Merge pull request #117 from bcgsc/task/KBDEV-1163-tests-ensembl-protein

Task/kbdev 1163 tests ensembl protein
bcgsc · Mar 2, 2024 · 73473c6 · 73473c6
2 parents 15180c6 + 16195a0
commit 73473c6
Show file tree

Hide file tree

Showing 6 changed files with 143 additions and 7 deletions.
diff --git a/graphkb/genes.py b/graphkb/genes.py
@@ -1,4 +1,5 @@
 """Methods for retrieving gene annotation lists from GraphKB."""
+
 from typing import Any, Dict, List, Sequence, Set, Tuple, cast
 
 from . import GraphKBConnection

diff --git a/graphkb/match.py b/graphkb/match.py
@@ -1,6 +1,7 @@
 """
 Functions which return Variants from GraphKB which match some input variant definition
 """
+
 from typing import Dict, List, Optional, Set, Union, cast
 
 from . import GraphKBConnection
@@ -303,9 +304,11 @@ def compare_positional_variants(
     if not positions_overlap(
         cast(BasicPosition, variant["break1Start"]),
         cast(BasicPosition, reference_variant["break1Start"]),
-        None
-        if "break1End" not in reference_variant
-        else cast(BasicPosition, reference_variant["break1End"]),
+        (
+            None
+            if "break1End" not in reference_variant
+            else cast(BasicPosition, reference_variant["break1End"])
+        ),
     ):
         return False
 
@@ -318,9 +321,11 @@ def compare_positional_variants(
         if not positions_overlap(
             cast(BasicPosition, variant["break2Start"]),
             cast(BasicPosition, reference_variant["break2Start"]),
-            None
-            if "break2End" not in reference_variant
-            else cast(BasicPosition, reference_variant["break2End"]),
+            (
+                None
+                if "break2End" not in reference_variant
+                else cast(BasicPosition, reference_variant["break2End"])
+            ),
         ):
             return False
 

diff --git a/tests/data.py b/tests/data.py
@@ -5,7 +5,6 @@
         Array of variants (diplayName and type) that MUST NOT be matching, but not restricted to
 """
 
-
 # Screening structural variant to rule out small events [KBDEV_1056]
 structuralVariants = {
     # Unambiguous structural variations

diff --git a/tests/test_genes.py b/tests/test_genes.py
@@ -1,6 +1,7 @@
 """
 Tests here depend on specific data in GraphKB which can change. To avoid this, expected/stable values are chosen
 """
+
 import os
 
 import pytest

diff --git a/tests/test_match.py b/tests/test_match.py
@@ -79,6 +79,135 @@ def test_checks_by_source_id_kras(self, conn):
         ]
         assert "KRAS" in kras
 
+    # KBDEV-1163
+    # Testing if the addition of Ensembl protein Features are limiting results
+    # returned by get_equivalent_features() since SimilatTo queryType queries
+    # aren't traversing the graph to it's whole depth.
+    @pytest.mark.skipif(EXCLUDE_INTEGRATION_TESTS, reason="excluding data-specific test")
+    def test_ensembl_protein(self, conn):
+        for feature, expected in [
+            # Sample
+            (
+                'EGFR',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'ENST00000275493',
+                    'ENST00000275493.6',
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',
+                    'NP_001333826.1',
+                ],
+            ),
+            (
+                'NM_001346897',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',
+                    'NP_001333826.1',
+                ],
+            ),
+            (
+                'NM_001346897.2',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',
+                    'NP_001333826.1',
+                ],
+            ),
+            (
+                'NP_001333826',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',  # Warn: Versionized ENSG won't be returned due to API limitations
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',
+                    'NP_001333826.1',
+                ],
+            ),
+            (
+                'NP_001333826.1',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',  # Warn: Versionized ENSG won't be returned due to API limitations
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',
+                    'NP_001333826.1',
+                ],
+            ),
+            (
+                'ENSG00000146648',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'ENST00000275493',
+                    'ENST00000275493.6',
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',  # Warn: Versionized NP won't be returned due to API limitations
+                ],
+            ),
+            (
+                'ENSG00000146648.17',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'ENST00000275493',
+                    'ENST00000275493.6',
+                    'NM_001346897',
+                    'NM_001346897.2',
+                    'NP_001333826',  # Warn: Versionized NP won't be returned due to API limitations
+                ],
+            ),
+            (
+                'ENST00000275493',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'ENST00000275493',
+                    'ENST00000275493.6',
+                ],
+            ),
+            (
+                'ENST00000275493.6',
+                [
+                    'EGFR',
+                    'ERBB',
+                    'ENSG00000146648',
+                    'ENSG00000146648.17',
+                    'ENST00000275493',
+                    'ENST00000275493.6',
+                ],
+            ),
+        ]:
+            equivalent_features = match.get_equivalent_features(conn, feature)
+            equivalent_features = [el['displayName'] for el in equivalent_features]
+            for equivalent_feature in expected:
+                assert equivalent_feature in equivalent_features
+
 
 class TestMatchCopyVariant:
     def test_bad_category(self, conn):

diff --git a/tests/test_vocab.py b/tests/test_vocab.py
@@ -1,6 +1,7 @@
 """
 Tests here depend on specific data in GraphKB which can change. To avoid this, expected/stable values are chosen
 """
+
 import os
 
 import pytest