diff --git a/tests/translation/test_predict_writer.py b/tests/translation/test_predict_writer.py
index c757774ae..0a71dbcea 100644
--- a/tests/translation/test_predict_writer.py
+++ b/tests/translation/test_predict_writer.py
@@ -1,4 +1,9 @@
-from viscy.translation.predict_writer import _pad_shape
+from iohub import open_ome_zarr
+
+from viscy.data.hcs import HCSDataModule
+from viscy.trainer import VisCyTrainer
+from viscy.translation.engine import VSUNet
+from viscy.translation.predict_writer import HCSPredictionWriter, _pad_shape
 
 
 def test_pad_shape():
@@ -6,3 +11,52 @@ def test_pad_shape():
     assert _pad_shape((4, 5), 4) == (1, 1, 4, 5)
     full_shape = tuple(range(1, 6))
     assert _pad_shape(full_shape, 5) == full_shape
+
+
+def test_predict_writer(preprocessed_hcs_dataset, tmp_path):
+    z_window_size = 5
+    data_path = preprocessed_hcs_dataset
+    channel_split = 2
+    with open_ome_zarr(data_path) as dataset:
+        channel_names = dataset.channel_names
+    dm = HCSDataModule(
+        data_path=data_path,
+        source_channel=channel_names[:channel_split],
+        target_channel=channel_names[channel_split:],
+        z_window_size=z_window_size,
+        target_2d=bool(z_window_size == 1),
+        batch_size=2,
+        num_workers=0,
+    )
+
+    model = VSUNet(
+        architecture="fcmae",
+        model_config=dict(
+            in_channels=channel_split,
+            out_channels=len(channel_names) - channel_split,
+            encoder_blocks=[2, 2, 2, 2],
+            dims=[4, 8, 16, 32],
+            decoder_conv_blocks=2,
+            stem_kernel_size=[z_window_size, 4, 4],
+            in_stack_depth=z_window_size,
+            pretraining=False,
+        ),
+    )
+
+    output_path = tmp_path / "predictions.zarr"
+    prediction_writer = HCSPredictionWriter(
+        output_store=str(output_path), write_input=False
+    )
+
+    trainer = VisCyTrainer(
+        logger=False,
+        callbacks=[prediction_writer],
+        fast_dev_run=False,
+        default_root_dir=tmp_path,
+    )
+
+    trainer.predict(model, datamodule=dm)
+    assert output_path.exists()
+    with open_ome_zarr(output_path) as result:
+        for _, pos in result.positions():
+            assert pos["0"][:].any()
diff --git a/viscy/translation/predict_writer.py b/viscy/translation/predict_writer.py
index 75d6d4152..eb593e4f4 100644
--- a/viscy/translation/predict_writer.py
+++ b/viscy/translation/predict_writer.py
@@ -33,10 +33,12 @@ def _resize_image(image: ImageArray, t_index: int, z_slice: slice) -> None:
             f"T={t_index}, Z-sclice={z_slice}."
         )
         image.resize(
-            max(t_index + 1, image.shape[0]),
-            image.channels,
-            max(z_slice.stop, image.shape[2]),
-            *image.shape[-2:],
+            (
+                max(t_index + 1, image.shape[0]),
+                image.channels,
+                max(z_slice.stop, image.shape[2]),
+                *image.shape[-2:],
+            )
         )
 
 
@@ -142,7 +144,7 @@ def on_predict_start(self, trainer: Trainer, pl_module: LightningModule) -> None
             self.plate = open_ome_zarr(
                 self.output_store, layout="hcs", mode="a", channel_names=channel_names
             )
-        _logger.info(f"Writing prediction to: '{self.plate.zgroup.store.path}'.")
+        _logger.info(f"Writing prediction to: '{self.plate.zgroup.store.root}'.")
         if self.write_input:
             self.source_index = self._get_channel_indices(source_channel)
             self.target_index = self._get_channel_indices(target_channel)