def test_nisin_complete(self):
with TemporaryDirectory() as output_dir:
args = [
"--minimal", "--enable-tta", "--tta-threshold", "0",
"--output-dir", output_dir,
helpers.get_path_to_nisin_genbank()
]
options = build_config(args,
isolated=True,
modules=antismash.get_all_modules())
antismash.run_antismash(helpers.get_path_to_nisin_genbank(),
options)
# regen the results
update_config(
{"reuse_results": os.path.join(output_dir, "nisin.json")})
prior_results = read_data(None, options)
record = prior_results.records[0]
results = prior_results.results[0]
tta_results = tta.regenerate_previous_results(
results.get("antismash.modules.tta"), record, options)
assert isinstance(tta_results, tta.TTAResults)
assert len(tta_results.features) == 174
# raise the threshold above the gc_content and ensure regenned has no hits
update_config({"tta_threshold": 0.65})
tta_results = tta.regenerate_previous_results(
results.get("antismash.modules.tta"), record, options)
assert isinstance(tta_results, tta.TTAResults)
assert not tta_results.features
def test_trim(self):
nisin_path = helpers.get_path_to_nisin_genbank()
records = record_processing.parse_input_sequence(nisin_path, start=10, end=5000)
assert len(records) == 1
assert isinstance(records[0], Record)
assert len(records[0].get_cds_features()) == 2
assert len(records[0].seq) == 4990
def test_nisin_minimal(self):
# make sure the output directory isn't filled
out_dir = self.default_options.output_dir
assert not list(glob.glob(os.path.join(out_dir, "*")))
# die with neither inputs provided
with self.assertRaisesRegex(
ValueError, "No sequence file or prior results to read"):
run_antismash(None, self.default_options)
# make sure no files created
assert not list(glob.glob(os.path.join(out_dir, "*")))
# do a normal run
run_antismash(get_path_to_nisin_genbank(), self.default_options)
self.check_output_files()
# remove html file and make sure it's recreated
os.unlink(os.path.join(self.default_options.output_dir, "index.html"))
update_config({
"reuse_results":
os.path.join(self.default_options.output_dir, "nisin.json")
})
run_antismash(None, self.default_options)
self.check_output_files()
def test_add_to_record(self):
nisin = helpers.get_path_to_nisin_genbank()
record = record_processing.parse_input_sequence(nisin)[0]
assert not record.get_pfam_domains()
# add a test PFAM
pfam = PFAMDomain(FeatureLocation(2, 5),
description="test",
protein_start=5,
protein_end=10,
identifier="PF00005",
domain="PF00005",
tool="test")
pfam.domain_id = "test"
record.add_pfam_domain(pfam)
assert len(record.get_pfam_domains()) == 1
# run pfam2go and add the results
results = pfam2go.run_on_record(record, None, self.options)
assert pfam in results.pfam_domains_with_gos
assert not pfam.gene_ontologies
results.add_to_record(record)
assert pfam.gene_ontologies
# check the contents of the annotation
for domain in record.get_pfam_domains():
assert domain.gene_ontologies
assert sorted(domain.gene_ontologies.ids) == sorted(
results.get_all_gos(domain))
def test_missing_locations_caught(self):
rec = list(Bio.SeqIO.parse(get_path_to_nisin_genbank(), "genbank"))[0]
Record.from_biopython(rec, taxon="bacteria")
rec.features.append(SeqFeature(None, type="broken"))
with self.assertRaisesRegex(SecmetInvalidInputError,
"missing or invalid location"):
Record.from_biopython(rec, taxon="bacteria")
def test_record_conversion_from_biopython(self):
before = list(Bio.SeqIO.parse(helpers.get_path_to_nisin_genbank(), "genbank"))[0]
# sort notes, because direct comparisons otherwise are awful
for feature in before.features:
if "note" in feature.qualifiers:
feature.qualifiers["note"] = sorted(feature.qualifiers["note"])
before_features = sorted(map(str, before.features))
type_counts = defaultdict(lambda: 0)
for feature in before.features:
type_counts[feature.type] += 1
record = Record.from_biopython(before, taxon="bacteria")
after = record.to_biopython()
# ensure new features are correct
assert len(before_features) == len(after.features)
for bef, aft in zip(before_features, sorted(map(str, after.features))):
assert bef == aft
# ensure we haven't changed the original record or feature list
assert id(before) != id(after)
assert id(before.features) != id(after.features)
for i in range(len(before.features)):
assert id(before.features[i]) != id(after.features[i])
for bef, aft in zip(before_features, sorted(map(str, before.features))):
assert bef == aft
# ensure that the counts of each match
assert type_counts["CDS"] == len(record.get_cds_features())
assert type_counts["PFAM_domain"] == len(record.get_pfam_domains())
assert type_counts["cluster"] == len(record.get_clusters())
assert type_counts["aSDomain"] == len(record.get_antismash_domains())
def test_nisin(self):
nisin_path = helpers.get_path_to_nisin_genbank()
records = record_processing.parse_input_sequence(nisin_path)
assert len(records) == 1
assert isinstance(records[0], Record)
assert len(records[0].get_cds_features()) == 11
assert len(records[0].seq) == 15016
def test_record_to_json_and_back(self):
filename = get_path_to_nisin_genbank()
records = list(seqio.parse(open(filename), "genbank"))
records = [
Record.from_biopython(rec, taxon="bacteria") for rec in records
]
rec_results = [{}, {}, {}]
results = serialiser.AntismashResults(filename, records, rec_results,
"dummy")
json_handle = StringIO()
results.write_to_file(json_handle)
json_handle.seek(0)
new_results = serialiser.AntismashResults.from_file(json_handle,
taxon="bacteria")
assert results.to_json() == new_results.to_json()
# check no records were lost
assert len(new_results.records) == len(results.records)
# check that the contents of the records is the same
# by converting to biopython and writing to genbanks
original = self.create_data_stream(results.records)
new = self.create_data_stream(new_results.records)
oldvalue = original.getvalue()
newvalue = new.getvalue()
with TemporaryDirectory(change=True):
open("old.json", "w").write(oldvalue)
open("new.json", "w").write(newvalue)
for oldline, newline in zip(oldvalue.split('\n'),
newvalue.split('\n')):
assert oldline == newline
def test_reuse(self):
nisin = helpers.get_path_to_nisin_genbank()
record = record_processing.parse_input_sequence(nisin)[0]
results = helpers.run_and_regenerate_results_for_module(
nisin, cluster_hmmer, self.options)
json = results.to_json()
assert len(results.hits) == 24
self.check_add_to_record(nisin, results)
# test regeneration when thresholds are less restrictive
new_score_threshold = self.original_min_score - .1
self.set_min_score(new_score_threshold)
new_results = cluster_hmmer.regenerate_previous_results(
json, record, self.options)
assert new_results is None
self.set_min_score(self.original_min_score)
new_evalue_threshold = self.original_max_evalue + .1
self.set_max_evalue(new_evalue_threshold)
new_results = cluster_hmmer.regenerate_previous_results(
json, record, self.options)
assert new_results is None
self.set_max_evalue(self.original_max_evalue)
# test regeneration when evalue threshold is more restrictive
new_evalue_threshold = sorted(hit["evalue"]
for hit in results.hits)[12]
assert new_evalue_threshold < self.original_max_evalue
new_hits = []
for hit in results.hits:
if hit["evalue"] <= new_evalue_threshold:
new_hits.append(hit)
new_hits.sort(key=lambda x: x["evalue"])
assert len(new_hits) < 24
self.set_max_evalue(new_evalue_threshold)
new_results = cluster_hmmer.regenerate_previous_results(
json, record, self.options)
self.set_max_evalue(self.original_max_evalue)
assert sorted(new_results.hits, key=lambda x: x["evalue"]) == new_hits
self.check_add_to_record(nisin, results)
# test regeneration when score threshold is more restrictive
new_score_threshold = sorted(hit["score"] for hit in results.hits)[12]
assert new_score_threshold > cluster_hmmer.MIN_SCORE
new_hits = []
for hit in results.hits:
if hit["score"] >= new_score_threshold:
new_hits.append(hit)
new_hits.sort(key=lambda x: x["score"])
assert len(new_hits) < 24
self.set_min_score(new_score_threshold)
new_results = cluster_hmmer.regenerate_previous_results(
json, record, self.options)
self.set_min_score(self.original_min_score)
assert sorted(new_results.hits, key=lambda x: x["score"]) == new_hits
self.check_add_to_record(nisin, results)
def test_read_from_file(self):
# very basic testing to ensure that the file IO itself functions
recs = Record.from_genbank(get_path_to_nisin_genbank())
assert len(recs) == 1
rec = recs[0]
assert rec.get_feature_count() == 24
assert len(rec.get_cds_features()) == 11
assert isinstance(rec.get_cds_by_name("nisB"), CDSFeature)
def test_trees_complete(self):
with TemporaryDirectory() as output_dir:
args = [
"--minimal", "--enable-genefunctions", "--smcog-trees",
"--output-dir", output_dir,
helpers.get_path_to_nisin_genbank()
]
options = build_config(args,
isolated=True,
modules=antismash.get_all_modules())
antismash.run_antismash(helpers.get_path_to_nisin_genbank(),
options)
with open(os.path.join(output_dir, "nisin.json")) as res_file:
assert "antismash.modules.smcog_trees" in res_file.read()
tree_files = list(
glob.glob(os.path.join(output_dir, "smcogs", "*.png")))
assert len(tree_files) == 7
sample_tree = tree_files[0]
# regen the results
update_config(
{"reuse_results": os.path.join(output_dir, "nisin.json")})
prior_results = read_data(None, options)
record = prior_results.records[0]
results = prior_results.results[0]
tree_results = results["antismash.modules.smcog_trees"]
smcogs_results = smcog_trees.regenerate_previous_results(
tree_results, record, options)
assert len(smcogs_results.tree_images) == 7
assert os.path.exists(sample_tree)
os.unlink(sample_tree)
assert not os.path.exists(sample_tree)
# attempt to regen the results, the deleted tree image will prevent it
prior_results = read_data(None, options)
record = prior_results.records[0]
results = prior_results.results[0]
smcogs_results = smcog_trees.regenerate_previous_results(
tree_results, record, options)
assert smcogs_results is None
def test_protein_sequences_caught(self):
before = list(Bio.SeqIO.parse(get_path_to_nisin_genbank(), "genbank"))[0]
# as a sanity check, make sure it's a seq and it functions as expected
assert isinstance(before.seq, Seq)
Record.from_biopython(before, taxon="bacteria")
before.seq = Seq("AAAA", IUPACProtein())
with self.assertRaisesRegex(ValueError, "protein records are not supported"):
Record.from_biopython(before, taxon="bacteria")
def test_nisin_end_to_end(self):
# skip fimo being disabled for this, we already test the computational
# side elsewhere
if self.options.without_fimo:
return
nisin = helpers.get_path_to_nisin_genbank()
result = helpers.run_and_regenerate_results_for_module(
nisin, lanthipeptides, self.options)
assert list(result.motifs_by_locus) == ["nisB"]
prepeptide = result.motifs_by_locus["nisB"][0]
self.assertAlmostEqual(3336.0, prepeptide.molecular_weight, delta=0.05)
def test_dna_casing(self):
before = list(Bio.SeqIO.parse(get_path_to_nisin_genbank(),
"genbank"))[0]
for molecule in ["DNA", "dna", "Dna"]:
before.annotations["molecule_type"] = molecule
Record.from_biopython(before, taxon="bacteria")
before.annotations["molecule_type"] = molecule + "x"
with self.assertRaisesRegex(ValueError,
"records are not supported"):
Record.from_biopython(before, taxon="bacteria")
def test_nisin_complete(self):
with TemporaryDirectory() as output_dir:
args = [
"--minimal", "--tta", "--output-dir", output_dir,
helpers.get_path_to_nisin_genbank()
]
options = build_config(args,
isolated=True,
modules=antismash.get_all_modules())
antismash.run_antismash(helpers.get_path_to_nisin_genbank(),
options)
# regen the results
update_config(
{"reuse_results": os.path.join(output_dir, "nisin.json")})
prior_results = read_data(None, options)
record = prior_results.records[0]
results = prior_results.results[0]
regenned = regenerate_results_for_record(record, options, [tta],
results)
tta_results = regenned["antismash.modules.tta"]
assert isinstance(tta_results, tta.TTAResults)
assert len(tta_results.features) == 174
def test_subregions_annotated(self, _patched_rules):
record = Record.from_genbank(get_path_to_nisin_genbank())[0]
record.strip_antismash_annotations()
assert not record.get_regions()
assert not record.get_subregions()
results = hmm_detection.run_on_record(record, None, self.options)
assert not results.get_predicted_protoclusters()
for cds in ["nisB", "nisC"]:
assert not record.get_cds_by_name(cds).sec_met
cutoff = record.get_cds_by_name("nisB").location.end + 10
record.add_subregion(DummySubRegion(end=cutoff))
results = hmm_detection.run_on_record(record, None, self.options)
assert record.get_cds_by_name("nisB").sec_met
assert not record.get_cds_by_name("nisC").sec_met
# and then a json conversion, even without subregion added
record = Record.from_genbank(get_path_to_nisin_genbank())[0]
record.strip_antismash_annotations()
raw = json.loads(json.dumps(results.to_json()))
hmm_detection.regenerate_previous_results(raw, record, self.options)
assert record.get_cds_by_name("nisB").sec_met
assert not record.get_cds_by_name("nisC").sec_met
def test_nisin_complete(self):
with TemporaryDirectory() as output_dir:
args = [
"run_antismash.py", "--minimal", "--enable-lanthipeptides",
"--output-dir", output_dir
]
options = build_config(args,
isolated=True,
modules=antismash.get_all_modules())
antismash.run_antismash(helpers.get_path_to_nisin_genbank(),
options)
# make sure the html_output section was tested
with open(os.path.join(output_dir, "index.html")) as handle:
content = handle.read()
assert "nisA leader / core peptide" in content
def test_minimum_length(self):
nisin_path = helpers.get_path_to_nisin_genbank()
records = record_processing.parse_input_sequence(nisin_path,
minimum_length=-16)
assert len(records) == 1
records = record_processing.parse_input_sequence(nisin_path,
minimum_length=15016)
assert len(records) == 1
with self.assertRaisesRegex(AntismashInputError, "no valid records found"):
record_processing.parse_input_sequence(nisin_path, minimum_length=15017)
for bad_len in [5.6, None, "5"]:
with self.assertRaisesRegex(TypeError, "minimum_length must be an int"):
record_processing.parse_input_sequence(path, minimum_length=bad_len)
def test_reuse(self):
nisin = helpers.get_path_to_nisin_genbank()
results = helpers.run_and_regenerate_results_for_module(
nisin, pfam2go, self.options)
# are the expected go ids for pfams found/no wrong ids for pfams?
expected_pfams_and_gos_with_descs = {
"PF00005": {
"GO:0005524": "ATP binding",
"GO:0016887": "ATPase activity"
},
"PF00072": {
"GO:0000160": "phosphorelay signal transduction system"
},
"PF00486": {
"GO:0003677": "DNA binding",
"GO:0000160": "phosphorelay signal transduction system",
"GO:0006355": "regulation of transcription, DNA-templated"
}
}
expected_pfams_found = set()
for pfam, all_ontologies in results.pfam_domains_with_gos.items():
pfam_ids_without_versions = [
pfam_id.partition(".")[0] for pfam_id in pfam.db_xref
]
# make sure the Pfams without gos aren't in the results
assert "PF05147" not in pfam_ids_without_versions and "PF04738" not in pfam_ids_without_versions
for ontologies in all_ontologies:
# make sure GeneOntologies' pfam id actually is one found in the domain's ids
assert ontologies.pfam in pfam_ids_without_versions
# did it find the right amount of GO IDs for the sample Pfams, and did it find the right ones?
if ontologies.pfam in expected_pfams_and_gos_with_descs:
expected_pfams_found.add(ontologies.pfam)
go_ids = [
str(go_entry) for go_entry in ontologies.go_entries
]
assert len(go_ids) == len(
expected_pfams_and_gos_with_descs[ontologies.pfam])
for go_id in go_ids:
assert go_id in expected_pfams_and_gos_with_descs[
ontologies.pfam]
# make sure all expected pfams have been found
assert len(expected_pfams_found) == len(
expected_pfams_and_gos_with_descs)
self.check_add_to_record(nisin, results)
def test_nisin(self):
"Test lanthipeptide prediction for nisin A"
expected_html_snippet = "nisA leader / core peptide"
genbank = helpers.get_path_to_nisin_genbank()
rec, _ = self.run_lanthi(genbank, expected_html_snippet)
prepeptide = rec.get_cds_motifs()[0]
# real monoisotopic mass is 3351.51, but we overpredict a Dha
self.assertAlmostEqual(3333.6, prepeptide.monoisotopic_mass, delta=0.05)
# real mw is 3354.5, see above
self.assertAlmostEqual(3336.0, prepeptide.molecular_weight, delta=0.05)
for expected, calculated in zip([3354.0, 3372.1, 3390.1, 3408.1],
prepeptide.alternative_weights):
self.assertAlmostEqual(expected, calculated, delta=0.05)
assert prepeptide.detailed_information.lan_bridges == 5
self.assertEqual("MSTKDFNLDLVSVSKKDSGASPR", prepeptide.leader)
self.assertEqual("ITSISLCTPGCKTGALMGCNMKTATCHCSIHVSK", prepeptide.core)
self.assertEqual('Class I', prepeptide.peptide_subclass)
def test_reuse(self):
raw_results = sideloader.loader.load_validated_json(
GOOD_FILE, sideloader.general._SCHEMA_FILE)
nisin = helpers.get_path_to_nisin_genbank()
results = helpers.run_and_regenerate_results_for_module(
nisin, sideloader, self.options)
assert results.record_id == raw_results["records"][0]["name"]
record_section = raw_results["records"][0]
for result, raw in zip(results.subregions,
record_section["subregions"]):
assert result.tool.name == raw_results["tool"]["name"]
assert result.start == raw["start"]
assert result.end == raw["end"]
assert result.label == raw["label"]
assert result.details == {
"score": ["6.5"],
"some_option_name": ["yes"],
"some_other_detail": ["first", "second", "etc"]
}
def test_add_to_record(self):
nisin = helpers.get_path_to_nisin_genbank()
record = record_processing.parse_input_sequence(nisin)[0]
assert not record.get_pfam_domains()
# add a test PFAM
pfam = helpers.DummyPFAMDomain(identifier="PF00005", domain="PF00005")
record.add_pfam_domain(pfam)
assert len(record.get_pfam_domains()) == 1
# run pfam2go and add the results
results = pfam2go.run_on_record(record, None, self.options)
assert pfam in results.pfam_domains_with_gos
assert not pfam.gene_ontologies
results.add_to_record(record)
assert pfam.gene_ontologies
# check the contents of the annotation
for domain in record.get_pfam_domains():
assert domain.gene_ontologies
assert sorted(domain.gene_ontologies.ids) == sorted(
results.get_all_gos(domain))
def test_results_from_invalid_file_raises_error(self):
filename = get_path_to_nisin_genbank()
self.assertRaisesRegex(ValueError, "Cannot load results to reuse",
serialiser.AntismashResults.from_file, filename)
def read_nisin(self):
records = record_processing.parse_input_sequence(
helpers.get_path_to_nisin_genbank())
assert len(records) == 1
return records
def check_nisin(self, expected):
return self.run_antismash(helpers.get_path_to_nisin_genbank(),
expected)
def test_nisin_minimal(self):
run_antismash(get_path_to_nisin_genbank(), self.default_options)
self.check_output_files()
def test_nisin_minimal(self):
run_antismash(get_path_to_nisin_genbank(), self.config)
self.check_output_files()