organoids

chanzuckerberg · Jan 9, 2025 · 5f05cba · 5f05cba
1 parent e15138b
commit 5f05cba
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 4 deletions.
diff --git a/tools/cellxgene_census_builder/src/cellxgene_census_builder/build_soma/anndata.py b/tools/cellxgene_census_builder/src/cellxgene_census_builder/build_soma/anndata.py
@@ -330,7 +330,7 @@ def make_anndata_cell_filter(filter_spec: AnnDataFilterSpec) -> AnnDataFilterFun
     def _filter(ad: AnnDataProxy) -> AnnDataProxy:
         """Filter observations and features per Census schema."""
         var_mask = ad.var.feature_biotype == "gene"
-        obs_mask = pd.Series(np.ones(ad.n_obs, dtype=bool), index=ad.obs.index)
+        obs_mask = ad.obs.tissue_type.isin(["tissue", "organoid"])
 
         # Handle multi-species edge case
         var_organisms = set(ad.var.feature_reference[var_mask].unique())

diff --git a/tools/cellxgene_census_builder/src/cellxgene_census_builder/build_soma/validate_soma.py b/tools/cellxgene_census_builder/src/cellxgene_census_builder/build_soma/validate_soma.py
@@ -239,9 +239,9 @@ def validate_axis_dataframes_global_ids(
                 (obs_unique_joinids[0] == 0) and (obs_unique_joinids[-1] == (len(obs_unique_joinids) - 1))
             )
 
-            # Validate that we only contain primary tissue cells, no organoid, cell culture, etc.
+            # Validate that we only contain primary tissue cells and organoids, no cell culture, etc.
             # See census schema for more info.
-            # assert (census_obs_df.tissue_type == "tissue").all()
+            assert (census_obs_df.tissue_type.isin(["tissue", "organoid"])).all()
 
             # Assert the stats values look reasonable
             assert all(

diff --git a/tools/cellxgene_census_builder/tests/anndata/test_anndata.py b/tools/cellxgene_census_builder/tests/anndata/test_anndata.py
@@ -108,7 +108,8 @@ def test_make_anndata_cell_filter_filters_out_organoids_cell_culture(
     filtered_adata_with_organoids_and_cell_culture = func(adata_with_organoids_and_cell_culture)
 
     assert adata_with_organoids_and_cell_culture.var.equals(filtered_adata_with_organoids_and_cell_culture.var)
-    assert filtered_adata_with_organoids_and_cell_culture.obs.shape[0] == 4
+    # tissue(2) and organoid(1). cell_culture should be excluded
+    assert filtered_adata_with_organoids_and_cell_culture.obs.shape[0] == 3
 
 
 def test_make_anndata_cell_filter_organism(tmp_path: pathlib.Path, h5ad_with_organism: str) -> None: