-hot encoding turns base-pairs into vectors of length 5 (ATCGN). # There is one "image channel". assert sequences.X.shape == (3, 5, 58, 1)
deepchem/utils/save.py +29 −7 Original line number Diff line number Diff line Loading @@ -110,14 +110,15 @@ def load_csv_files(filenames, shard_size=None, verbose=True): def seq_one_hot_encode(sequences, letters='ATCGN'): """One hot encodes list of genomic sequences. Sequences encoded have shape (N_sequences, 4, sequence_length, 1). Here 4 is for the 4 basepairs (ACGT) present in genomic sequences. Sequences encoded have shape (N_sequences, N_letters, sequence_length, 1). These sequences will be processed as images with one color channel. Parameters ---------- sequences: np.ndarray Array of genetic sequences letters: str String with the set of possible letters in the sequences. Raises ------ Loading @@ -126,7 +127,7 @@ def seq_one_hot_encode(sequences, letters='ATCGN'): Returns ------- np.ndarray: Shape (N_sequences, 4, sequence_length, 1). np.ndarray: Shape (N_sequences, N_letters, sequence_length, 1). """ # The label encoder is given characters for ACGTN Loading Loading @@ -168,9 +169,11 @@ def _seq_to_encoded(seq, letter_encoder, alphabet_length, sequence_length): return b # This could just be ambiguous_dna_letters, but that would be much higher dim. class IUPACUnambiguousDNAWithN(Alphabet): letters = unambiguous_dna_letters + "N" """ A deepchem specific DNA Alphabet with the 4 main letters and N. """ letters = "ACGTN" def encode_fasta_sequence(fname): Loading @@ -181,6 +184,10 @@ def encode_fasta_sequence(fname): ---------- fname: str Filename of fasta file. Returns ------- np.ndarray: Shape (N_sequences, 4, sequence_length, 1). """ dna_alphabet = IUPACUnambiguousDNAWithN() Loading @@ -188,12 +195,27 @@ def encode_fasta_sequence(fname): def encode_bio_sequence(fname, file_type="fasta", alphabet=None): # np.ndarray: Shape (N_sequences, 4, sequence_length, 1). """ Loads a sequence file and returns an array of one-hot sequences. Parameters ---------- fname: str Filename of fasta file. file_type: str The type of file encoding to process, e.g. fasta or fastq, this is passed to Biopython.SeqIO.parse. alphabet: Bio.Alphabet.Alphabet A Biopython Alphabet, which should have a letter class property. Returns ------- np.ndarray: Shape (N_sequences, N_letters, sequence_length, 1). """ if alphabet is None: alphabet = IUPACUnambiguousDNAWithN() # TODO: if None, then get from the filename sequences = SeqIO.parse(fname, file_type, alphabet) return seq_one_hot_encode(sequences, alphabet.letters) Loading
deepchem/utils/test/data/example.fasta +2 −2 Original line number Diff line number Diff line >seq0 AC XY >seq1 GA ZX
deepchem/utils/test/data/example.fastq 0 → 100644 +8 −0 Original line number Diff line number Diff line @seq0 XY + hh @seq1 ZX + hh
deepchem/utils/test/test_seq.py +33 −5 Original line number Diff line number Diff line Loading @@ -7,16 +7,28 @@ from __future__ import unicode_literals __author__ = "Bharath Ramsundar" __license__ = "MIT" import numpy as np import unittest import os import numpy as np from Bio.Alphabet import Alphabet import deepchem as dc class MockAlphabet(Alphabet): letters = "XYZ" class TestSeq(unittest.TestCase): """ Tests sequence handling utilities. """ def setUp(self): super(TestSeq, self).setUp() self.current_dir = os.path.dirname(os.path.abspath(__file__)) def test_one_hot_simple(self): sequences = np.array(["ACGT", "GATA", "CGCG"]) sequences = dc.utils.save.seq_one_hot_encode(sequences) Loading @@ -31,13 +43,29 @@ class TestSeq(unittest.TestCase): sequences = dc.utils.save.seq_one_hot_encode(sequences) def test_encode_fasta_sequence(self): fname = "./data/example.fasta" # Test it's possible to load a sequence with an aribrary alphabet from a fasta file. fname = os.path.join(self.current_dir, "./data/example.fasta") encoded_seqs = dc.utils.save.encode_bio_sequence( fname, alphabet=MockAlphabet()) expected = np.expand_dims( np.array([ [[1, 0], [0, 1], [0, 0]], [[0, 1], [0, 0], [1, 0]], ]), -1) np.testing.assert_array_equal(expected, encoded_seqs) def test_encode_fastq_sequence(self): fname = os.path.join(self.current_dir, "./data/example.fastq") encoded_seqs = dc.utils.save.encode_bio_sequence( fname, file_type="fastq", alphabet=MockAlphabet()) encoded_seqs = dc.utils.save.encode_fasta_sequence(fname) expected = np.expand_dims( np.array([ [[0, 0], [1, 0], [0, 0], [0, 1], [0, 0]], [[1, 0], [0, 1], [0, 0], [0, 0], [0, 0]], [[1, 0], [0, 1], [0, 0]], [[0, 1], [0, 0], [1, 0]], ]), -1) np.testing.assert_array_equal(expected, encoded_seqs)
