chore: consolidate variant annotation paths into list

opentargets · Feb 12, 2025 · 35e6ba0 · 35e6ba0
1 parent 89e1785
commit 35e6ba0
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 28 deletions.
diff --git a/src/gentropy/config.py b/src/gentropy/config.py
@@ -464,8 +464,7 @@ class _ConsequenceToPathogenicityScoreMap(TypedDict):
     )
     vep_output_json_path: str = MISSING
     variant_index_path: str = MISSING
-    gnomad_variant_annotations_path: str | None = None
-    lof_curation_variant_annotations_path: str | None = None
+    variant_annotations_path: list[str] | None = None
     hash_threshold: int = 300
     consequence_to_pathogenicity_score: ClassVar[
         list[_ConsequenceToPathogenicityScoreMap]

diff --git a/src/gentropy/variant_index.py b/src/gentropy/variant_index.py
@@ -27,8 +27,7 @@ def __init__(
         vep_output_json_path: str,
         variant_index_path: str,
         hash_threshold: int,
-        gnomad_variant_annotations_path: str | None = None,
-        lof_curation_variant_annotations_path: str | None = None,
+        variant_annotations_path: list[str] | None = None,
         amino_acid_change_annotations: list[str] | None = None,
     ) -> None:
         """Run VariantIndex step.
@@ -38,8 +37,7 @@ def __init__(
             vep_output_json_path (str): Variant effect predictor output path (in json format).
             variant_index_path (str): Variant index dataset path to save resulting data.
             hash_threshold (int): Hash threshold for variant identifier length.
-            gnomad_variant_annotations_path (str | None): Path to extra variant annotation dataset.
-            lof_curation_variant_annotations_path (str | None): Path to loss-of-function variant annotation dataset.
+            variant_annotations_path (list[str] | None): List of paths to extra variant annotation datasets.
             amino_acid_change_annotations (list[str] | None): list of paths to amino-acid based variant annotations.
         """
         # Extract variant annotations from VEP output:
@@ -48,30 +46,20 @@ def __init__(
         )
 
         # Process variant annotations if provided:
-        if gnomad_variant_annotations_path:
-            # Read variant annotations from parquet:
-            annotations = VariantIndex.from_parquet(
-                session=session,
-                path=gnomad_variant_annotations_path,
-                recursiveFileLookup=True,
-                id_threshold=hash_threshold,
-            )
-
-            # Update index with extra annotations:
-            variant_index = variant_index.add_annotation(annotations)
-
-        # Process LOF annotations if provided:
-        if lof_curation_variant_annotations_path:
-            # Read LOF annotations from parquet:
-            lof_annotations = VariantIndex.from_parquet(
-                session=session,
-                path=lof_curation_variant_annotations_path,
-            )
+        if variant_annotations_path:
+            for annotation_path in variant_annotations_path:
+                # Read variant annotations from parquet:
+                annotations = VariantIndex.from_parquet(
+                    session=session,
+                    path=annotation_path,
+                    recursiveFileLookup=True,
+                    id_threshold=hash_threshold,
+                )
 
-            # Update variant index with LOF annotations:
-            variant_index = variant_index.add_annotation(lof_annotations)
+                # Update index with extra annotations:
+                variant_index = variant_index.add_annotation(annotations)
 
-        # If provided read amion-acid based annotation and enrich variant index:
+        # If provided read amino-acid based annotation and enrich variant index:
         if amino_acid_change_annotations:
             for annotation_path in amino_acid_change_annotations:
                 annotation_data = AminoAcidVariants.from_parquet(