def get_gene_sets(gene_set_names: list, organism: str = ORGANISM, go_slims: bool = False,
set_sizes: tuple = (-np.inf, np.inf), reference: set = None) -> dict:
"""
Get all gene sets.
:param gene_set_names: Names of ontologies for which to get gene sets (as returned by list_gene_sets)
:param organism: organism id
:param go_slims: If ontology type (first element from tuples retured by list_gene_sets) is GO then output
only gene sets that are in 'goslim_generic'
:param set_sizes: Use only gene sets with size greater or equal to 1st element of set_sizes and smaller or equal to
2nd element of set_sizes, default is -inf,inf
:param reference: List of gene EIDs to use as a reference set. Gene set filtering based on gene set size takes in
account only genes present in reference.
:return: Dict with key is ontology name and values are its GeneSet objects
"""
gene_set_ontology = dict()
if go_slims:
anno = go.Annotations(organism)
anno._ensure_ontology()
anno._ontology.set_slims_subset('goslim_generic')
slims = anno._ontology.slims_subset
for gene_set_name in gene_set_names:
gene_sets = load_gene_sets(gene_set_name, str(organism))
if reference is None:
gene_sets = [gene_set for gene_set in gene_sets if set_sizes[0] <= len(gene_set.genes) <= set_sizes[1]]
else:
gene_sets = [gene_set for gene_set in gene_sets if
set_sizes[0] <= len(gene_set.genes & reference) <= set_sizes[1]]
if go_slims and gene_set_name[0] == 'GO':
gene_sets = [gene_set for gene_set in gene_sets if gene_set.gs_id in slims]
gene_set_ontology[gene_set_name] = gene_sets
return gene_set_ontology
def go_gene_sets(tax_id: str) -> None:
domain = 'go'
ontology = go.Ontology(filename=f'{data_path}/{domain}/gene_ontology.obo')
annotations = go.Annotations(tax_id,
filename=f'{data_path}/{domain}/{tax_id}.tab',
ontology=ontology)
def to_gene_set(term: go.Term) -> Optional[GeneSet]:
genes = annotations.get_genes_by_go_term(term.id)
if len(genes) > 0:
return GeneSet(
gs_id=term.id,
name=term.name,
genes=set(genes),
hierarchy=('GO', term.namespace),
organism=tax_id,
link=f'http://amigo.geneontology.org/amigo/term/{term.id}')
gene_sets = GeneSets([
gs for gs in [to_gene_set(term) for term in ontology.terms.values()]
if gs is not None
])
for gs_group in gene_sets.split_by_hierarchy():
hierarchy = gs_group.common_hierarchy()
gs_group.to_gmt_file_format(
f'{data_path}/gene_sets/{filename(hierarchy, tax_id)}')
def go_gene_sets(org):
""" Returns gene sets from GO.
"""
ontology = go.Ontology()
annotations = go.Annotations(org, ontology=ontology)
gene_sets = []
for termn, term in ontology.terms.items():
genes = annotations.get_genes_by_go_term(termn)
hier = ('GO', term.namespace)
if len(genes) > 0:
gs = GeneSet(gs_id=termn, name=term.name, genes=genes, hierarchy=hier,
organism=org, link=GO_TERM_LINK.format(termn))
gene_sets.append(gs)
return GeneSets(gene_sets)
def Load(self):
a = self.available_annotations[self.annotation_index]
if self.ontology is None:
self.ontology = go.Ontology()
if a.taxid != self.loaded_annotation_code:
self.annotations = None
gc.collect() # Force run garbage collection
self.annotations = go.Annotations(a.taxid)
self.loaded_annotation_code = a.taxid
count = defaultdict(int)
geneSets = defaultdict(set)
for anno in self.annotations.annotations:
count[anno.evidence] += 1
geneSets[anno.evidence].add(anno.gene_id)
for etype in go.evidenceTypesOrdered:
ecb = self.evidenceCheckBoxDict[etype]
ecb.setEnabled(bool(count[etype]))
ecb.setText(etype + ": %i annots(%i genes)" % (count[etype], len(geneSets[etype])))
def GO_enrichment(entrez_ids: list, organism: int = ORGANISM, fdr=0.25, slims: bool = True,
aspect: str = None) -> OrderedDict:
"""
Calulate onthology enrichment for list of genes
:param entrez_ids: entrez IDs of gene group to be analysed for enrichemnt
:param organism: organism ID
:param fdr: For retention of enriched gene sets
:param slims: From Orange Annotations
:param aspect: Which GO aspect to use. From Orange Annotations: None: all, 'Process', 'Function', 'Component'
:return: Dict: key ontology term, value FDR. Sorted by FDR ascendingly.
"""
anno = go.Annotations(organism)
enrichment = anno.get_enriched_terms(entrez_ids, slims_only=slims, aspect=aspect)
filtered = go.filter_by_p_value(enrichment, fdr)
enriched_data = dict()
for go_id, data in filtered.items():
terms = anno.get_annotations_by_go_id(go_id)
for term in terms:
if term.go_id == go_id:
padj = data[1]
enriched_data[term.go_term] = padj
break
enriched_data = OrderedDict(sorted(enriched_data.items(), key=lambda x: x[1]))
return enriched_data
from orangecontrib.bioinformatics import go
ontology = go.Ontology()
annotations = go.Annotations("4932", ontology=ontology)
# keys are symbol names, values are Entrez IDs
genes_ids = {'Yta7p': '853186', 'RPN2': '854735', 'RPT2': '851557'}
res = annotations.get_enriched_terms(genes_ids.values())
print(res)
print("Enriched terms:")
for go_id, (genes, p_value, ref) in res.items():
if p_value < 0.05:
print(ontology[go_id].name + " with p-value: %.4f " % p_value +
", ".join(genes))
# And again for slims
annotations.ontology.set_slims_subset('goslim_yeast')
res = annotations.get_enriched_terms(genes_ids.values(), slims_only=True)
print("\n\nEnriched slim terms:")
for go_id, (genes, p_value, _) in res.items():
if p_value < 0.2:
print(ontology[go_id].name + " with p-value: %.4f " % p_value +
", ".join(genes))
