def test_semantic_similarity(usr_assc=None):
"""Computing basic semantic similarities between GO terms."""
not_these = {'goa_uniprot_all.gaf', 'goa_uniprot_all_noiea.gaf'}
assc_names = sorted(ASSOCIATIONS.difference(not_these))
go2obj = get_go2obj()
# http://current.geneontology.org/annotations/
if usr_assc is not None:
assc_names = [usr_assc]
not_found = set()
gaf2errs = cx.defaultdict(list)
for assc_name in assc_names: # Limit test numbers for speed
tic = timeit.default_timer()
# Get all the annotations from arabidopsis.
fin_gaf = os.path.join(REPO, assc_name)
if not os.path.exists(fin_gaf):
dnld_annotation(fin_gaf)
annoobj = GafReader(fin_gaf)
for nta in annoobj.associations:
if nta.GO_ID in go2obj:
goterm = go2obj[nta.GO_ID]
namespace_anno = NS2NAMESPACE.get(nta.NS)
if namespace_anno != goterm.namespace:
gaf2errs[assc_name].append(nta)
else:
not_found.add(nta.GO_ID)
print('{HMS} {N} Associations'.format(HMS=_hms(TIC), N=len(assc_names)))
if not_found:
_prt_not_found(not_found)
if gaf2errs:
_wr_errs('namespace_errors.txt', gaf2errs, go2obj)
def test_semantic_similarity(usr_assc=None):
"""Computing basic semantic similarities between GO terms."""
not_these = {'goa_uniprot_all.gaf', 'goa_uniprot_all_noiea.gaf'}
associations = sorted(ASSOCIATIONS.difference(not_these))
go2obj = get_go2obj()
# goids = go2obj.keys()
# http://current.geneontology.org/annotations/
if usr_assc is not None:
associations = [usr_assc]
not_found = set()
errs = []
for assc_name in associations: # Limit test numbers for speed
tic = timeit.default_timer()
# Get all the annotations from arabidopsis.
fin_gaf = os.path.join(REPO, assc_name)
if not os.path.exists(fin_gaf):
dnld_annotation(fin_gaf)
annoobj = GafReader(fin_gaf)
#### for nspc in ['BP', 'MF', 'CC']:
assc_gene2gos = annoobj.get_id2gos('all')
if not assc_gene2gos:
not_found.add(assc_name)
continue
# Calculate the information content of the single term, GO:0048364
# "Information content (GO:0048364) = 7.75481392334
# Initialize the counts of each GO term.
tcntobj = TermCounts(go2obj, assc_gene2gos)
go_cnt = tcntobj.gocnts.most_common()
#print tcntobj.gocnts.most_common()
if go_cnt:
print("{ASSC}".format(ASSC=assc_name))
print(tcntobj.aspect_counts)
gocnt_max = go_cnt[0][1]
prt_info(tcntobj, go_cnt, None)
prt_info(tcntobj, go_cnt, gocnt_max / 2.0)
prt_info(tcntobj, go_cnt, gocnt_max / 10.0)
print("{HMS} {hms} {ASSC}\n".format(ASSC=assc_name,
HMS=_hms(TIC),
hms=_hms(tic)))
print('{HMS} {N} Associations'.format(HMS=_hms(TIC), N=len(associations)))
if not_found:
_prt_not_found(not_found)
if errs:
fout_err = 'namespace_errors.txt'
with open(fout_err, 'w') as prt:
for err in errs:
prt.write(err)
print(' {N} ERRORS WROTE: {TXT}'.format(N=len(errs),
TXT=fout_err))
def test_termcnt_init():
"""Compare GOATOOLS Resnik scores and Yang Resnik scores"""
godag = get_godag(os.path.join(REPO, 'go-basic.obo'))
fin_gpad = os.path.join(REPO, 'goa_human.gpad')
dnld_annotation(fin_gpad)
# Load all annoations (BP, MF, CC)
top_cnt_all = _run_full(fin_gpad, godag)
# Load one annoation (BP, MF, CC) at a time
top_cnt_ns = _run_each(fin_gpad, godag)
# Compare different load methods
assert top_cnt_all == top_cnt_ns
def test_find_enrichment(run_all=False):
"""RUn an enrichments using all annotation file formats"""
if run_all:
fin_obo = join(REPO, 'go-basic.obo')
get_godag(fin_obo, optional_attrs={'relationship'}, loading_bar=None)
fin_gaf = join(REPO, 'goa_human.gaf')
dnld_annotation(fin_gaf)
for idx, cmd in enumerate(_get_cmds()):
print('------------------- TEST {I} ------------------------------------'.format(I=idx))
print('CMD: {CMD}'.format(CMD=cmd))
assert system(cmd) == 0
print("TEST PASSED")
else:
print('RUN THIS TEST WITH AN ARGUMENT')
def test_tcntobj_relationships(prt=sys.stdout):
"""Test loading of relationships, like part_of, into TermCounts"""
fin_obo = os.path.join(REPO, "go-basic.obo")
fin_anno = os.path.join(REPO, 'goa_human.gpad')
download_go_basic_obo(fin_obo, prt, loading_bar=None)
dnld_annotation(fin_anno)
# Load ontologies
go2obj_r0 = GODag(fin_obo)
go2obj_r1 = GODag(fin_obo, optional_attrs=['relationship'])
# Load annotations
annoobj = GpadReader(fin_anno, godag=go2obj_r0)
# Create TermCounts objects
ns2tcntobj_r0 = {ns:TermCounts(go2obj_r0, annoobj.get_id2gos(ns)) for ns in NSS}
ns2tcntobj_r1 = {ns:TermCounts(go2obj_r1, annoobj.get_id2gos(ns), RELS) for ns in NSS}
_chk_pass_fail(ns2tcntobj_r0, ns2tcntobj_r1)
def get_anno_fullname(fin_anno):
"""Get annotation filename"""
fin_full = join(REPO, fin_anno)
dnld_annotation(fin_full)
return fin_full