def test_sequence_conversion():
path = os.path.join(data_dir, "nuc.fasta")
file = fasta.FastaFile()
file.read(path)
assert seq.NucleotideSequence("ACGCTACGT") == fasta.get_sequence(file)
seq_dict = fasta.get_sequences(file)
file2 = fasta.FastaFile()
fasta.set_sequences(file2, seq_dict)
seq_dict2 = fasta.get_sequences(file2)
assert seq_dict == seq_dict2
file3 = fasta.FastaFile()
fasta.set_sequence(file3, seq.NucleotideSequence("AACCTTGG"))
assert file3["sequence"] == "AACCTTGG"
path = os.path.join(data_dir, "prot.fasta")
file4 = fasta.FastaFile()
file4.read(path)
assert seq.ProteinSequence("YAHGFRTGS") == fasta.get_sequence(file4)
path = os.path.join(data_dir, "invalid.fasta")
file5 = fasta.FastaFile()
file5.read(path)
with pytest.raises(ValueError):
seq.NucleotideSequence(fasta.get_sequence(file5))
def test_sequence_conversion():
path = os.path.join(data_dir("sequence"), "nuc.fasta")
file = fasta.FastaFile.read(path)
assert seq.NucleotideSequence("ACGCTACGT") == fasta.get_sequence(file)
seq_dict = fasta.get_sequences(file)
file2 = fasta.FastaFile()
fasta.set_sequences(file2, seq_dict)
seq_dict2 = fasta.get_sequences(file2)
# Cannot compare dicts directly, since the original RNA sequence is
# now guessed as protein sequence
for seq1, seq2 in zip(seq_dict.values(), seq_dict2.values()):
assert str(seq1) == str(seq2)
file3 = fasta.FastaFile()
fasta.set_sequence(file3, seq.NucleotideSequence("AACCTTGG"))
assert file3["sequence"] == "AACCTTGG"
path = os.path.join(data_dir("sequence"), "prot.fasta")
file4 = fasta.FastaFile.read(path)
# Expect a warning for selenocysteine conversion
with pytest.warns(UserWarning):
assert seq.ProteinSequence("YAHCGFRTGS") == fasta.get_sequence(file4)
path = os.path.join(data_dir("sequence"), "invalid.fasta")
file5 = fasta.FastaFile.read(path)
with pytest.raises(ValueError):
seq.NucleotideSequence(fasta.get_sequence(file5))
def test_fetch_single_file(as_file_like):
file_name = None if as_file_like else biotite.temp_file("fa")
file = entrez.fetch_single_file(["1L2Y_A", "3O5R_A"], file_name, "protein",
"fasta")
fasta_file = fasta.FastaFile.read(file)
prot_seqs = fasta.get_sequences(fasta_file)
assert len(prot_seqs) == 2
def test_fetch_single_file():
file = entrez.fetch_single_file(["1L2Y_A", "3O5R_A"],
biotite.temp_file("fa"), "protein",
"fasta")
fasta_file = fasta.FastaFile()
fasta_file.read(file)
prot_seqs = fasta.get_sequences(fasta_file)
assert len(prot_seqs) == 2
def test_access_high_level():
path = os.path.join(data_dir("sequence"), "nuc.fasta")
file = fasta.FastaFile.read(path)
sequences = fasta.get_sequences(file)
assert sequences == {
"dna sequence": seq.NucleotideSequence("ACGCTACGT", False),
"another dna sequence": seq.NucleotideSequence("A", False),
"third dna sequence": seq.NucleotideSequence("ACGT", False),
"rna sequence": seq.NucleotideSequence("ACGT", False),
"ambiguous rna sequence": seq.NucleotideSequence("ACGTNN", True),
}
def test_fetch(format, as_file_like):
path = None if as_file_like else biotite.temp_dir()
file_path_or_obj = rcsb.fetch("1l2y", format, path, overwrite=True)
if format == "pdb":
file = pdb.PDBFile.read(file_path_or_obj)
pdb.get_structure(file)
elif format == "pdbx":
file = pdbx.PDBxFile.read(file_path_or_obj)
pdbx.get_structure(file)
elif format == "mmtf":
file = mmtf.MMTFFile.read(file_path_or_obj)
mmtf.get_structure(file)
elif format == "fasta":
file = fasta.FastaFile.read(file_path_or_obj)
# Test if the file contains any sequences
assert len(fasta.get_sequences(file)) > 0
def test_fetch_single_file(as_file_like):
if as_file_like:
file_name = None
else:
file = tempfile.NamedTemporaryFile("r", suffix=".fa")
file_name = file.name
downloaded_file_name = entrez.fetch_single_file(
["1L2Y_A", "3O5R_A"], file_name, "protein", "fasta"
)
fasta_file = fasta.FastaFile.read(downloaded_file_name)
prot_seqs = fasta.get_sequences(fasta_file)
assert len(prot_seqs) == 2
if not as_file_like:
file.close()
def main():
parser = argparse.ArgumentParser(
description='Score sequences based on a given structure.'
)
parser.add_argument(
'pdbfile', type=str,
help='input filepath, either .pdb or .cif',
)
parser.add_argument(
'seqfile', type=str,
help='input filepath for variant sequences in a .fasta file',
)
parser.add_argument(
'--outpath', type=str,
help='output filepath for scores of variant sequences',
default='output/sequence_scores.csv',
)
parser.add_argument(
'--chain', type=str,
help='chain id for the chain of interest', default='A',
)
args = parser.parse_args()
model, alphabet = esm.pretrained.esm_if1_gvp4_t16_142M_UR50()
coords, seq = esm.inverse_folding.util.load_coords(args.pdbfile, args.chain)
print('Native sequence loaded from structure file:')
print(seq)
print('\n')
ll, _ = esm.inverse_folding.util.score_sequence(
model, alphabet, coords, seq)
print('Native sequence')
print(f'Log likelihood: {ll:.2f}')
print(f'Perplexity: {np.exp(-ll):.2f}')
print('\nScoring variant sequences from sequence file..\n')
infile = FastaFile()
infile.read(args.seqfile)
seqs = get_sequences(infile)
Path(args.outpath).parent.mkdir(parents=True, exist_ok=True)
with open(args.outpath, 'w') as fout:
fout.write('seqid,log_likelihood\n')
for header, seq in tqdm(seqs.items()):
ll, _ = esm.inverse_folding.util.score_sequence(
model, alphabet, coords, str(seq))
fout.write(header + ',' + str(ll) + '\n')
print(f'Results saved to {args.outpath}')
spacing=spacing)
twin = axes.get_shared_x_axes().get_siblings(axes)[0]
for ax in (axes, twin):
ax.set_yticklabels(ax.get_yticklabels(), fontdict={"color": "white"})
axes.get_figure().patch.set_facecolor("#181818")
# Using cyclotide sequences as example
query = (entrez.SimpleQuery("Cyclotide") & entrez.SimpleQuery("cter")
& entrez.SimpleQuery("srcdb_swiss-prot", field="Properties")
^ entrez.SimpleQuery("Precursor"))
uids = entrez.search(query, "protein")
fasta_file = fasta.FastaFile.read(
entrez.fetch_single_file(uids, None, "protein", "fasta"))
sequence_dict = fasta.get_sequences(fasta_file)
headers = list(sequence_dict.keys())
sequences = list(sequence_dict.values())
labels = [header[-1] for header in headers]
# Perform a multiple sequence alignment
matrix = align.SubstitutionMatrix.std_protein_matrix()
alignment, order, _, _ = align.align_multiple(sequences, matrix)
# Order alignment according to guide tree
alignment = alignment[:, order.tolist()]
labels = [labels[i] for i in order]
# Visualize the alignment using the new alignment plotter
fig = plt.figure(figsize=(8.0, 3.7))
ax = fig.add_subplot(111)
plot_alignment_shapes(ax,
