def test_embl_cds_interaction(self):
"""Test EMBL CDS interaction, parse CDS features on embl files."""
embl_s = Scanner.EmblScanner()
# Test parse CDS features on embl_file
with open("EMBL/AE017046.embl") as handle_embl7046:
l_cds_f = list(embl_s.parse_cds_features(handle_embl7046))
# number of records, should be 10
self.assertEqual(len(l_cds_f), 10)
# Seq ID
self.assertEqual(l_cds_f[0].id, 'AAS58758.1')
self.assertEqual(l_cds_f[0].description, 'putative transposase')
def test_embl_record_interaction(self):
"""Test EMBL Record interaction on embl files."""
embl_s = Scanner.EmblScanner()
# Test parse records on embl_file
with open("EMBL/AE017046.embl") as handle_embl7046:
l_embl_r = list(embl_s.parse_records(handle_embl7046, do_features=True))
# number of records, should be 1
self.assertEqual(len(l_embl_r), 1)
self.assertEqual(l_embl_r[0].id, 'AE017046.1')
self.assertEqual(l_embl_r[0].description, 'Yersinia pestis biovar Microtus '
'str. 91001 plasmid pPCP1, complete '
'sequence.')
self.assertEqual(len(l_embl_r[0].features), 29)
def test_topology_embl(self):
"""Check EMBL ID line parsing."""
# This is a bit low level, but can test pasing the ID line only
tests = [
# Modern examples with sequence version
("ID X56734; SV 1; linear; mRNA; STD; PLN; 1859 BP.", "linear",
"mRNA", "PLN"),
("ID CD789012; SV 4; linear; genomic DNA; HTG; MAM; 500 BP.",
"linear", "genomic DNA", "MAM"),
# Example to match GenBank example used above:
("ID U49845; SV 1; linear; genomic DNA; STD; FUN; 5028 BP.",
"linear", "genomic DNA", "FUN"),
# Old examples:
("ID BSUB9999 standard; circular DNA; PRO; 4214630 BP.",
"circular", "DNA", "PRO"),
("ID SC10H5 standard; DNA; PRO; 4870 BP.", None, "DNA", "PRO"),
# Patent example from 2016-06-10
# ftp://ftp.ebi.ac.uk/pub/databases/embl/patent/
("ID A01679; SV 1; linear; unassigned DNA; PAT; MUS; 12 BP.",
"linear", "unassigned DNA", "MUS"),
# Old patent examples
("ID NRP_AX000635; PRT; NR1; 15 SQ", None, None, "NR1"),
("ID NRP0000016E; PRT; NR2; 5 SQ", None, None, "NR2"),
# KIPO patent examples
("ID DI500001 STANDARD; PRT; 111 AA.", None, None,
None),
("ID DI644510 standard; PRT; 1852 AA.", None, None, None),
]
for (line, topo, mol_type, div) in tests:
scanner = Scanner.EmblScanner()
consumer = GenBank._FeatureConsumer(1, GenBank.FeatureValueCleaner)
scanner._feed_first_line(consumer, line)
t = consumer.data.annotations.get('topology', None)
self.assertEqual(
t, topo, "Wrong topology %r not %r from %r" % (t, topo, line))
mt = consumer.data.annotations.get('molecule_type', None)
self.assertEqual(
mt, mol_type,
"Wrong molecule_type %r not %r from %r" % (mt, mol_type, line))
d = consumer.data.annotations.get('data_file_division', None)
self.assertEqual(
d, div, "Wrong division %r not %r from %r" % (d, div, line))