def test_genbank_to_fasta(self):
"""Conversion of GenBank to FASTA."""
filename = "GenBank/NC_005816.gb"
old = SeqIO.read(filename, "gb")
with open(filename) as handle:
new = SeqIO.read(TogoWS.convert(handle, "genbank", "fasta"), "fasta")
self.assertEqual(str(old.seq), str(new.seq))
def test_databases(self):
"""Check supported databases"""
dbs = set(TogoWS._get_entry_dbs())
expected = set(
[
"nuccore",
"nucest",
"nucgss",
"nucleotide",
"protein",
"gene",
"homologene",
"snp",
"mesh",
"pubmed", # 'embl',
"uniprot",
"uniparc",
"uniref100",
"uniref90",
"uniref50",
"ddbj",
"dad",
"pdb",
"compound",
"drug",
"enzyme",
"genes",
"glycan",
"orthology",
"reaction",
"module",
"pathway",
]
)
self.assertTrue(dbs.issuperset(expected), "Missing DB: %s" % ", ".join(sorted(expected.difference(dbs))))
def test_ncbi_protein(self):
"""Check supported fields for NCBI protein database"""
fields = set(TogoWS._get_entry_fields("ncbi-protein"))
self.assertTrue(
fields.issuperset(
[
"entry_id",
"length",
"strand",
"moltype",
"linearity",
"division",
"date",
"definition",
"accession",
"accessions",
"version",
"versions",
"acc_version",
"gi",
"keywords",
"organism",
"common_name",
"taxonomy",
"comment",
"seq",
]
),
fields,
)
def test_pubmed_16381885_title(self):
"""Bio.TogoWS.entry("pubmed", "16381885", field="title")"""
handle = TogoWS.entry("pubmed", "16381885", field="title")
data = handle.read().strip()
handle.close()
self.assertEqual(data,
'From genomics to chemical genomics: new developments in KEGG.')
def test_genbank_gff3(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", format="gff")"""
# Note - Using manual URL with genbank instead of nucleotide works
handle = TogoWS.entry("nucleotide", "X52960", format="gff")
data = handle.read()
handle.close()
self.assertTrue(data.startswith("##gff-version 3\nX52960\tGenbank\t"), data)
def test_pubmed_16381885_and_19850725(self):
"""Bio.TogoWS.entry("pubmed", "16381885,19850725")"""
handle = TogoWS.entry("pubmed", "16381885,19850725")
records = list(Medline.parse(handle))
handle.close()
self.assertEqual(len(records), 2)
self.assertEqual(records[0]["TI"], "From genomics to chemical genomics: new developments in KEGG.")
self.assertEqual(
records[0]["AU"],
[
"Kanehisa M",
"Goto S",
"Hattori M",
"Aoki-Kinoshita KF",
"Itoh M",
"Kawashima S",
"Katayama T",
"Araki M",
"Hirakawa M",
],
)
self.assertEqual(
records[1]["TI"],
"DDBJ launches a new archive database with analytical tools " + "for next-generation sequence data.",
)
self.assertEqual(
records[1]["AU"],
["Kaminuma E", "Mashima J", "Kodama Y", "Gojobori T", "Ogasawara O", "Okubo K", "Takagi T", "Nakamura Y"],
)
def test_ddbj_genbank(self):
"""Bio.TogoWS.entry("ddbj", "X52960")"""
handle = TogoWS.entry("ddbj", "X52960") # Returns "genbank" format
record = SeqIO.read(handle, "gb")
handle.close()
self.assertEqual(record.id, "X52960.1")
self.assertEqual(record.name, "X52960")
self.assertEqual(len(record), 248)
self.assertEqual(seguid(record.seq), "Ktxz0HgMlhQmrKTuZpOxPZJ6zGU")
def test_ddbj_fasta(self):
"""Bio.TogoWS.entry("ddbj", "X52960", "fasta")"""
handle = TogoWS.entry("ddbj", "X52960", "fasta")
record = SeqIO.read(handle, "fasta")
handle.close()
self.assertTrue("X52960" in record.id, record.id)
self.assertTrue("X52960" in record.name, record.name)
self.assertEqual(len(record), 248)
self.assertEqual(seguid(record.seq), "Ktxz0HgMlhQmrKTuZpOxPZJ6zGU")
def test_nucleotide_fasta(self):
"""Bio.TogoWS.entry("nucleotide", "6273291", "fasta")"""
handle = TogoWS.entry("nucleotide", "6273291", "fasta")
record = SeqIO.read(handle, "fasta")
handle.close()
self.assertTrue("6273291" in record.id, record.id)
self.assertTrue("6273291" in record.name, record.name)
self.assertEqual(len(record), 902)
self.assertEqual(seguid(record.seq), "bLhlq4mEFJOoS9PieOx4nhGnjAQ")
def check(self, database, search_term, expected_matches=(), limit=None):
if expected_matches and limit:
raise ValueError("Bad test - TogoWS makes no promises about order")
search_count = TogoWS.search_count(database, search_term)
if expected_matches and search_count < len(expected_matches):
raise ValueError("Only %i matches, expected at least %i" % (search_count, len(expected_matches)))
if search_count > 5000 and not limit:
print("%i results, skipping" % search_count)
return
if limit:
count = min(search_count, limit)
else:
count = search_count
# Iteration should find everything... unless a limit is used
search_iter = list(TogoWS.search_iter(database, search_term, limit))
self.assertEqual(count, len(search_iter))
for match in expected_matches:
self.assertTrue(match in search_iter, "Expected %s in results but not" % match)
def test_embl_AM905444(self):
"""Bio.TogoWS.entry("embl", "AM905444")"""
handle = TogoWS.entry("embl", "AM905444")
record = SeqIO.read(handle, "embl")
handle.close()
self.assertTrue("AM905444" in record.id, record.id)
self.assertTrue("AM905444" in record.name, record.name)
self.assertTrue("porin" in record.description, record.description)
self.assertEqual(len(record), 1164)
self.assertEqual(seguid(record.seq), "G0HtLpwF7i4FXUaUjDUPTjok79c")
def test_pubmed_16381885_au(self):
"""Bio.TogoWS.entry("pubmed", "16381885", field="au")"""
# Gives one name per line (i.e. \n separated), no dots
handle = TogoWS.entry("pubmed", "16381885", field="au")
data = handle.read().strip().split("\n")
handle.close()
self.assertEqual(data, ['Kanehisa M', 'Goto S', 'Hattori M',
'Aoki-Kinoshita KF', 'Itoh M',
'Kawashima S', 'Katayama T', 'Araki M',
'Hirakawa M'])
def test_pubmed_16381885_authors(self):
"""Bio.TogoWS.entry("pubmed", "16381885", field="authors")"""
# Gives names tab separated (i.e. \t separated)
handle = TogoWS.entry("pubmed", "16381885", field="authors")
data = handle.read().strip().split("\t")
handle.close()
self.assertEqual(data, ['Kanehisa, M.', 'Goto, S.', 'Hattori, M.',
'Aoki-Kinoshita, K. F.', 'Itoh, M.',
'Kawashima, S.', 'Katayama, T.', 'Araki, M.',
'Hirakawa, M.'])
def test_protein_fasta(self):
"""Bio.TogoWS.entry("protein", "16130152", "fasta")"""
handle = TogoWS.entry("protein", "16130152", "fasta")
record = SeqIO.read(handle, "fasta")
handle.close()
# Could use assertIn but requires Python 2.7+
self.assertTrue("16130152" in record.id, record.id)
self.assertTrue("16130152" in record.name, record.name)
self.assertTrue("porin protein" in record.description, record.description)
self.assertEqual(len(record), 367)
self.assertEqual(seguid(record.seq), "fCjcjMFeGIrilHAn6h+yju267lg")
def test_ddbj(self):
"""Check supported fields for ddbj database"""
fields = set(TogoWS._get_entry_fields("ddbj"))
self.assertTrue(fields.issuperset(['entry_id', 'length', 'strand',
'moltype', 'linearity', 'division',
'date', 'definition', 'accession',
'accessions', 'version', 'versions',
'acc_version', 'gi', 'keywords',
'organism', 'common_name',
'taxonomy', 'comment', 'seq']),
fields)
def test_nucleotide_fasta(self):
"""Bio.TogoWS.entry("nucleotide", "6273291", "fasta")"""
handle = TogoWS.entry("nucleotide", "6273291", "fasta")
record = SeqIO.read(handle, "fasta")
handle.close()
# NCBI is phasing out GI numbers, so no longer true:
# self.assertIn("6273291", record.id)
# self.assertIn("6273291", record.name)
self.assertIn("AF191665.1", record.id)
self.assertIn("AF191665.1", record.name)
self.assertEqual(len(record), 902)
self.assertEqual(seguid(record.seq), "bLhlq4mEFJOoS9PieOx4nhGnjAQ")
def test_pubmed_16381885(self):
"""Bio.TogoWS.entry("pubmed", "16381885")"""
# Gives Medline plain text
handle = TogoWS.entry("pubmed", "16381885")
data = Medline.read(handle)
handle.close()
self.assertEqual(data["TI"],
'From genomics to chemical genomics: new developments in KEGG.')
self.assertEqual(data["AU"], ['Kanehisa M', 'Goto S', 'Hattori M',
'Aoki-Kinoshita KF', 'Itoh M',
'Kawashima S', 'Katayama T', 'Araki M',
'Hirakawa M'])
def test_databases(self):
"""Check supported databases"""
dbs = set(TogoWS._get_entry_dbs())
self.assert_(dbs.issuperset(['nuccore', 'nucest', 'nucgss',
'nucleotide', 'protein', 'gene',
'omim', 'homologene', 'snp',
'mesh', 'pubmed', 'embl',
'uniprot', 'uniparc', 'uniref100',
'uniref90', 'uniref50', 'ddbj',
'dad', 'pdb', 'compound', 'drug',
'enzyme', 'genes', 'glycan',
'orthology', 'reaction', 'module',
'pathway']), dbs)
def test_protein_fasta(self):
"""Bio.TogoWS.entry("protein", "16130152", "fasta")"""
handle = TogoWS.entry("protein", "16130152", "fasta")
record = SeqIO.read(handle, "fasta")
handle.close()
# NCBI is phasing out GI numbers, so no longer true:
# self.assertIn("16130152", record.id)
# self.assertIn("16130152", record.name)
self.assertIn("NP_416719.1", record.id)
self.assertIn("NP_416719.1", record.name)
self.assertIn("porin protein", record.description)
self.assertEqual(len(record), 367)
self.assertEqual(seguid(record.seq), "fCjcjMFeGIrilHAn6h+yju267lg")
def test_databases(self):
"""Check supported databases"""
dbs = set(TogoWS._get_entry_dbs())
expected = set(['nuccore', 'nucest', 'nucgss',
'nucleotide', 'protein', 'gene',
'homologene', 'snp',
'mesh', 'pubmed', # 'embl',
'uniprot', 'uniparc', 'uniref100',
'uniref90', 'uniref50', 'ddbj',
'dad', 'pdb', 'compound', 'drug',
'enzyme', 'genes', 'glycan',
'orthology', 'reaction', 'module',
'pathway'])
self.assertTrue(dbs.issuperset(expected),
"Missing DB: %s" % ", ".join(sorted(expected.difference(dbs))))
def test_uniprot_swiss(self):
"""Bio.TogoWS.entry("uniprot", ["A1AG1_HUMAN","A1AG1_MOUSE"])"""
# Returns "swiss" format:
handle = TogoWS.entry("uniprot", ["A1AG1_HUMAN", "A1AG1_MOUSE"])
record1, record2 = SeqIO.parse(handle, "swiss")
handle.close()
self.assertEqual(record1.id, "P02763")
self.assertEqual(record1.name, "A1AG1_HUMAN")
self.assertEqual(len(record1), 201)
self.assertEqual(seguid(record1.seq), "LHDJJ6oC7gUXo8CC7Xn6EUeA8Gk")
self.assertEqual(record2.id, "Q60590")
self.assertEqual(record2.name, "A1AG1_MOUSE")
self.assertEqual(len(record2), 207)
self.assertEqual(seguid(record2.seq), "FGcj+RFQhP2gRusCmwPFty5PJT0")
def test_pubmed_16381885_and_19850725_authors(self):
"""Bio.TogoWS.entry("pubmed", "16381885,19850725", field="authors")"""
handle = TogoWS.entry("pubmed", "16381885,19850725", field="authors")
# Little hack to remove blank lines...
# names = handle.read().replace("\n\n", "\n").strip().split("\n")
names = handle.read().strip().split("\n")
handle.close()
self.assertEqual(2, len(names))
names1, names2 = names
self.assertEqual(names1.split("\t"),
['Kanehisa, M.', 'Goto, S.', 'Hattori, M.',
'Aoki-Kinoshita, K. F.', 'Itoh, M.',
'Kawashima, S.', 'Katayama, T.',
'Araki, M.', 'Hirakawa, M.'])
self.assertEqual(names2.split("\t"),
['Kaminuma, E.', 'Mashima, J.', 'Kodama, Y.',
'Gojobori, T.', 'Ogasawara, O.', 'Okubo, K.',
'Takagi, T.', 'Nakamura, Y.'])
def test_pubmed_16381885_and_19850725(self):
"""Bio.TogoWS.entry("pubmed", "16381885,19850725")"""
handle = TogoWS.entry("pubmed", "16381885,19850725")
records = list(Medline.parse(handle))
handle.close()
self.assertEqual(len(records), 2)
self.assertEqual(records[0]["TI"],
'From genomics to chemical genomics: new developments in KEGG.')
self.assertEqual(records[0]["AU"], ['Kanehisa M', 'Goto S',
'Hattori M', 'Aoki-Kinoshita KF',
'Itoh M', 'Kawashima S',
'Katayama T', 'Araki M',
'Hirakawa M'])
self.assertEqual(records[1]["TI"],
'DDBJ launches a new archive database with analytical tools ' +
'for next-generation sequence data.')
self.assertEqual(records[1]["AU"], ['Kaminuma E', 'Mashima J',
'Kodama Y', 'Gojobori T',
'Ogasawara O', 'Okubo K',
'Takagi T', 'Nakamura Y'])
def test_pubmed_16381885(self):
"""Bio.TogoWS.entry("pubmed", "16381885")"""
# Gives Medline plain text
handle = TogoWS.entry("pubmed", "16381885")
data = Medline.read(handle)
handle.close()
self.assertEqual(data["TI"], "From genomics to chemical genomics: new developments in KEGG.")
self.assertEqual(
data["AU"],
[
"Kanehisa M",
"Goto S",
"Hattori M",
"Aoki-Kinoshita KF",
"Itoh M",
"Kawashima S",
"Katayama T",
"Araki M",
"Hirakawa M",
],
)
def test_databases(self):
"""Check supported databases"""
dbs = set(TogoWS._get_entry_dbs())
self.assertTrue(
dbs.issuperset(
[
"nuccore",
"nucest",
"nucgss",
"nucleotide",
"protein",
"gene",
"omim",
"homologene",
"snp",
"mesh",
"pubmed",
"embl",
"uniprot",
"uniparc",
"uniref100",
"uniref90",
"uniref50",
"ddbj",
"dad",
"pdb",
"compound",
"drug",
"enzyme",
"genes",
"glycan",
"orthology",
"reaction",
"module",
"pathway",
]
),
dbs,
)
def test_genbank_to_embl(self):
"""Conversion of GenBank to EMBL."""
filename = "GenBank/NC_005816.gb"
old = SeqIO.read(filename, "gb")
new = SeqIO.read(TogoWS.convert(open(filename), "genbank", "embl"), "embl")
self.assertEqual(str(old.seq), str(new.seq))
def test_pdb(self):
"""Check supported fields for pdb database."""
fields = set(TogoWS._get_entry_fields("pdb"))
self.assertTrue(
fields.issuperset(["accession", "chains", "keywords", "models"]),
fields)
def test_uniprot(self):
"""Check supported fields for uniprot database."""
fields = set(TogoWS._get_entry_fields("uniprot"))
self.assertTrue(fields.issuperset(["definition", "entry_id", "seq"]),
fields)
def test_nucleotide_genbank_length(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", field="length")"""
handle = TogoWS.entry("nucleotide", "X52960", field="length")
data = handle.read().strip() # ignore trailing \n
handle.close()
self.assertEqual(data, "248")
def test_embl_AM905444_seq(self):
"""Bio.TogoWS.entry("embl", "AM905444", field="seq")"""
handle = TogoWS.entry("embl", "AM905444", field="seq")
data = handle.read().strip() # ignore any trailing \n
handle.close()
self.assertEqual(seguid(data), "G0HtLpwF7i4FXUaUjDUPTjok79c")
def test_embl_AM905444_definition(self):
"""Bio.TogoWS.entry("embl", "AM905444", field="definition")"""
handle = TogoWS.entry("embl", "AM905444", field="definition")
data = handle.read().strip() #ignore any trailing \n
handle.close()
self.assertEqual(data, "Herbaspirillum seropedicae locus tag HS193.0074 for porin")
def test_nucleotide_genbank_seq(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", field="seq")"""
handle = TogoWS.entry("nucleotide", "X52960", field="seq")
data = handle.read().strip() # ignore trailing \n
handle.close()
self.assertEqual(seguid(data), "Ktxz0HgMlhQmrKTuZpOxPZJ6zGU")
def test_embl(self):
"""Check supported fields for embl database"""
fields = set(TogoWS._get_entry_fields("embl"))
self.assertTrue(fields.issuperset(["definition", "entry_id", "seq"]),
fields)
def test_nucleotide_genbank_definition(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", field="definition")"""
handle = TogoWS.entry("nucleotide", "X52960", field="definition")
data = handle.read().strip() # ignore trailing \n
handle.close()
self.assertEqual(data, "Coleus blumei viroid 1 (CbVd) RNA.")
def test_nucleotide_genbank_acc_version(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", field="acc_version")"""
handle = TogoWS.entry("nucleotide", "X52960", field="acc_version")
data = handle.read().strip() # ignore trailing \n
handle.close()
self.assertEqual(data, "X52960.1")
def test_nucleotide_genbank_organism(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", field="organism")"""
handle = TogoWS.entry("nucleotide", "X52960", field="organism")
data = handle.read().strip() # ignore trailing \n
handle.close()
self.assertEqual(data, "Coleus blumei viroid 1")
def test_uniprot(self):
"""Check supported fields for uniprot database"""
fields = set(TogoWS._get_entry_fields("uniprot"))
self.assertTrue(fields.issuperset(["definition", "entry_id", "seq"]),
fields)
def test_pubmed(self):
"""Check supported fields for pubmed database"""
fields = set(TogoWS._get_entry_fields("pubmed"))
self.assert_(fields.issuperset(['abstract', 'au', 'authors',
'doi', 'mesh', 'so', 'ti',
'title']), fields)
def test_pubmed(self):
"""Check supported fields for pubmed database"""
fields = set(TogoWS._get_entry_fields("pubmed"))
self.assertTrue(fields.issuperset(['abstract', 'au', 'authors',
'doi', 'mesh', 'so',
'title']), fields)
def test_nucleotide_genbank_organism(self):
"""Bio.TogoWS.entry("nucleotide", "X52960", field="organism")"""
handle = TogoWS.entry("nucleotide", "X52960", field="organism")
data = handle.read().strip() # ignore trailing \n
handle.close()
self.assertEqual(data, "Coleus blumei viroid 1")
def test_ddbj_gff3(self):
"""Bio.TogoWS.entry("ddbj", "X52960", format="gff")"""
handle = TogoWS.entry("ddbj", "X52960", format="gff")
data = handle.read()
handle.close()
self.assertTrue(data.startswith("##gff-version 3\nX52960\tDDBJ\t"), data)
def test_embl_AM905444_gff3(self):
"""Bio.TogoWS.entry("embl", "AM905444", format="gff")"""
handle = TogoWS.entry("embl", "AM905444", format="gff")
data = handle.read()
handle.close()
self.assert_(data.startswith("##gff-version 3\nAM905444\tembl\t"), data)
def test_pdb(self):
"""Check supported fields for pdb database"""
fields = set(TogoWS._get_entry_fields("pdb"))
self.assertTrue(fields.issuperset(["accession", "chains", "keywords",
"models"]), fields)
def test_ddbj_gff3(self):
"""Bio.TogoWS.entry("ddbj", "X52960", format="gff")"""
handle = TogoWS.entry("ddbj", "X52960", format="gff")
data = handle.read()
handle.close()
self.assertTrue(data.startswith("##gff-version 3\nX52960\tDDBJ\t"), data)
