Merge pull request #48 from Mye-InfoBank/solo-on-small-batches

Fix bug with SOLO if batch size is smaller than 128

modules/solo.nf

@@ -27,6 +27,16 @@ process SOLO {
 
     adata = sc.read_h5ad("${adata}")
 
+    adata_batch = adata[adata.obs.batch == "${batch}"]
+    batch_size = len(adata_batch)
+
+    if batch_size < 128:
+        print(f"Batch size too small ({batch_size}), skipping SOLO")
+        solo_res = pd.DataFrame(index=adata_batch.obs.index)
+        solo_res["doublet_label"] = "Unknown"
+        solo_res.to_pickle("${new_meta.id}.solo.pkl")
+        exit(0)
+
     if ${has_celltypes ? "True" : "False"}:
         scvi.model.SCANVI.setup_anndata(adata, batch_key="batch", labels_key="cell_type", unlabeled_category="Unknown")
         scvi_model = scvi.model.SCANVI.load("${scvi_model}", adata=adata)