def to_bel(self) -> BELGraph:
"""Convert miRBase to BEL."""
result = BELGraph()
for sequence in self._get_query(Sequence):
mirbase_node = sequence.as_pybel()
for xref in sequence.xrefs:
result.add_equivalence(
mirbase_node,
MicroRna(
namespace=xref.database,
identifier=xref.database_id,
))
return result
def test_mirna_reference(self):
self.assertEqual('m(HGNC:MIR1)',
str(MicroRna(namespace='HGNC', name='MIR1')))
CHEBI = 'CHEBI'
g1 = Gene(namespace=HGNC, name='1')
r1 = Rna(namespace=HGNC, name='1')
p1 = Protein(HGNC, name='1')
g2 = Gene(HGNC, name='2')
r2 = Rna(HGNC, name='2')
p2 = Protein(HGNC, name='2')
g3 = Gene(namespace=HGNC, name='3')
r3 = Rna(namespace=HGNC, name='3')
p3 = Protein(namespace=HGNC, name='3')
g4 = Gene(namespace=HGNC, name='4')
m4 = MicroRna(namespace=HGNC, name='4')
a5 = Abundance(namespace=CHEBI, name='5')
p5 = Pathology(namespace=GO, name='5')
class TestCollapseProteinInteractions(unittest.TestCase):
def test_protein_interaction_1(self):
graph = BELGraph()
graph.add_node_from_data(p1)
graph.add_node_from_data(p2)
graph.add_node_from_data(a5)
graph.add_node_from_data(p5)
graph.add_qualified_edge(p1, p2, relation=POSITIVE_CORRELATION, citation=n(), evidence=n())
)
BEL_THOROUGH_NODES = {
oxygen_atom,
tmprss2_erg_rna_fusion,
tmprss2_erg_rna_fusion_unspecified,
akt_methylated,
bcr_jak2_rna_fusion,
chchd4_aifm1_rna_fusion,
akt1_gene,
akt1_phe_508_del,
akt1,
Gene('HGNC', 'AKT1', variants=Hgvs('c.308G>A')),
tmprss2_erg_gene_fusion,
Gene('HGNC', 'AKT1', variants=[Hgvs('c.1521_1523delCTT'), Hgvs('c.308G>A'), Hgvs('p.Phe508del')]),
MicroRna('HGNC', 'MIR21'),
bcr_jak2_gene_fusion,
Gene('HGNC', 'CFTR', variants=Hgvs('c.1521_1523delCTT')),
Gene('HGNC', 'CFTR'),
Gene('HGNC', 'CFTR', variants=Hgvs('g.117199646_117199648delCTT')),
Gene('HGNC', 'CFTR', variants=Hgvs('c.1521_1523delCTT')),
Protein('HGNC', 'AKT1', variants=ProteinModification('Ph', 'Ser', 473)),
MicroRna('HGNC', 'MIR21', variants=Hgvs('p.Phe508del')),
Protein('HGNC', 'AKT1', variants=Hgvs('p.C40*')),
Protein('HGNC', 'AKT1', variants=[Hgvs('p.Ala127Tyr'), ProteinModification('Ph', 'Ser')]),
chchd4_aifm1_gene_fusion,
tmprss2_erg_protein_fusion,
Protein('HGNC', 'AKT1', variants=Hgvs('p.Arg1851*')),
bcr_jak2_protein_fusion,
Protein('HGNC', 'AKT1', variants=Hgvs('p.40*')),
chchd4_aifm1_protein_fusion,
def _rela(x, y=None):
return {RELATION: x, OBJECT: activity(y)}
def _assoc(y):
return {RELATION: ASSOCIATION, 'association_type': y}
a1 = Abundance('CHEBI', '1')
p1 = Protein('HGNC', '1')
pf1 = Protein('INTERPRO', '1')
d1 = Pathology('MESH', '1')
b1 = BiologicalProcess('GO', '1')
b2 = BiologicalProcess('GO', '2')
m1 = MicroRna('MIRBASE', '1')
r1 = Rna('HGNC', '1')
r2 = Rna('HGNC', '2')
nca1 = NamedComplexAbundance('FPLX', '1')
pop1 = Population('taxonomy', '1')
p2 = Protein('HGNC', identifier='9236')
p3 = Protein('HGNC', identifier='9212')
r3 = p3.get_rna()
g3 = r3.get_gene()
c1 = ComplexAbundance([p2, g3])
c2 = ComplexAbundance([p1, p2])
c3 = ComplexAbundance([a1, p2])
converters_true_list = [
def normalize_graph_names(graph: BELGraph, database: str) -> None:
"""Normalize graph names."""
# Victim to Survivor (one to one node) mapping
one_to_one_mapping = {}
# Victim to Survivors (one to many nodes) mapping
one_to_many_mapping = defaultdict(set)
for node in graph.nodes():
# Skip ListAbundances and Reactions since they do not have a name
if isinstance(node, ListAbundance) or isinstance(
node, Reaction) or not node.name:
continue
# Normalize names: Lower case name and strip quotes or white spaces
lower_name = node.name.lower().strip('"').strip()
# Dealing with Genes/miRNAs
if isinstance(node, CentralDogma):
##################
# miRNA entities #
##################
if lower_name.startswith("mir"):
# Reactome preprocessing to flat multiple identifiers
if database == REACTOME:
reactome_cell = munge_reactome_gene(lower_name)
if isinstance(reactome_cell, list):
for lower_name in reactome_cell:
one_to_many_mapping[node].add(
MicroRna(
node.namespace,
name=lower_name.replace("mir-", "mir"),
identifier=node.identifier,
), )
if lower_name.endswith(' genes'):
lower_name = lower_name[:-len(' genes')]
elif lower_name.endswith(' gene'):
lower_name = lower_name[:-len(' gene')]
one_to_one_mapping[node] = MicroRna(
node.namespace,
name=lower_name.replace(
"mir-", "mir"), # Special case for Reactome
)
continue
# KEGG and Reactome
one_to_one_mapping[node] = MicroRna(
node.namespace,
name=node.name.replace("mir-", "mir"),
identifier=node.identifier,
)
##################
# Genes entities #
##################
else:
# Reactome preprocessing to flat multiple identifiers
if database == REACTOME:
reactome_cell = munge_reactome_gene(lower_name)
if isinstance(reactome_cell, list):
for lower_name in reactome_cell:
if lower_name in BLACK_LIST_REACTOME: # Filter entities in black list
continue
elif lower_name.startswith(
"("): # remove redundant parentheses
lower_name = lower_name.strip("(").strip(")")
one_to_many_mapping[node].add(
Protein(node.namespace,
name=lower_name,
identifier=node.identifier), )
else:
one_to_one_mapping[node] = Protein(
node.namespace,
name=lower_name,
identifier=node.identifier)
continue
# WikiPathways and KEGG do not require any processing of genes
elif database == WIKIPATHWAYS and lower_name in WIKIPATHWAYS_BIOL_PROCESS:
one_to_one_mapping[node] = BiologicalProcess(
node.namespace,
name=lower_name,
identifier=node.identifier,
)
continue
one_to_one_mapping[node] = Protein(node.namespace,
name=lower_name,
identifier=node.identifier)
#######################
# Metabolite entities #
#######################
elif isinstance(node, Abundance):
if database == 'wikipathways':
# Biological processes that are captured as abundance in
# BEL since they were characterized wrong in WikiPathways
if lower_name in WIKIPATHWAYS_BIOL_PROCESS:
one_to_one_mapping[node] = BiologicalProcess(
node.namespace,
name=lower_name,
identifier=node.identifier,
)
continue
# Abundances to BiologicalProcesses
elif (node.namespace
in {'WIKIDATA', 'WIKIPATHWAYS', 'REACTOME'}
and lower_name not in WIKIPATHWAYS_METAB):
one_to_one_mapping[node] = BiologicalProcess(
node.namespace,
name=lower_name,
identifier=node.identifier,
)
continue
# Fix naming in duplicate entity
if lower_name in WIKIPATHWAYS_NAME_NORMALIZATION:
lower_name = WIKIPATHWAYS_NAME_NORMALIZATION[lower_name]
elif database == REACTOME:
# Curated proteins that were coded as metabolites
if lower_name in REACTOME_PROT:
one_to_one_mapping[node] = Protein(
node.namespace,
name=lower_name,
identifier=node.identifier,
)
continue
# Flat multiple identifiers (this is not trivial because most of ChEBI names contain commas,
# so a clever way to fix some of the entities is to check that all identifiers contain letters)
elif "," in lower_name and all(
string.isalpha() for string in lower_name.split(",")):
for string in lower_name.split(","):
one_to_many_mapping[node].add(
Abundance(node.namespace,
name=string,
identifier=node.identifier), )
continue
one_to_one_mapping[node] = Abundance(node.namespace,
name=lower_name,
identifier=node.identifier)
#################################
# Biological Processes entities #
#################################
elif isinstance(node, BiologicalProcess):
# KEGG normalize name by removing the title prefix
if lower_name.startswith('title:'):
lower_name = lower_name[len('title:'):]
one_to_one_mapping[node] = BiologicalProcess(
node.namespace,
name=lower_name,
identifier=node.identifier,
)
relabel_nodes(graph, one_to_one_mapping)
multi_relabel(graph, one_to_many_mapping)
'HGNC', 'TMPRSS2'),
partner_3p=Rna('HGNC', 'ERG'))
BEL_THOROUGH_NODES = {
oxygen_atom, tmprss2_erg_rna_fusion, tmprss2_erg_rna_fusion_unspecified,
akt_methylated, bcr_jak2_rna_fusion, chchd4_aifm1_rna_fusion, akt1_gene,
akt1_phe_508_del, akt1,
Gene('HGNC', 'AKT1', variants=Hgvs('c.308G>A')), tmprss2_erg_gene_fusion,
Gene('HGNC',
'AKT1',
variants=[
Hgvs('c.1521_1523delCTT'),
Hgvs('c.308G>A'),
Hgvs('p.Phe508del')
]),
MicroRna('HGNC', 'MIR21'), bcr_jak2_gene_fusion,
Gene('HGNC', 'CFTR', variants=Hgvs('c.1521_1523delCTT')),
Gene('HGNC', 'CFTR'),
Gene('HGNC', 'CFTR', variants=Hgvs('g.117199646_117199648delCTT')),
Gene('HGNC', 'CFTR', variants=Hgvs('c.1521_1523delCTT')),
Protein('HGNC', 'AKT1', variants=ProteinModification('Ph', 'Ser', 473)),
MicroRna('HGNC', 'MIR21', variants=Hgvs('p.Phe508del')),
Protein('HGNC', 'AKT1', variants=Hgvs('p.C40*')),
Protein('HGNC',
'AKT1',
variants=[Hgvs('p.Ala127Tyr'),
ProteinModification('Ph', 'Ser')]),
chchd4_aifm1_gene_fusion, tmprss2_erg_protein_fusion,
Protein('HGNC', 'AKT1',
variants=Hgvs('p.Arg1851*')), bcr_jak2_protein_fusion,
Protein('HGNC', 'AKT1',