def __init__(self):
self.prt = sys.stdout
_fin_assc = os.path.join(REPO, "goa_human.gaf")
self.gene2gos_orig = dnld_assc(_fin_assc, go2obj=None, prt=self.prt)
self.go2genes_orig = get_b2aset(self.gene2gos_orig)
_num_genes = [len(gs) for gs in self.go2genes_orig.values()]
self.min_genes = min(_num_genes)
self.max_genes = max(_num_genes)
assert self.gene2gos_orig == get_b2aset(self.go2genes_orig)
def __init__(self, go2obj, annots, relationships=None, **kws):
'''
Initialise the counts and
'''
_prt = kws.get('prt')
# Backup
self.go2obj = go2obj # Full GODag
self.annots, go_alts = clean_anno(annots, go2obj, _prt)[:2]
# Genes annotated to all associated GO, including inherited up ancestors'
_relationship_set = RelationshipCombos(go2obj).get_set(relationships)
self.go2genes = self._init_go2genes(_relationship_set, go2obj)
self.gene2gos = get_b2aset(self.go2genes)
# Annotation main GO IDs (prefer main id to alt_id)
self.goids = set(self.go2genes.keys())
self.gocnts = Counter({go:len(geneset) for go, geneset in self.go2genes.items()})
# Get total count for each branch: BP MF CC
self.aspect_counts = {
'biological_process': self.gocnts.get(NAMESPACE2GO['biological_process'], 0),
'molecular_function': self.gocnts.get(NAMESPACE2GO['molecular_function'], 0),
'cellular_component': self.gocnts.get(NAMESPACE2GO['cellular_component'], 0)}
self._init_add_goid_alt(go_alts)
self.gosubdag = GoSubDag(
set(self.gocnts.keys()),
go2obj,
tcntobj=self,
relationships=_relationship_set,
prt=None)
if _prt:
self.prt_objdesc(_prt)
def _run_get_id2gos(annoobjs):
"""Test get_id2gos"""
for idx, obj in enumerate(annoobjs):
print('\n{I}) get_id2gos {DESC} {NSs} {N:,} annotations'.format(
I=idx,
DESC=obj.get_desc(),
NSs=obj.namespaces,
N=len(obj.associations)))
# If all namespaces are loaded, returns BP, else returns loaded NS
print('Load all evidence codes')
id2gos = obj.get_id2gos()
assert id2gos, 'NO ANNOTATIONS FOUND'
## print(next(iter(obj.associations)))
print('Load all evidence codes, except IEA')
id2gos_inc = obj.get_id2gos(ev_include=INC_GOOD)
id2gos_exc = obj.get_id2gos(ev_exclude={'IEA'})
assert id2gos_exc, 'NO NON-IEA ANNOTATIONS FOUND'
assert id2gos_inc == id2gos_exc, \
'INC ALL({A}) != EXC IEA({I}): {DIF}'.format(
A=len(id2gos_inc),
I=len(id2gos_exc),
# DIF=set(obj.get_id2gos(ev_inc=INC_GOOD).keys()).
# symmetric_difference(obj.get_id2gos(ev_exclude={'IEA'})))
DIF='')
num_ids = len(id2gos)
print('>>>>> {I} >>>>> get_id2gos {N:6,} go2ids[{B:6,}] {ANNO}'.format(
I=idx, N=num_ids, ANNO=obj.get_desc(), B=len(get_b2aset(id2gos))))
assert next(iter(next(iter(id2gos.values()))))[:3] == "GO:"
if obj.filename[-16:] == 'data/association' and obj.godag is None:
assert num_ids == 34284
def _get_id2gos(file_id2gos, godag, name2go):
"""Get annotations"""
if os.path.exists(file_id2gos):
return IdToGosReader(file_id2gos, godag=godag).get_id2gos('CC')
id2num = {
name2go['A']: 10,
name2go['B']: 10,
name2go['C']: 10,
name2go['D']: 10,
name2go['E']: 10,
name2go['F']: 10,
name2go['G']: 10,
name2go['H']: 10,
name2go['I']: 30,
name2go['L']: 30,
name2go['M']: 20,
name2go['N']: 30,
}
go2genes = cx.defaultdict(set)
genenum = 0
for goid, qty in id2num.items():
for _ in range(qty):
go2genes[goid].add(genenum)
genenum += 1
id2gos = get_b2aset(go2genes)
IdToGosReader.wr_id2gos(file_id2gos, id2gos)
return id2gos
def _adj_for_assc(self):
"""Print only GO IDs from associations and their ancestors."""
if self.gene2gos:
gos_assoc = set(get_b2aset(self.gene2gos).keys())
if 'item_marks' not in self.kws:
self.kws['item_marks'] = {go: '>' for go in gos_assoc}
if 'include_only' not in self.kws:
gosubdag = GoSubDag(gos_assoc, self.gosubdag.go2obj,
self.gosubdag.relationships)
self.kws['include_only'] = gosubdag.go2obj
def _get_id2gos(file_id2gos, godag, name2go, name2num):
"""Get annotations"""
if os.path.exists(file_id2gos):
return IdToGosReader(file_id2gos, godag=godag).get_id2gos('CC')
go2genes = cx.defaultdict(set)
genenum = 0
for name, qty in name2num.items():
goid = name2go[name]
for _ in range(qty):
go2genes[goid].add(genenum)
genenum += 1
id2gos = get_b2aset(go2genes)
IdToGosReader.wr_id2gos(file_id2gos, id2gos)
return id2gos
def describe_assc(org, fin_assc, go2obj, obj, prt):
"""Report statistics for a single association."""
# Assc. | # Assc| range | 25th | median | 75th | mean | stddev
# ------------|-------|------------|------|--------|------|------|-------
# hsa GO/gene | 19394 | 1 to 212 | 5 | 9 | 17 | 13 | 14
# hsa gene/GO | 17277 | 1 to 8,897 | 1 | 3 | 8 | 15 | 120
#
# mus GO/gene | 19870 | 1 to 261 | 5 | 10 | 18 | 14 | 15
# mus gene/GO | 17491 | 1 to 7,009 | 1 | 3 | 8 | 16 | 129
#
# dme GO/gene | 12551 | 1 to 137 | 2 | 4 | 8 | 6 | 7
# dme gene/GO | 7878 | 1 to 1,675 | 1 | 3 | 7 | 10 | 41
gene2gos = dnld_assc(fin_assc, go2obj, prt=None) # Associations
go2genes = get_b2aset(gene2gos)
assert gene2gos
assert go2genes
cnts_gos_p_gene = [len(gos) for gos in gene2gos.values()]
cnts_genes_p_go = [len(genes) for genes in go2genes.values()]
obj.prt_data("{ORG} GO/gene".format(ORG=org), cnts_gos_p_gene, prt)
obj.prt_data("{ORG} gene/GO".format(ORG=org), cnts_genes_p_go, prt)
def get_annotations_reversed(self):
"""Return go2geneset for all GO IDs explicitly annotated to a gene"""
return set.union(*get_b2aset(self.annots))
def get_go2chrs(sec2gos, sec2chr):
"""Dict: given a GO return a set of letters representing it's section membership(s)."""
go2chrs = {}
for goid, sections in get_b2aset(sec2gos).items():
go2chrs[goid] = set(sec2chr[s] for s in sections)
return go2chrs
