def get_gene_ontology_access(self):
"""
Loads all of the relations between the UNIPROTs and GOs as one giant dictionary
"""
uniprots_without_gene_ontology_terms = 0
log.info('Starting GO matrix mapping starting from %s uniprots', len(self.InitSet))
for uniprot_bulbs_id in self.InitSet:
uniprot_specific_gos = []
up_node = DatabaseGraph.UNIPORT.get(uniprot_bulbs_id)
# Yeah, this is a dangerous one, we need to avoid failure on that one
self.UP_Names[uniprot_bulbs_id] = [up_node.ID, up_node.displayName]
attached_go_nodes = up_node.bothV("is_go_annotation")
if attached_go_nodes:
for go_node in attached_go_nodes:
if go_node.Namespace in self.go_namespace_filter:
go_node_bulbs_id = get_bulbs_id(go_node)
uniprot_specific_gos.append(go_node_bulbs_id)
self.GO2UP[go_node_bulbs_id].append(uniprot_bulbs_id)
self.SeedSet.add(go_node_bulbs_id)
if not uniprot_specific_gos:
uniprots_without_gene_ontology_terms += 1
log.debug("UP without GO was found. UP bulbs_id: %s, \t name: %s",
uniprot_bulbs_id, self.UP_Names[uniprot_bulbs_id])
self.UPs_without_GO.add(uniprot_bulbs_id)
else:
self.UP2GO_Dict[uniprot_bulbs_id] = copy(uniprot_specific_gos)
log.info('total number of UPs without a go_node annotation: %s out of %s',
uniprots_without_gene_ontology_terms, len(self.InitSet))
def compute_uniprot_attachments(self):
"""
Computes the dictionary of attachments between the reached_uniprots_bulbs_id_list and
Reactome proteins
"""
log.info('attaching reactome proteins to uniprot nodes')
uniprot_attachments_counter = 0
reactome_attachments_counter = 0
for uniprot_bulbs_id in self.reached_uniprots_bulbs_id_list:
bulbs_node = DatabaseGraph.UNIPORT.get(uniprot_bulbs_id)
reactome_bulbs_id_generator = bulbs_node.bothV("is_same")
if reactome_bulbs_id_generator is not None:
self.Uniprot_attachments[uniprot_bulbs_id] = []
for uniprot_alias in reactome_bulbs_id_generator:
self.Uniprot_attachments[uniprot_bulbs_id].append(get_bulbs_id(uniprot_alias))
uniprot_attachments_counter += 1
reactome_attachments_counter += len(self.Uniprot_attachments[uniprot_bulbs_id])
log.debug('attached %s Reactome proteins to the node %s',
len(self.Uniprot_attachments[uniprot_bulbs_id]), uniprot_bulbs_id)
else:
log.debug('No attachment for the node %s', uniprot_bulbs_id)
log.info('Attached %s reactome protein nodes to %s / %s uniprot nodes',
reactome_attachments_counter,
uniprot_attachments_counter, len(self.reached_uniprots_bulbs_id_list))
def insert_meta_objects(bulbs_graph_class, meta_id_2_property_dict):
"""
Inserst a Meta-Object (I.e. any physical entity or collection thereof) as a member of a
bulbs class and pumping the bioflow information from the property bioflow
:param bulbs_graph_class:
:param meta_id_2_property_dict:
"""
total_properties = len(meta_id_2_property_dict)
config = bulbs_graph_class.client.config
element_type = bulbs_graph_class.element_class.get_element_type(config)
for i, (meta_name,
property_dict) in enumerate(meta_id_2_property_dict.iteritems()):
if i * 20 % total_properties < 21:
log.info('insert %s: %s out of %s', element_type, i,
total_properties)
primary = bulbs_graph_class.create(
ID=meta_name,
displayName=property_dict['displayName'],
localization=property_dict['cellularLocation'],
main_connex=False)
if meta_name in Leg_ID_Filter:
ForbiddenIDs.append(get_bulbs_id(primary))
memoization_dict[meta_name] = primary
insert_minimal_annotations(
memoization_dict[meta_name],
meta_id_2_property_dict[meta_name]['references'])
if 'cellularLocation' in meta_id_2_property_dict[meta_name].keys():
secondary = memoization_dict[meta_id_2_property_dict[meta_name]
['cellularLocation']]
DatabaseGraph.is_localized.create(primary,
secondary,
costum_from=primary.ID,
costum_to=secondary.ID)
if 'modification' in meta_id_2_property_dict[meta_name].keys():
for modification in meta_id_2_property_dict[meta_name][
'modification']:
if 'location' in modification.keys(
) and 'modification' in modification.keys():
located_modification = DatabaseGraph.ModificationFeature.create(
ID=modification['ID'],
type="post-translational_Mod",
location=modification['location'],
displayName=modification['modification'])
DatabaseGraph.is_able_to_modify.create(
primary,
located_modification,
costum_from=primary.ID,
costum_to=located_modification.ID)
def pull_up_acc_nums_from_reactome():
"""
Attempts to retrieve accession numbers nums from the neo4j database
:return: dict that maps acnums to nodes in the database to which they point (Reactome proteins)
:raise Exception: if the generator that performs a lookup of existing uniprot acnum annotation
nodes is null (Reactome wasn't imported yet) or a wrong uniprot is crosslinked with a wrong
reactome
"""
acc_num_annot_nodes = DatabaseGraph.AnnotNode.index.lookup(ptype='UniProt')
if acc_num_annot_nodes is None:
raise Exception(
"Reactome was not loaded or contains no acc_num cross-references to Uniprot"
)
acc_num_dict = {} # acnum to AnnotNode
for annotation_node in acc_num_annot_nodes:
if annotation_node is not None:
annot_obj = DatabaseGraph.vertices.get(
get_bulbs_id(annotation_node))
acc_num_dict[str(annot_obj.payload)] = annot_obj
if len(acc_num_dict) < 10:
raise Exception(
"Reactome was not loaded or contains no acc_num cross-references to Uniprot"
)
reactome_proteins = {}
for acc_num in acc_num_dict.keys():
reactome_proteins_generator = acc_num_dict[acc_num].bothV()
reactome_proteins[acc_num] = []
if reactome_proteins_generator is not None:
for vertex in reactome_proteins_generator:
if vertex is not None:
reactome_proteins[acc_num].append(vertex)
if reactome_proteins[acc_num] != 1:
log.debug(
'Cross-linking reactome v.s. acc_num %s mapped to %s proteins',
acc_num, len(reactome_proteins))
log.info('Cross-linked %s proteins from reactome v.s. Uniprot',
len(reactome_proteins))
return reactome_proteins
def get_gene_ontology_structure(self):
"""
Loads all of the relations between the GOs that are generalisation of the seedList
GOs and that are withing the types specified in go_namespace_filter
"""
visited_set = set()
seeds_list = copy(list(self.SeedSet))
log.info('Starting gene ontology structure retrieval from the set of %s seeds',
len(self.SeedSet))
while seeds_list:
node_id = seeds_list.pop()
visited_set.add(node_id)
local_uniprot_list = []
local_regulation_list = []
local_up_regulation_list = []
local_down_regulation_list = []
gene_ontology_node = DatabaseGraph.GOTerm.get(node_id)
self.GO_Names[node_id] = str(gene_ontology_node.displayName)
self.GO_Legacy_IDs[node_id] = str(gene_ontology_node.ID)
self.rev_GO_IDs[str(gene_ontology_node.ID)] = node_id
for relation_type in chain(self._GOUpTypes, self._GORegTypes):
related_go_nodes = gene_ontology_node.outV(relation_type)
if not related_go_nodes:
continue # skip in case GO Node has no outgoing relations to other GO nodes
for go_node in related_go_nodes:
if go_node.Namespace not in self.go_namespace_filter:
continue
node_bulbs_id = get_bulbs_id(go_node)
if node_bulbs_id not in visited_set:
seeds_list.append(node_bulbs_id)
if relation_type in self._GOUpTypes:
local_uniprot_list.append(node_bulbs_id)
else:
local_regulation_list.append(node_bulbs_id)
rev_generator = gene_ontology_node.inV(relation_type)
if not rev_generator:
continue
for go_node in rev_generator:
if go_node.Namespace not in self.go_namespace_filter:
continue
node_bulbs_id = get_bulbs_id(go_node)
if relation_type in self._GOUpTypes:
local_down_regulation_list.append(node_bulbs_id)
else:
local_up_regulation_list.append(node_bulbs_id)
self.Reachable_nodes_dict[node_id] = (
list(set(local_uniprot_list)),
list(set(local_regulation_list)),
list(set(local_down_regulation_list)),
list(set(local_up_regulation_list)))
self.All_GOs = list(visited_set)
self.Num2GO = dict((i, val) for i, val in enumerate(self.All_GOs))
self.GO2Num = dict((val, i) for i, val in enumerate(self.All_GOs))
def map_rows_to_names(self):
"""
Maps Node Database IDs, Legacy IDs, display names and types to matrix row/column indexes;
"""
def request_location(_location_buffer_dict, location):
"""
Just a Buffered lookup of location, since the number of cellular location
is relatively small (~80), it makes sense to buffer the IOs on it.
Normally should be moved out as a buffering decorator
:param _location_buffer_dict: Buffered location
:param location: location Node Legacy ID we are willing to verify
:return: displayName of the requested location
"""
location = str(location)
if location in _location_buffer_dict.keys():
return _location_buffer_dict[location]
else:
generator = DatabaseGraph.Location.index.lookup(ID=location)
if generator is not None:
for elt in generator:
_location_buffer_dict[location] = str(elt.displayName)
return str(elt.displayName)
#######################################################################
counter = 0
location_buffer_dict = {}
self.bulbs_id_2_matrix_index = {}
self.matrix_index_2_bulbs_id = {}
log.info('nodes in Highest Set: %s', len(self.Highest_Set))
for bulbs_node_id in self.Highest_Set:
self.bulbs_id_2_matrix_index[bulbs_node_id] = counter
self.matrix_index_2_bulbs_id[counter] = bulbs_node_id
node = DatabaseGraph.vertices.get(bulbs_node_id)
self.bulbs_id_2_display_name[bulbs_node_id] = node.displayName
self.bulbs_id2_node_type[bulbs_node_id] = node.element_type
self.bulbs_id_2_legacy_id[bulbs_node_id] = node.ID
if node.element_type == "UNIPROT":
# TODO: there is a problem: there is no UNIPROT nodes reached during the expansion.
self.reached_uniprots_bulbs_id_list.append(bulbs_node_id)
self.uniprot_matrix_index_list.append(counter)
if node.localization is not None:
self.bulbs_id_2_localization[bulbs_node_id] = request_location(
location_buffer_dict, node.localization)
counter += 1
self.all_uniprots_bulbs_id_list += self.reached_uniprots_bulbs_id_list
self.reached_uniprots_bulbs_id_list = list(set(self.reached_uniprots_bulbs_id_list))
log.info("reached uniprots: %s", len(self.reached_uniprots_bulbs_id_list))
up_generator = stable_get_all(DatabaseGraph.UNIPORT)
if up_generator:
for up_node in up_generator:
bulbs_node_id = get_bulbs_id(up_node)
if bulbs_node_id not in self.reached_uniprots_bulbs_id_list:
self.all_uniprots_bulbs_id_list.append(bulbs_node_id)
self.bulbs_id_2_display_name[bulbs_node_id] = up_node.displayName
self.bulbs_id2_node_type[bulbs_node_id] = up_node.element_type
self.bulbs_id_2_legacy_id[bulbs_node_id] = up_node.ID
self.all_uniprots_bulbs_id_list = list(set(self.all_uniprots_bulbs_id_list))
log.info("analyzable uniprots: %s", len(self.all_uniprots_bulbs_id_list))