Fix gene constraints for RefSeq intergenic vars (#622).

Related-Issue: #622
Closes: #622
Projected-Results-Impact: none

HISTORY.rst

@@ -49,6 +49,7 @@ End-User Summary
 - Adjusting SV filtration presets (#616).
 - Fix bug with thousand genomes frequencies in SV filtration (#619).
 - Displaying disease gene icon also for SVs (#620).
+- Fix bug with gene constraint display for intergenic variants (#620).
 
 Full Change List
 ================
@@ -93,6 +94,7 @@ Full Change List
 - Adjusting SV filtration presets (#616).
 - Fix bug with thousand genomes frequencies in SV filtration (#619).
 - Displaying disease gene icon also for SVs (#620).
+- Fix bug with gene constraint display for intergenic variants (#620).
 
 ------
 v1.2.0

svs/templates/svs/_filter_form.html

@@ -23,8 +23,8 @@
         </a>
 
         <div class="dropdown-menu" aria-labelledby="dropdownFreqs">
-          <a class="dropdown-item" data-preset="freq-in_house_only">in-house only (default)</a>
-          <a class="dropdown-item" data-preset="freq-relaxed">relaxed</a>
+          <a class="dropdown-item" data-preset="freq-in_house_only">in-house only</a>
+          <a class="dropdown-item" data-preset="freq-relaxed">relaxed (default)</a>
           <a class="dropdown-item" data-preset="freq-strict">strict</a>
         </div>
       </div>

variants/queries.py

@@ -28,6 +28,7 @@
     ExacConstraints,
     MgiMapping,
     RefseqToGeneSymbol,
+    RefseqToEnsembl,
     EnsemblToGeneSymbol,
     GeneIdInHpo,
 )
@@ -1360,49 +1361,122 @@ def __init__(self, *args, **kwargs):
             "oe_lof_upper",
             "oe_lof_lower",
         ]
-        self.subquery = (
-            select(
-                [
-                    func.max(getattr(GnomadConstraints.sa, field)).label(field)
-                    for field in self.fields
-                ]
-                + [func.max(GnomadConstraints.sa.oe_lof_upper + 0.001).label("loeuf")]
-            )
-            .select_from(GnomadConstraints.sa)
-            .where(SmallVariant.sa.ensembl_gene_id == GnomadConstraints.sa.ensembl_gene_id)
-            .group_by(GnomadConstraints.sa.ensembl_gene_id)
-            .lateral("gnomad_constraints_subquery")
-        )
+        self.subquery = self._build_subquery()
         self.fields.append("loeuf")
 
+    def _build_subquery(self):
+        """Build sub query, depending on selected database (refseq/ensembl)."""
+        if self.kwargs["database_select"] == "ensembl":
+            return (
+                select(
+                    [
+                        func.max(getattr(GnomadConstraints.sa, field)).label(field)
+                        for field in self.fields
+                    ]
+                    + [func.max(GnomadConstraints.sa.oe_lof_upper + 0.001).label("loeuf")]
+                )
+                .select_from(GnomadConstraints.sa)
+                .where(SmallVariant.sa.ensembl_gene_id == GnomadConstraints.sa.ensembl_gene_id)
+                .group_by(GnomadConstraints.sa.ensembl_gene_id)
+                .lateral("gnomad_constraints_subquery")
+            )
+        else:
+            self.subquery_refseq_to_ensembl = (
+                select([func.max(RefseqToEnsembl.sa.ensembl_gene_id).label("ensembl_gene_id")])
+                .select_from(RefseqToEnsembl.sa)
+                .where(SmallVariant.sa.refseq_gene_id == RefseqToEnsembl.sa.entrez_id)
+                .group_by(RefseqToEnsembl.sa.entrez_id)
+                .lateral("refseqtoensembl_subquery_gnomad_constraints")
+            )
+            link = (
+                self.subquery_refseq_to_ensembl.c.ensembl_gene_id
+                == GnomadConstraints.sa.ensembl_gene_id
+            )
+            return (
+                select(
+                    [
+                        func.max(getattr(GnomadConstraints.sa, field)).label(field)
+                        for field in self.fields
+                    ]
+                    + [func.max(GnomadConstraints.sa.oe_lof_upper + 0.001).label("loeuf")]
+                )
+                .select_from(GnomadConstraints.sa)
+                .where(link)
+                .group_by(GnomadConstraints.sa.ensembl_gene_id)
+                .lateral("gnomad_constraints_subquery")
+            )
+
     def extend_fields(self, _query_parts):
         return [getattr(self.subquery.c, field).label("gnomad_%s" % field) for field in self.fields]
 
     def extend_selectable(self, query_parts):
+        if self.kwargs["database_select"] == "refseq":
+            query_parts = query_parts.selectable.outerjoin(self.subquery_refseq_to_ensembl, true())
         return query_parts.selectable.outerjoin(self.subquery, true())
 
 
 class ExtendQueryPartsExacConstraintsJoin(ExtendQueryPartsBase):
     def __init__(self, *args, **kwargs):
         super().__init__(*args, **kwargs)
         self.fields = ["pLI", "mis_z", "syn_z"]
-        self.subquery = (
-            select(
-                [func.max(getattr(ExacConstraints.sa, field)).label(field) for field in self.fields]
+        self.subquery = self._build_subquery()
+        self.subquery = self._build_subquery()
+
+    def _build_subquery(self):
+        """Build sub query, depending on selected database (refseq/ensembl)."""
+        if self.kwargs["database_select"] == "ensembl":
+            return (
+                select(
+                    [
+                        func.max(getattr(ExacConstraints.sa, field)).label(field)
+                        for field in self.fields
+                    ]
+                )
+                .select_from(ExacConstraints.sa)
+                .where(
+                    func.split_part(SmallVariant.sa.ensembl_transcript_id, ".", 1)
+                    == ExacConstraints.sa.ensembl_transcript_id
+                )
+                .group_by(ExacConstraints.sa.ensembl_transcript_id)
+                .lateral("exac_constraints_subquery")
             )
-            .select_from(ExacConstraints.sa)
-            .where(
-                func.split_part(SmallVariant.sa.ensembl_transcript_id, ".", 1)
+        else:
+            self.subquery_refseq_to_ensembl = (
+                select(
+                    [
+                        func.max(RefseqToEnsembl.sa.ensembl_transcript_id).label(
+                            "ensembl_transcript_id"
+                        )
+                    ]
+                )
+                .select_from(RefseqToEnsembl.sa)
+                .where(SmallVariant.sa.refseq_gene_id == RefseqToEnsembl.sa.entrez_id)
+                .group_by(RefseqToEnsembl.sa.entrez_id)
+                .lateral("refseqtoensembl_subquery_exac_constraints")
+            )
+            link = (
+                self.subquery_refseq_to_ensembl.c.ensembl_transcript_id
                 == ExacConstraints.sa.ensembl_transcript_id
             )
-            .group_by(ExacConstraints.sa.ensembl_transcript_id)
-            .lateral("exac_constraints_subquery")
-        )
+            return (
+                select(
+                    [
+                        func.max(getattr(ExacConstraints.sa, field)).label(field)
+                        for field in self.fields
+                    ]
+                )
+                .select_from(ExacConstraints.sa)
+                .where(link)
+                .group_by(ExacConstraints.sa.ensembl_transcript_id)
+                .lateral("exac_constraints_subquery")
+            )
 
     def extend_fields(self, _query_parts):
         return [getattr(self.subquery.c, field).label("exac_%s" % field) for field in self.fields]
 
     def extend_selectable(self, query_parts):
+        if self.kwargs["database_select"] == "refseq":
+            query_parts = query_parts.selectable.outerjoin(self.subquery_refseq_to_ensembl, true())
         return query_parts.selectable.outerjoin(self.subquery, true())
 
 

variants/tests/test_file_export.py

@@ -16,7 +16,7 @@
 from clinvar.tests.factories import ClinvarFactory
 from cohorts.tests.factories import TestCohortBase
 from extra_annos.tests.factories import ExtraAnnoFieldFactory, ExtraAnnoFactory
-from geneinfo.tests.factories import GnomadConstraintsFactory
+from geneinfo.tests.factories import GnomadConstraintsFactory, RefseqToEnsemblFactory
 from variants.tests.factories import (
     SmallVariantFactory,
     ResubmitFormDataFactory,
@@ -93,6 +93,11 @@ def setUp(self):
                 pathogenicity_summary="uncertain significance",
             )
             GnomadConstraintsFactory(ensembl_gene_id=small_var.ensembl_gene_id)
+            RefseqToEnsemblFactory(
+                entrez_id=small_var.refseq_gene_id,
+                ensembl_gene_id=small_var.ensembl_gene_id,
+                ensembl_transcript_id=small_var.ensembl_transcript_id,
+            )
 
     def _set_janno_mocker(self, database, mock_):
         if database == "refseq":

variants/tests/test_queries.py

@@ -35,6 +35,7 @@
     MgiMappingFactory,
     RefseqToGeneSymbolFactory,
     EnsemblToGeneSymbolFactory,
+    RefseqToEnsemblFactory,
 )
 from dbsnp.tests.factories import DbsnpFactory
 from .factories import (
@@ -112,6 +113,11 @@ def setUp(self):
             project=case.project
         )
         self.projectcasessmallvariantquery.query_results.add(small_vars[0].id, small_vars[2].id)
+        RefseqToEnsemblFactory(
+            entrez_id=small_vars[0].refseq_gene_id,
+            ensembl_gene_id=small_vars[0].ensembl_gene_id,
+            ensembl_transcript_id=small_vars[0].ensembl_transcript_id,
+        )
 
     def test_load_prefetched_case_results(self):
         results = self.run_query(
@@ -150,6 +156,88 @@ def test_load_prefetched_project_cases_results(self):
         self.assertIsNone(results[1].mgi_id)
 
 
+class TestCaseRefSeqIntergenicPLI(SupportQueryTestBase):
+    """Test the case that the entrez ID does not correspond to the ensembl ID
+
+    cf. bug in https://github.com/bihealth/varfish-server/issues/622
+    """
+
+    def setUp(self):
+        super().setUp()
+        case, variant_set, _ = CaseWithVariantSetFactory.get("small")
+        self.acmg_0 = AcmgFactory(entrez_id="1000", ensembl_gene_id="ENSG1001")
+        self.acmg_1 = AcmgFactory(entrez_id="1000", ensembl_gene_id="ENSG1000")
+        small_vars = [
+            SmallVariantFactory(
+                chromosome=normalize_chrom("1", case.release),
+                ensembl_effect=["missense_variant"],
+                refseq_effect=["intergenic_variant"],
+                refseq_gene_id=self.acmg_1.entrez_id,
+                ensembl_gene_id=self.acmg_0.ensembl_gene_id,
+                variant_set=variant_set,
+            ),
+            SmallVariantFactory(
+                chromosome=normalize_chrom("1", case.release),
+                ensembl_effect=["missense_variant"],
+                refseq_effect=["missense_variant_variant"],
+                refseq_gene_id=self.acmg_1.entrez_id,
+                ensembl_gene_id=self.acmg_1.ensembl_gene_id,
+                variant_set=variant_set,
+            ),
+        ]
+        # Prepare constraints
+        self.gnomad_constraints = [
+            GnomadConstraintsFactory(ensembl_gene_id=small_vars[0].ensembl_gene_id, pLI=0.8,),
+            GnomadConstraintsFactory(ensembl_gene_id=small_vars[1].ensembl_gene_id, pLI=0.4,),
+        ]
+        self.exac_constraints = [
+            ExacConstraintsFactory(
+                ensembl_transcript_id=small_vars[0].ensembl_transcript_id, pLI=1.0,
+            ),
+            ExacConstraintsFactory(
+                ensembl_transcript_id=small_vars[1].ensembl_transcript_id, pLI=0.5,
+            ),
+        ]
+        # Prepare smallvariant query results
+        self.smallvariantquery = SmallVariantQueryFactory(case=case)
+        self.smallvariantquery.query_results.add(small_vars[0].id, small_vars[1].id)
+        # Prepare projectcases smallvariant query results
+        self.projectcasessmallvariantquery = ProjectCasesSmallVariantQueryFactory(
+            project=case.project
+        )
+        self.projectcasessmallvariantquery.query_results.add(small_vars[0].id, small_vars[1].id)
+        # Create appropriate refseq to ensembl mappings
+        RefseqToEnsemblFactory(
+            entrez_id="1000",
+            ensembl_gene_id="ENSG1000",
+            ensembl_transcript_id=small_vars[1].ensembl_transcript_id,
+        )
+
+    def test_run_query_refseq(self):
+        results = self.run_query(
+            ProjectLoadPrefetchedQuery,
+            {"filter_job_id": self.projectcasessmallvariantquery.id, "database_select": "refseq"},
+            2,
+            query_type="project",
+        )
+        self.assertEqual(results[0].exac_pLI, 0.5)
+        self.assertEqual(results[1].exac_pLI, 0.5)
+        self.assertEqual(results[0].gnomad_pLI, 0.4)
+        self.assertEqual(results[1].gnomad_pLI, 0.4)
+
+    def test_run_query_ensembl(self):
+        results = self.run_query(
+            ProjectLoadPrefetchedQuery,
+            {"filter_job_id": self.projectcasessmallvariantquery.id, "database_select": "ensembl"},
+            2,
+            query_type="project",
+        )
+        self.assertEqual(results[0].exac_pLI, 1.0)
+        self.assertEqual(results[1].exac_pLI, 0.5)
+        self.assertEqual(results[0].gnomad_pLI, 0.8)
+        self.assertEqual(results[1].gnomad_pLI, 0.4)
+
+
 class TestCaseLoadPrefetchedSorting(SupportQueryTestBase):
     def setUp(self):
         super().setUp()