qurying opentargets for variatns; fix issue with possibly missing gene name for a variant in dbsnp

associations/minerva_variants.py

@@ -31,6 +31,12 @@ def get_dbsnp(ids: List[str]) -> List[Dict]:
             if "dbsnp" not in res:
                 continue
 
+            if "gene" not in res["dbsnp"]:
+                # This can happen for variants which are not mapped onto a gene in dbSNP but for example OpenTargets can
+                # obtain this information from a difference source. Example would be rs9388451 which is not associated
+                # with a gene in dbSNP but it is in GWAS Catalog (or at least this was the situation as of 2019-12-02)
+                continue
+
             genes = res["dbsnp"]["gene"]
             if not isinstance(genes, list):
                 genes = [genes]

associations/opentargets.py

@@ -29,14 +29,14 @@ def get_genes_and_evidence(orphanet_ids:List[str], association_score: float) ->
 
     return genes
 
-def get_variants(genes_evidence: Dict):
-    for ge in genes_evidence:
-        s_ge = ge['evidence_id'].split('-')
+def get_variants(genes: Dict[str, Dict]):
+    for gene_name in genes:
+        s_ge = genes[gene_name]['evidence_id'].split('-')
         target = s_ge[0]
         disease = s_ge[1]
 
-        res = requests.os(
-            "https://platform-api.opentargets.io/v3/platform/public/association/filter",
+        res = requests.get(
+            "https://platform-api-qc.opentargets.io/v3/platform/public/evidence/filter",
             params={'target': target,
                     'disease': disease,
                     'data_type': 'genetic_association',
@@ -45,6 +45,15 @@ def get_variants(genes_evidence: Dict):
         )
         if res.status_code == requests.codes.ok:
             content = json.loads(res.text)
+            if 'data' in content:
+                for content_item in content['data']:
+                    if 'variant' in content_item:
+                        if 'dbsnp' in content_item['variant']['id']:
+                            rsId = content_item['variant']['id'].split("/")[-1]
+                            genes[gene_name]['variants'].append(rsId)
+
+        # Not sure whether we can't get duplicities from the API so we better uniquify
+        genes[gene_name]['variants'] = list(set(genes[gene_name]['variants']))
 
 
 if __name__ == '__main__':
@@ -71,7 +80,6 @@ if __name__ == '__main__':
     args = parser.parse_args()
 
     genes_evidence = get_genes_and_evidence(args.orphanet_ids.split(","), args.threshold_score)
-    # get_variants(genes_evidence)
     genes_sorted = {}
     cnt = args.top_n
     i = 0
@@ -81,6 +89,8 @@ if __name__ == '__main__':
         i += 1
         genes_sorted[k] = v
 
+    get_variants(genes_sorted)
+
     print(json.dumps({"name": "opentargets", "genes": genes_sorted}, indent=2))