fix: raise ValueError when combat batch has fewer than 2 cells

docs/release-notes/3994.fix.md

@@ -0,0 +1 @@
+{func}`scanpy.pp.combat` now raises a {class}`ValueError` when a batch contains fewer than 2 cells, instead of silently producing NaN values in the corrected data {smaller}`L Zhang`

src/scanpy/preprocessing/_combat.py

@@ -202,10 +202,22 @@ def combat(  # noqa: PLR0915
     sanitize_anndata(adata)
 
     # construct a pandas series of the batch annotation
-    model = adata.obs[[key, *(covariates if covariates else [])]]
-    batch_info = model.groupby(key, observed=True).indices.values()
+    model: pd.DataFrame = adata.obs[[key, *(covariates if covariates else [])]]
+    batch_info = model.groupby(key, observed=True).indices
     n_batch = len(batch_info)
-    n_batches = np.array([len(v) for v in batch_info])
+    n_batches = np.array([len(v) for v in batch_info.values()])
+
+    # check for batches with fewer than 2 cells
+    small_batches = [
+        batch for batch, size in zip(batch_info, n_batches, strict=True) if size < 2
+    ]
+    if small_batches:
+        msg = (
+            f"Batches {small_batches!r} have fewer than 2 cells. "
+            "ComBat requires at least 2 cells per batch to estimate "
+            "within-batch variance. Filter these batches before running combat."
+        )
+        raise ValueError(msg)
     n_array = float(sum(n_batches))
 
     # standardize across genes using a pooled variance estimator
@@ -220,7 +232,9 @@ def combat(  # noqa: PLR0915
         la.inv(batch_design.T @ batch_design) @ batch_design.T @ s_data.T
     ).values
     # first estimate for the multiplicative batch effect
-    delta_hat = [s_data.iloc[:, batch_idxs].var(axis=1) for batch_idxs in batch_info]
+    delta_hat = [
+        s_data.iloc[:, batch_idxs].var(axis=1) for batch_idxs in batch_info.values()
+    ]
 
     # empirically fix the prior hyperparameters
     gamma_bar = gamma_hat.mean(axis=1)
@@ -233,7 +247,7 @@ def combat(  # noqa: PLR0915
     # gamma star and delta star will be our empirical bayes (EB) estimators
     # for the additive and multiplicative batch effect per batch and cell
     gamma_star, delta_star = [], []
-    for i, batch_idxs in enumerate(batch_info):
+    for i, batch_idxs in enumerate(batch_info.values()):
         # temp stores our estimates for the batch effect parameters.
         # temp[0] is the additive batch effect
         # temp[1] is the multiplicative batch effect
@@ -257,7 +271,7 @@ def combat(  # noqa: PLR0915
 
     # we now apply the parametric adjustment to the standardized data from above
     # loop over all batches in the data
-    for j, batch_idxs in enumerate(batch_info):
+    for j, batch_idxs in enumerate(batch_info.values()):
         # we basically subtract the additive batch effect, rescale by the ratio
         # of multiplicative batch effect to pooled variance and add the overall gene
         # wise mean

tests/test_combat.py

@@ -75,6 +75,20 @@ def test_combat_obs_names():
     assert_equal(a, b)
 
 
+def test_combat_single_cell_batch():
+    """Test that combat raises an error when a batch has fewer than 2 cells.
+
+    Regression test for https://github.com/scverse/scanpy/issues/1175
+    """
+    adata = sc.datasets.blobs()
+    # Create a batch where one category has only 1 cell
+    batch = pd.Categorical(["single"] + ["other"] * (adata.n_obs - 1))
+    adata.obs["batch"] = batch
+
+    with pytest.raises(ValueError, match="fewer than 2 cells"):
+        sc.pp.combat(adata, key="batch")
+
+
 def test_silhouette():
     # this test checks wether combat can align data from several gaussians
     # it checks this by computing the silhouette coefficient in a pca embedding