Merge pull request #2 from bcgsc/bugfix/DEVSU-2348-add-gene-assoc-to-…

…pharmacogenelist

Bugfix/devsu 2348 add gene assoc to pharmacogenelist

pori_python/graphkb/genes.py

@@ -226,10 +226,19 @@ def get_preferred_gene_name(
     return gene_names[0]
 
 
-# DEVSU-2348 - relate the genes to the variants
 def get_cancer_predisposition_info(
     conn: GraphKBConnection, source: str = PREFERRED_GENE_SOURCE
 ) -> Tuple[List[str], Dict[str, str]]:
+    newval = get_gene_linked_cancer_predisposition_info(conn, source)
+    genes = newval[0]
+    allvardata = newval[1]
+    variants = {key: allvardata[key][0] for key in allvardata.keys()}
+    return newval[0], variants
+
+
+def get_gene_linked_cancer_predisposition_info(
+    conn: GraphKBConnection, source: str = PREFERRED_GENE_SOURCE
+) -> Tuple[List[str], Dict[str, Tuple[str, List[str]]]]:
     """
     Return two lists from GraphKB, one of cancer predisposition genes and one of associated variants.
 
@@ -242,6 +251,8 @@ def get_cancer_predisposition_info(
 
     Example: https://graphkb.bcgsc.ca/view/Statement/155:11616
 
+
+
     Returns:
         genes: list of cancer predisposition genes
         variants: dictionary mapping pharmacogenomic variant IDs to variant display names
@@ -284,21 +295,24 @@ def get_cancer_predisposition_info(
     ):
         for condition in record["conditions"]:  # type: ignore
             if condition["@class"] == "PositionalVariant":
-                variants[condition["@rid"]] = condition["displayName"]
+                assoc_gene_list = []
                 for reference in ["reference1", "reference2"]:
                     name = (condition.get(reference) or {}).get("displayName", "")
                     biotype = (condition.get(reference) or {}).get("biotype", "")
                     if name and biotype == "gene":
                         genes.add(name)
+                        assoc_gene_list.append(name)
                     elif name:
                         gene = get_preferred_gene_name(conn, name, source)
                         if gene:
                             infer_genes.add((gene, name, biotype))
+                            assoc_gene_list.append(gene)
                         else:
                             non_genes.add((name, biotype))
                             logger.error(
                                 f"Non-gene cancer predisposition {biotype}: {name} for {condition['displayName']}"
                             )
+                variants[condition["@rid"]] = [condition["displayName"], assoc_gene_list]
 
     for gene, name, biotype in infer_genes:
         logger.debug(f"Found gene '{gene}' for '{name}' ({biotype})")
@@ -310,10 +324,19 @@ def get_cancer_predisposition_info(
     return sorted(genes), variants
 
 
-# DEVSU-2348 - relate the genes to the variants
 def get_pharmacogenomic_info(
     conn: GraphKBConnection, source: str = PREFERRED_GENE_SOURCE
 ) -> Tuple[List[str], Dict[str, str]]:
+    newval = get_gene_linked_pharmacogenomic_info(conn, source)
+    genes = newval[0]
+    allvardata = newval[1]
+    variants = {key: allvardata[key][0] for key in allvardata.keys()}
+    return newval[0], variants
+
+
+def get_gene_linked_pharmacogenomic_info(
+    conn: GraphKBConnection, source: str = PREFERRED_GENE_SOURCE
+) -> Tuple[List[str], Dict[str, Tuple[str, List[str]]]]:
     """
     Return two lists from GraphKB, one of pharmacogenomic genes and one of associated variants.
 
@@ -362,22 +385,24 @@ def get_pharmacogenomic_info(
 
         for condition in record["conditions"]:  # type: ignore
             if condition["@class"] == "PositionalVariant":
-                variants[condition["@rid"]] = condition["displayName"]
+                assoc_gene_list = []
                 for reference in ["reference1", "reference2"]:
                     name = (condition.get(reference) or {}).get("displayName", "")
                     biotype = (condition.get(reference) or {}).get("biotype", "")
                     if name and biotype == "gene":
                         genes.add(name)
+                        assoc_gene_list.append(name)
                     elif name:
                         gene = get_preferred_gene_name(conn, name, source)
                         if gene:
                             infer_genes.add((gene, name, biotype))
+                            assoc_gene_list.append(gene)
                         else:
                             non_genes.add((name, biotype))
                             logger.error(
                                 f"Non-gene pharmacogenomic {biotype}: {name} for {condition['displayName']}"
                             )
-
+                variants[condition["@rid"]] = [condition["displayName"], assoc_gene_list]
     for gene, name, biotype in infer_genes:
         logger.debug(f"Found gene '{gene}' for '{name}' ({biotype})")
         genes.add(gene)

tests/test_graphkb/test_genes.py

@@ -10,11 +10,13 @@
 from pori_python.graphkb.genes import (
     get_cancer_genes,
     get_cancer_predisposition_info,
+    get_gene_linked_cancer_predisposition_info,
     get_gene_information,
     get_genes_from_variant_types,
     get_oncokb_oncogenes,
     get_oncokb_tumour_supressors,
     get_pharmacogenomic_info,
+    get_gene_linked_pharmacogenomic_info,
     get_preferred_gene_name,
     get_therapeutic_associated_genes,
 )
@@ -141,14 +143,28 @@ def test_cancer_genes(conn):
         assert gene not in names
 
 
-@pytest.mark.skip(reason="DEVSU-2348")
 def test_get_pharmacogenomic_info(conn):
     genes, matches = get_pharmacogenomic_info(conn)
     for gene in PHARMACOGENOMIC_INITIAL_GENES:
         assert gene in genes, f"{gene} not found in get_pharmacogenomic_info"
         for rid, variant_display in matches.items():
             if variant_display.startswith(gene):
                 break
+        else:  # no break called
+            # failing on this version of the func; addressed in 'new' version
+            if gene == 'ACYP2':
+                continue
+            assert False, f"No rid found for a pharmacogenomic with {gene}"
+
+
+def test_get_gene_linked_pharmacogenomic_info(conn):
+    genes, matches = get_gene_linked_pharmacogenomic_info(conn)
+    for gene in PHARMACOGENOMIC_INITIAL_GENES:
+        assert gene in genes, f"{gene} not found in get_pharmacogenomic_info"
+        for rid, variant_info in matches.items():
+            variant_gene_assoc = variant_info[1]
+            if gene in variant_gene_assoc:
+                break
         else:  # no break called
             assert False, f"No rid found for a pharmacogenomic with {gene}"
 
@@ -160,6 +176,13 @@ def test_get_cancer_predisposition_info(conn):
         assert gene in genes, f"{gene} not found in get_cancer_predisposition_info"
 
 
+@pytest.mark.skipif(EXCLUDE_INTEGRATION_TESTS, reason="excluding long running integration tests")
+def test_get_gene_linked_cancer_predisposition_info(conn):
+    genes, matches = get_gene_linked_cancer_predisposition_info(conn)
+    for gene in CANCER_PREDISP_INITIAL_GENES:
+        assert gene in genes, f"{gene} not found in get_cancer_predisposition_info"
+
+
 @pytest.mark.parametrize(
     "alt_rep", ("NM_033360.4", "NM_033360", "ENSG00000133703.11", "ENSG00000133703")
 )