def get_db_refs_by_name(ns, name, node_data): """Return standard name and grounding based on a namespace and a name. Parameters ---------- ns : str A name space in which the given name is interpreted. name : str The name in the given name space to get grounding for. node_data : dict Node data for logging purposes. Returns ------- name : str The standardized name for the given entity. db_refs : dict The grounding for the given entity. """ db_refs = None if ns == 'HGNC': # Assumption: name is an HGNC symbol hgnc_id = hgnc_client.get_current_hgnc_id(name) if not hgnc_id: logger.info("Invalid HGNC name: %s (%s)" % (name, node_data)) return name, None elif isinstance(hgnc_id, list): logger.info('More than one current HGNC ID for %s, choosing %s' % (name, hgnc_id[0])) hgnc_id = hgnc_id[0] name = hgnc_client.get_hgnc_name(hgnc_id) db_refs = {'HGNC': hgnc_id} up_id = _get_up_id(hgnc_id) if up_id: db_refs['UP'] = up_id mirbase_id = mirbase_client.get_mirbase_id_from_hgnc_id(hgnc_id) if mirbase_id: db_refs['MIRBASE'] = mirbase_id elif ns in ('UNIPROT', 'UP'): up_id = None # This is a simple test to see if name is a valid UniProt ID, # if we can't get a mnemonic, we assume it's not a UP ID if uniprot_client.get_mnemonic(name, web_fallback=False): up_id = name # We next check if it's a mnemonic else: up_id_from_mnem = uniprot_client.get_id_from_mnemonic(name) if up_id_from_mnem: up_id = up_id_from_mnem if not up_id: logger.info('Couldn\'t get UP ID from %s' % name) return name, None db_refs = {'UP': up_id} hgnc_id = uniprot_client.get_hgnc_id(up_id) if hgnc_id: db_refs['HGNC'] = hgnc_id name = hgnc_client.get_hgnc_name(hgnc_id) else: name = uniprot_client.get_gene_name(up_id) elif ns == 'FPLX': db_refs = {'FPLX': name} elif ns in ('GO', 'GOBP', 'GOCC'): if name == 'cell proliferation': name = 'cell population proliferation' go_id = go_client.get_go_id_from_label(name) if not go_id: logger.info('Could not find GO ID for %s' % name) return name, None db_refs = {'GO': go_id} name = go_client.get_go_label(go_id) elif ns in ('MESHPP', 'MESHD', 'MESH'): mesh_id, mesh_name = mesh_client.get_mesh_id_name(name) if not mesh_id: logger.info('Could not find MESH ID from %s' % name) return name, None name = mesh_name db_refs = {'MESH': mesh_id} # For now, handle MGI/RGD but putting the name into the db_refs so # it's clear what namespace the name belongs to # FIXME: Full implementation would look up MGI/RGD identifiers from # the names, and obtain corresponding Uniprot IDs elif ns == 'MGI': up_id = mouse_lookup.get(name) if up_id: db_refs = {'UP': up_id} elif ns == 'RGD': up_id = rat_lookup.get(name) if up_id: db_refs = {'UP': up_id} # Map Selventa families and complexes to FamPlex elif ns == 'SFAM': db_refs = {'SFAM': name} indra_name = bel_to_indra.get(name) if indra_name is None: logger.info('Could not find mapping for BEL/SFAM family: ' '%s (%s)' % (name, node_data)) else: db_refs['FPLX'] = indra_name name = indra_name elif ns == 'SCOMP': db_refs = {'SCOMP': name} indra_name = bel_to_indra.get(name) if indra_name is None: logger.info('Could not find mapping for BEL/SCOMP complex: ' '%s (%s)' % (name, node_data)) else: db_refs['FPLX'] = indra_name name = indra_name # Map Entrez genes to HGNC/UP elif ns in ('EGID', 'ENTREZ', 'NCBIGENE'): hgnc_id = hgnc_client.get_hgnc_from_entrez(name) db_refs = {'EGID': name} if hgnc_id is not None: db_refs['HGNC'] = hgnc_id name = hgnc_client.get_hgnc_name(hgnc_id) up_id = hgnc_client.get_uniprot_id(hgnc_id) if up_id: db_refs['UP'] = up_id else: logger.info( 'HGNC entity %s with HGNC ID %s has no ' 'corresponding Uniprot ID.', name, hgnc_id) mirbase_id = mirbase_client.get_mirbase_id_from_hgnc_id(hgnc_id) if mirbase_id: db_refs['MIRBASE'] = mirbase_id else: logger.debug('Could not map EGID%s to HGNC.' % name) name = 'E%s' % name elif ns == 'MIRBASE': mirbase_id = mirbase_client.get_mirbase_id_from_mirbase_name(name) if not mirbase_id: logger.info('Could not map miRBase name %s to ID', name) return name, None db_refs = {'MIRBASE': mirbase_id} hgnc_id = mirbase_client.get_hgnc_id_from_mirbase_id(mirbase_id) if hgnc_id: db_refs['HGNC'] = hgnc_id name = hgnc_client.get_hgnc_name(hgnc_id) # CHEBI elif ns == 'CHEBI': # We first look up BEL's own namespace map for ChEBI names to IDs chebi_id = chebi_name_id.get(name) # If that fails, we look up INDRA's ChEBI name to ID mapping if not chebi_id: chebi_id = chebi_client.get_chebi_id_from_name(name) if chebi_id: db_refs = {'CHEBI': chebi_id} else: logger.info('CHEBI name %s not found in map.' % name) # These appear in the name slot but are actually IDs elif ns == 'CHEBIID': chebi_id = identifiers.ensure_chebi_prefix(name) db_refs = {'CHEBI': chebi_id} name = chebi_client.get_chebi_name_from_id(chebi_id) # SDIS, SCHEM: Include the name as the ID for the namespace elif ns in ('SDIS', 'SCHEM', 'TEXT'): db_refs = {ns: name} elif ns == 'TAX': tid = taxonomy_client.get_taxonomy_id(name) if tid: db_refs = {'TAXONOMY': tid} else: logger.info('Could not get taxonomy ID for %s' % name) else: logger.info("Unhandled namespace: %s: %s (%s)" % (ns, name, node_data)) return name, db_refs
def refinement_of(self, other, ontology, entities_refined=False): from indra.databases import go_client # Make sure the Agent types match if type(self) != type(other): return False # ENTITIES # Check that the basic entity of the agent either matches or is related # to the entity of the other agent. If not, no match. # If the entities, match, then we can continue if not (entities_refined or (self.entity_matches(other) or self.isa(other, ontology))): return False # BOUND CONDITIONS # Now check the bound conditions. For self to be a refinement of # other in terms of the bound conditions, it has to include all of the # bound conditions in the other agent, and add additional context. # TODO: For now, we do not check the bound conditions of the bound # conditions. # FIXME: This matching procedure will get confused if the same # entity is included more than once in one of the sets--this will # be picked up as a match # Iterate over the bound conditions in the other agent, and make sure # they are all matched in self. for bc_other in other.bound_conditions: # Iterate over the bound conditions in self to find a match bc_found = False for bc_self in self.bound_conditions: if (bc_self.is_bound == bc_other.is_bound) and \ bc_self.agent.refinement_of(bc_other.agent, ontology): bc_found = True # If we didn't find a match for this bound condition in self, then # no refinement if not bc_found: return False # MODIFICATIONS # Similar to the above, we check that self has all of the modifications # of other. # Here we need to make sure that a mod in self.mods is only matched # once to a mod in other.mods. Otherwise ('phoshporylation') would be # considered a refinement of ('phosphorylation', 'phosphorylation') matched_indices = [] # This outer loop checks that each modification in the other Agent # is matched. for other_mod in other.mods: mod_found = False # We need to keep track of indices for this Agent's modifications # to make sure that each one is used at most once to match # the modification of one of the other Agent's modifications. for ix, self_mod in enumerate(self.mods): if self_mod.refinement_of(other_mod, ontology): # If this modification hasn't been used for matching yet if ix not in matched_indices: # Set the index as used matched_indices.append(ix) mod_found = True break # If we didn't find an exact match for this mod in other, then # no refinement if not mod_found: return False # MUTATIONS # Similar to the above, we check that self has all of the mutations # of other. matched_indices = [] # This outer loop checks that each mutation in the other Agent # is matched. for other_mut in other.mutations: mut_found = False # We need to keep track of indices for this Agent's mutations # to make sure that each one is used at most once to match # the mutation of one of the other Agent's mutations. for ix, self_mut in enumerate(self.mutations): if self_mut.refinement_of(other_mut): # If this mutation hasn't been used for matching yet if ix not in matched_indices: # Set the index as used matched_indices.append(ix) mut_found = True break # If we didn't find an exact match for this mut in other, then # no refinement if not mut_found: return False # LOCATION # If the other location is specified and this one is not then self # cannot be a refinement if self.location is None: if other.location is not None: return False # If both this location and the other one is specified, we check the # hierarchy. elif other.location is not None: # If the other location is part of this location then # self.location is not a refinement sl = go_client.get_go_id_from_label(self.location) ol = go_client.get_go_id_from_label(other.location) if not ontology.isa_or_partof('GO', sl, 'GO', ol): return False # ACTIVITY if self.activity is None: if other.activity is not None: return False elif other.activity is not None: if not self.activity.refinement_of(other.activity, ontology): return False # Everything checks out return True
def get_db_refs_by_name(ns, name, node_data): """Return standard name and grounding based on a namespace and a name. Parameters ---------- ns : str A name space in which the given name is interpreted. name : str The name in the given name space to get grounding for. node_data : dict Node data for logging purposes. Returns ------- name : str The standardized name for the given entity. db_refs : dict The grounding for the given entity. """ db_refs = None if ns == 'HGNC': hgnc_id = hgnc_client.get_hgnc_id(name) if not hgnc_id: logger.info("Invalid HGNC name: %s (%s)" % (name, node_data)) return name, None db_refs = {'HGNC': hgnc_id} up_id = _get_up_id(hgnc_id) if up_id: db_refs['UP'] = up_id mirbase_id = mirbase_client.get_mirbase_id_from_hgnc_id(hgnc_id) if mirbase_id: db_refs['MIRBASE'] = mirbase_id elif ns in ('UNIPROT', 'UP'): up_id = None gene_name = uniprot_client.get_gene_name(name) if gene_name: up_id = name else: up_id_from_mnem = uniprot_client.get_id_from_mnemonic(name) if up_id_from_mnem: up_id = up_id_from_mnem gene_name = uniprot_client.get_gene_name(up_id) if not up_id: logger.info('Couldn\'t get UP ID from %s' % name) return name, None db_refs = {'UP': up_id} if uniprot_client.is_human(up_id): hgnc_id = hgnc_client.get_hgnc_id(gene_name) if not hgnc_id: logger.info('Uniprot ID linked to invalid human gene ' 'name %s' % name) else: db_refs['HGNC'] = hgnc_id elif ns == 'FPLX': db_refs = {'FPLX': name} elif ns in ('GO', 'GOBP', 'GOCC'): go_id = go_client.get_go_id_from_label(name) if not go_id: logger.info('Could not find GO ID for %s' % name) return name, None db_refs = {'GO': go_id} elif ns in ('MESHPP', 'MESHD', 'MESH'): mesh_id = mesh_client.get_mesh_id_name(name) if not mesh_id: logger.info('Could not find MESH ID fro %s' % name) return name, None db_refs = {'MESH': mesh_id} # For now, handle MGI/RGD but putting the name into the db_refs so # it's clear what namespace the name belongs to # FIXME: Full implementation would look up MGI/RGD identifiers from # the names, and obtain corresponding Uniprot IDs elif ns in ('MGI', 'RGD'): db_refs = {ns: name} # Map Selventa families to FamPlexes elif ns == 'SFAM': db_refs = {'SFAM': name} indra_name = bel_to_indra.get(name) if indra_name is None: logger.info('Could not find mapping for BEL/SFAM family: ' '%s (%s)' % (name, node_data)) else: db_refs['FPLX'] = indra_name name = indra_name # Map Entrez genes to HGNC/UP elif ns in ('EGID', 'ENTREZ', 'NCBIGENE'): hgnc_id = hgnc_client.get_hgnc_from_entrez(name) db_refs = {'EGID': name} if hgnc_id is not None: db_refs['HGNC'] = hgnc_id name = hgnc_client.get_hgnc_name(hgnc_id) up_id = hgnc_client.get_uniprot_id(hgnc_id) if up_id: db_refs['UP'] = up_id else: logger.info('HGNC entity %s with HGNC ID %s has no ' 'corresponding Uniprot ID.', name, hgnc_id) mirbase_id = mirbase_client.get_mirbase_id_from_hgnc_id(hgnc_id) if mirbase_id: db_refs['MIRBASE'] = mirbase_id else: logger.info('Could not map EGID%s to HGNC.' % name) name = 'E%s' % name elif ns == 'MIRBASE': mirbase_id = mirbase_client.get_mirbase_id_from_mirbase_name(name) if not mirbase_id: logger.info('Could not map miRBase name %s to ID', name) return db_refs = {'MIRBASE': mirbase_id} hgnc_id = mirbase_client.get_hgnc_id_from_mirbase_id(mirbase_id) if hgnc_id: db_refs['HGNC'] = hgnc_id # CHEBI elif ns == 'CHEBI': chebi_id = chebi_name_id.get(name) if not chebi_id: chebi_id = chebi_client.get_chebi_id_from_name(name) if chebi_id: db_refs = {'CHEBI': chebi_id} else: logger.info('CHEBI name %s not found in map.' % name) # SDIS, SCHEM: Include the name as the ID for the namespace elif ns in ('SDIS', 'SCHEM'): db_refs = {ns: name} else: logger.info("Unhandled namespace: %s: %s (%s)" % (ns, name, node_data)) return name, db_refs
def test_go_label_to_id(): assert go_client.get_go_id_from_label('mitochondrion inheritance') == \ 'GO:0000001'
def test_partof_comp_not(): assert not bio_ontology.isa_or_partof( 'GO', go_client.get_go_id_from_label('cellular_component'), 'GO', go_client.get_go_id_from_label('cytoplasm'))