exercise the protein specific code

Author: pryvkin10x

.github/workflows/prepare_test_data.yaml

@@ -35,6 +35,12 @@ jobs:
                   # 10x Genomics Xenium 3.0.0 (5K) Mouse ileum, nuclear expansion
                   curl -O https://cf.10xgenomics.com/samples/xenium/3.0.0/Xenium_Prime_Mouse_Ileum_tiny/Xenium_Prime_Mouse_Ileum_tiny_outs.zip
 
+                  # 10x Genomics Xenium 4.0.0 (v1) Human ovary, nuclear expansion
+                  curl -O https://cf.10xgenomics.com/samples/xenium/4.0.0/Xenium_V1_Human_Ovary_tiny/Xenium_V1_Human_Ovary_tiny_outs.zip
+
+                  # 10x Genomics Xenium 4.0.0 (v1) Human ovary, multimodal cell segmentation
+                  curl -O https://cf.10xgenomics.com/samples/xenium/4.0.0/Xenium_V1_MultiCellSeg_Human_Ovary_tiny/Xenium_V1_MultiCellSeg_Human_Ovary_tiny_outs.zip
+
                   # 10x Genomics Xenium 4.0.0 (v1+Protein) Human kidney, multimodal cell segmentation
                   curl -O https://cf.10xgenomics.com/samples/xenium/4.0.0/Xenium_V1_Protein_Human_Kidney_tiny/Xenium_V1_Protein_Human_Kidney_tiny_outs.zip
 

src/spatialdata_io/readers/xenium.py

@@ -68,6 +68,7 @@ def xenium(
     aligned_images: bool = True,
     cells_table: bool = True,
     n_jobs: int = 1,
+    gex_only: bool | None = True,
     imread_kwargs: Mapping[str, Any] = MappingProxyType({}),
     image_models_kwargs: Mapping[str, Any] = MappingProxyType({}),
     labels_models_kwargs: Mapping[str, Any] = MappingProxyType({}),
@@ -121,6 +122,8 @@ def xenium(
         Whether to read the cell annotations in the `AnnData` table.
     n_jobs
         Number of jobs to use for parallel processing.
+    gex_only
+        Whether to load only the "Gene Expression" feature type.
     imread_kwargs
         Keyword arguments to pass to the image reader.
     image_models_kwargs
@@ -186,7 +189,7 @@ def xenium(
             cells_table = True
 
     if cells_table:
-        return_values = _get_tables_and_circles(path, cells_as_circles, specs)
+        return_values = _get_tables_and_circles(path, cells_as_circles, specs, gex_only)
         if cells_as_circles:
             table, circles = return_values
         else:
@@ -561,9 +564,9 @@ def _get_points(path: Path, specs: dict[str, Any]) -> Table:
 
 
 def _get_tables_and_circles(
-    path: Path, cells_as_circles: bool, specs: dict[str, Any]
+    path: Path, cells_as_circles: bool, specs: dict[str, Any], gex_only: bool
 ) -> AnnData | tuple[AnnData, AnnData]:
-    adata = _read_10x_h5(path / XeniumKeys.CELL_FEATURE_MATRIX_FILE)
+    adata = _read_10x_h5(path / XeniumKeys.CELL_FEATURE_MATRIX_FILE, gex_only=gex_only)
     metadata = pd.read_parquet(path / XeniumKeys.CELL_METADATA_FILE)
     np.testing.assert_array_equal(metadata.cell_id.astype(str), adata.obs_names.values)
     circ = metadata[[XeniumKeys.CELL_X, XeniumKeys.CELL_Y]].to_numpy()

tests/test_xenium.py

@@ -188,3 +188,49 @@ def test_cli_xenium(runner: CliRunner, dataset: str) -> None:
         )
         assert result.exit_code == 0, result.output
         _ = read_zarr(output_zarr)
+
+@skip_if_below_python_version()
+@pytest.mark.parametrize(
+    ("dataset", "gex_only",),
+    [("Xenium_V1_human_Lung_2fov_outs", False),
+     ("Xenium_V1_human_Lung_2fov_outs", True),
+     ("Xenium_V1_Human_Ovary_tiny_outs", False),
+     ("Xenium_V1_Human_Ovary_tiny_outs", True),
+     ("Xenium_V1_MultiCellSeg_Human_Ovary_tiny_outs", False),
+     ("Xenium_V1_MultiCellSeg_Human_Ovary_tiny_outs", True),
+     ("Xenium_V1_Protein_Human_Kidney_tiny_outs", False),
+     ("Xenium_V1_Protein_Human_Kidney_tiny_outs", True)
+    ]
+)
+def test_xenium_other_feature_types(dataset: str, gex_only: bool) -> None:
+    f = Path("./data") / dataset
+    assert f.is_dir()
+    sdata = xenium(f, cells_as_circles=False, gex_only=gex_only)
+    if gex_only:
+        assert set(sdata["table"].var["feature_types"]) == {"Gene Expression"}
+    elif dataset == "Xenium_V1_human_Lung_2fov_outs":
+        assert set(sdata["table"].var["feature_types"]) == {"Deprecated Codeword",
+                                                            "Gene Expression",
+                                                            "Negative Control Codeword",
+                                                            "Negative Control Probe",
+                                                            "Unassigned Codeword"}
+    elif dataset in {"Xenium_V1_Human_Ovary_tiny_outs", "Xenium_V1_MultiCellSeg_Human_Ovary_tiny_outs"}:
+        assert set(sdata["table"].var["feature_types"]) == {"Gene Expression",
+                                                            "Genomic Control",
+                                                            "Negative Control Codeword",
+                                                            "Negative Control Probe",
+                                                            "Unassigned Codeword"}
+    elif dataset == "Xenium_V1_Protein_Human_Kidney_tiny_outs":
+        assert set(sdata["table"].var["feature_types"]) == {"Gene Expression",
+                                                            "Genomic Control",
+                                                            "Negative Control Codeword",
+                                                            "Negative Control Probe",
+                                                            "Protein Expression",
+                                                            "Unassigned Codeword"}
+        # Protein feature
+        assert np.allclose(sdata["table"].X[0:3, sdata["table"].var_names.str.match("VISTA")].toarray().squeeze(), [0.7, 1.2, 0.0])
+        # RNA feature
+        assert np.allclose(sdata["table"].X[[6,7,24], sdata["table"].var_names.str.match("ACTG2")].squeeze(), [1, 0, 2])
+
+    else:
+        assert ValueError(f"Unexpected dataset {dataset}")