Allow choosing preferred mappings

indra/databases/hgnc_client.py

@@ -422,6 +422,10 @@ def get_hgnc_name_from_mgi_name(mgi_name: str) -> Union[str, None]:
 def _read_hgnc_maps():
     hgnc_file = get_resource_path("hgnc_entries.tsv")
     csv_rows = read_unicode_csv(hgnc_file, delimiter='\t', encoding='utf-8')
+    hgnc_uniprot_preferred = get_resource_path("hgnc_uniprot_preferred.csv")
+    csv_rows_uniprot_preferred = \
+        read_unicode_csv(hgnc_uniprot_preferred, delimiter=',',
+                         encoding='utf-8')
     hgnc_names = {}
     hgnc_ids = {}
     hgnc_withdrawn = []
@@ -515,19 +519,26 @@ def _read_hgnc_maps():
     for old_id, new_id in hgnc_withdrawn_new_ids.items():
         hgnc_names[old_id] = hgnc_names[new_id]
 
+    uniprot_ids_preferred = {}
+    for row in csv_rows_uniprot_preferred:
+        hgnc_id = row[0]
+        uniprot_id = row[1]
+        uniprot_ids_preferred[hgnc_id] = uniprot_id
+
     return (
         hgnc_names, hgnc_ids, hgnc_withdrawn,
         uniprot_ids, entrez_ids, entrez_ids_reverse, mouse_map, rat_map,
         prev_sym_map, ensembl_ids, ensembl_ids_reverse, gene_types,
         dict(hgnc_to_enzymes), dict(enzyme_to_hgncs),
+        uniprot_ids_preferred
     )
 
 
 (
     hgnc_names, hgnc_ids, hgnc_withdrawn, uniprot_ids, entrez_ids,
     entrez_ids_reverse, mouse_map, rat_map, prev_sym_map, ensembl_ids,
     ensembl_ids_reverse, gene_type,
-    hgnc_to_enzymes, enzyme_to_hgncs,
+    hgnc_to_enzymes, enzyme_to_hgncs, uniprot_ids_preferred
 ) = _read_hgnc_maps()
 
 

indra/ontology/bio/ontology.py

@@ -26,7 +26,7 @@ class BioOntology(IndraOntology):
     # should be incremented to "force" rebuilding the ontology to be consistent
     # with the underlying resource files.
     name = 'bio'
-    version = '1.33'
+    version = '1.34'
     ontology_namespaces = [
         'go', 'efo', 'hp', 'doid', 'chebi', 'ido', 'mondo', 'eccode',
     ]
@@ -147,11 +147,15 @@ def add_hgnc_uniprot_entrez_xrefs(self):
         from indra.databases import hgnc_client
         from indra.databases import uniprot_client
         edges = []
-        for hid, uid in hgnc_client.uniprot_ids.items():
-            uids = uid.split(', ')
+        for hid, upid in hgnc_client.uniprot_ids.items():
+            uids = upid.split(', ')
+            preferred = hgnc_client.uniprot_ids_preferred.get(hid)
+            if preferred:
+                uids = [preferred]
             for uid in uids:
+                edge_data = {'type': 'xref', 'source': 'hgnc'}
                 edges.append((self.label('HGNC', hid), self.label('UP', uid),
-                              {'type': 'xref', 'source': 'hgnc'}))
+                              edge_data))
         self.add_edges_from(edges)
 
         edges = [(self.label('UP', uid), self.label('HGNC', hid),