def GOQLtoFindDrugs(TargetIds: list,
TargetType='Protein',
drugEffect=['negative']):
if TargetType == 'Protein':
return OQL.get_drugs(for_targets_with_ids=TargetIds)
elif TargetType == 'Small Molecule':
REL_TYPES = ['Regulation', 'MolSynthesis']
OQLquery = OQL.expand_entity(PropertyValues=TargetIds,
SearchByProperties=['id'],
expand_by_rel_types=REL_TYPES,
expand2neighbors=['Small Molecule'],
direction='upstream')
OQLquery += ' AND Effect = (' + ','.join(drugEffect) + ')'
return OQLquery
else:
REL_TYPES = ['Regulation']
OQLquery = OQL.expand_entity(PropertyValues=TargetIds,
SearchByProperties=['id'],
expand_by_rel_types=REL_TYPES,
expand2neighbors=['Small Molecule'],
direction='upstream')
OQLquery += ' AND Effect = (' + ','.join(drugEffect) + ')'
return OQLquery
def retreive_clinical_trials(drugs_name_file=None):
request_name = 'Find all completed clinical trials'
oql_query = 'SELECT Relation WHERE objectType = ClinicalTrial AND TrialStatus = Completed AND NeighborOf (SELECT Entity WHERE objectType = SmallMol)'
if isinstance(drugs_name_file, str):
with open(drugs_name_file) as f:
drug_names = [line.rstrip('\n') for line in f]
print('Read %s with %d drug names' %
(drugs_name_file, len(drug_names)))
search_by_prop, drug_names_str = OQL.get_search_strings(
['Name', 'Alias'], drug_names)
oql_query = oql_query + ' AND NeighborOf (SELECT Entity WHERE (Name,Alias)= ({names}))'
oql_query = oql_query.format(names=drug_names_str)
request_name = 'Find clinical trials for drugs in {fname}'
request_name = request_name.format(fname=drugs_name_file)
return ps_api.process_oql(oql_query, request_name, debug=False)
args.infile_has_header,
use_cache=False,
map2type=entity_types)
if len(args.pathways) > 0:
print("Begin linking entities mapped from infile to pathways")
LinkToPathways = str(args.pathways).split(",")
start_time = time.time()
for PathwayName in LinkToPathways:
PathwayMembersId2Entity = search.get_pathway_member_ids(
[PathwayName],
search_pathways_by=['Name'],
only_entities=['Protein', 'FunctionalClass', 'Complex'],
with_properties=['objectType'])
pathway_components = set(PathwayMembersId2Entity.keys())
QueryOntology = OQL.get_childs(list(pathway_components), ['id'])
pathway_components.update(search._obj_id_by_oql(QueryOntology))
if len(pathway_components) == 0:
print('No entity for %s found in the database' % PathwayName)
else:
search.link2concept(PathwayName, list(pathway_components))
exec_time = search.execution_time(start_time)
print("Entities in file %s were linked to %s pathway in %s" %
(EntityListFile, PathwayName, exec_time))
else:
print(
'No pathways were specified for semantic linking with entities from \"%s\"'
% (EntityListFile))
REL_PROPs = ['Name', 'Effect', 'Mechanism', 'ChangeType']# add here relation properties to retrieve
# if properties from NetworkxObjects.REF_ID_TYPES or NetworkxObjects.REF_PROPS are added to REL_PROPs then:
# output size may increase dramatically because it will contain one reference per row.
ENT_PROPs = ['Name', 'Description', 'Cell Localization']
ps_api = ps_api = open_api_session()
ps_api.PageSize = 10000
ps_api.add_rel_props(list(set(REL_PROPs)|PS_ID_TYPES))
ps_api.add_ent_props(ENT_PROPs)
# this dump file will list all proteins in the database with connectivity >0:
ps_api.add_dump_file('Proteins from database.tsv', replace_main_dump=True)
print('Fetching all proteins from the database')
ProteinsOnyGraph = ps_api.process_oql("Select Entity WHERE objectType = Protein AND Connectivity > 0 AND Name LIKE 'A%'", flush_dump=True)
ps_api.add_dump_file("Protein neighbors dump.tsv", replace_main_dump=True) # dump file accumulates all data in one big file
out_dir = 'csv'
counter = 0
for node_id, psObj in ProteinsOnyGraph.nodes(data=True):
protein_name = psObj['Name'][0]
counter += 1
print('Finding neighbors for \"%s\", node #%d from %d total' %
(protein_name, counter, ProteinsOnyGraph.number_of_nodes()))
oql_query = GOQL.expand_entity([node_id], SearchByProperties=['id'])
ProteinNeighborsGraph = ps_api.process_oql(oql_query)
protein_neighbors_file = out_dir + '/' + protein_name + '_neighbors.csv'
ps_api.to_csv(protein_neighbors_file)
ps_api.Graph.clear() # need to release memory when performing large dumps
import ElsevierAPI.ResnetAPI.PathwayStudioGOQL as OQL
from ElsevierAPI.ResnetAPI.NetworkxObjects import Reference
from ElsevierAPI import load_api_config
from ElsevierAPI.ResnetAPI.ResnetAPISession import APISession
import pandas as pd
APIconfig = load_api_config()
ps_api = APISession(APIconfig['ResnetURL'], APIconfig['PSuserName'],
APIconfig['PSpassword'])
fileIn = 'Drugs for Regulators in 4 patients.txt'
InDir = 'D:\\Python\\PBTA\\PNOC003\\4 patients analysis\\'
with open(InDir + fileIn) as f:
drugs = [line.rstrip('\n') for line in f]
print('Finding drugs in %s in Resnet' % (fileIn))
OQLquery = OQL.get_entities_by_props(drugs, ['Name', 'Alias'],
only_object_types=['Small Molecule'])
ps_api.add_ent_props(['Name', 'PharmaPendium ID'])
resnet_drugs = ps_api.process_oql(OQLquery, 'Find all drugs')
print('Found %d drugs in Resnet' % len(resnet_drugs))
#removing duplicates wiht no PharmaPendium ID
resnet2pharmapendium_map = dict()
for i, drug in resnet_drugs.nodes(data=True):
try:
resnet2pharmapendium_map[str(
drug['Name'][0]).lower()] = drug['PharmaPendium ID'][0]
except KeyError:
continue
all_drugs = list(resnet_drugs.nodes(data=True))
for i, drug in all_drugs:
api_config
) # specify here path to your APIconfig file. Defaults to ./ElsevierAPI/APIconfig.json
ps_api.add_ent_props(
['Alias']
) # need to retreive aliases from the database in case input metabolites are found by Alias
ps_api.PageSize = 10000
# dump file contains references for all relations retreived from database
# do not use dump file unless you need to include references into report:
ps_api.DumpFiles.clear()
# retreive all ChemicalReaction linked to metabolites in excel_file_name as ResnetGraph from the database:
step = 1000
for i in range(0, len(input_metabolite_names), step):
name_list = input_metabolite_names[i:i + step]
my_goql_query = OQL.expand_entity(name_list, ['Name', 'Alias'],
expand_by_rel_types=['ChemicalReaction'])
request_name = 'Retrieve metabolic reactions graph for {count} metabolites'.format(
count=len(name_list))
ps_api.process_oql(my_goql_query, request_name)
reactions_graph = ps_api.Graph
input_name2objs, objid2input_names = reactions_graph.get_prop2obj_dic(
'Name', input_metabolite_names)
aliasinput_2objs, objid2input_alias = reactions_graph.get_prop2obj_dic(
'Alias', input_metabolite_names)
objid2input_names.update(objid2input_alias)
metabolite_ids = list(objid2input_names.keys())
# objid2input_names = {obj_id:[input_names]} - allows for duplicates when mapping by name+alias
# find enzymes linked to ChemicalReactions and retreive their ontology children (proteins)
enzymes = reactions_graph.get_objects(PROTEIN_TYPES)
InputDiseaseNames = ','.join(SearchEntitiesBy)
# specify files used in ps_api.DiseaseNetwork.AddGraph to dump graph data in tab-delimited format:
myDir = '' # 'D:\\Python\\PS_API\\'
foutDiseaseSNPs = myDir + "Gene variants linked to " + InputDiseaseNames + '.tsv'
foutDiseaseProteins = myDir + "Genes with SNPs linked to " + InputDiseaseNames + '.tsv'
foutDrugsForDiseaseProteins = myDir + "Druggable targets for " + InputDiseaseNames + '.tsv'
ps_api.add_rel_props(REL_PROPs)
ps_api.add_ent_props(ENT_PROP_Neo4j)
print("Finding GeneticVariants linked to %s" % InputDiseaseNames)
ps_api.add_dump_file(foutDiseaseSNPs, replace_main_dump=True)
ps_api.process_oql(
GOQL.expand_entity(PropertyValues=SearchEntitiesBy,
SearchByProperties=['Name', 'Alias'],
expand_by_rel_types=[],
expand2neighbors=['GeneticVariant']))
SNPIds = list(set(ps_api.Graph.get_entity_ids(['GeneticVariant'])))
print("Finding Proteins containing GeneticVariants linked to %s" %
InputDiseaseNames)
ps_api.add_dump_file(foutDiseaseProteins, replace_main_dump=True)
ps_api.process_oql(GOQL.expand_entity(PropertyValues=SNPIds,
SearchByProperties=['id'],
expand_by_rel_types=['GeneticChange'],
expand2neighbors=['Protein']),
flush_dump=True)
foutDiseasePPI = myDir + "\\PPIs between genes linked to " + InputDiseaseNames + '.tsv'
PPIgraph = ps_api.get_ppi_graph(foutDiseasePPI)
# calculating centrality