DeepLiver Zonation
==================

.. sidebar:: Model Features

   - **Genome**: *mm10*
   - **Type**: Multi-label classification
   - **Parameters**: 12.2M
   - **Size**: 44MB
   - **Input shape**: (500, 4)
   - **Output shape**: (3,)

The **DeepLiver Zonation** is a fine-tuned hybrid CNN-RNN model starting from the trained **DeepLiver Accessibility** model to predict hepatocyte zonation (pericentral, periportal, general).

This transfer-learning strategy overcomes the limited number of regions available for zonation (4181, 1372, and 12122 respectively).

Details of the data and model can be found in the original publication.

-------------------

.. admonition:: Citation

    Bravo González-Blas, C. et al. Single-cell spatial multi-omics and deep learning dissect enhancer-driven gene regulatory networks in liver zonation. Nature Cell Biology 26, 153-167 (2024). https://doi.org/10.1038/s41556-023-01316-4

Usage
-------------------

.. code-block:: python
    :linenos:

    import crested
    import keras

    # download model
    model_path, output_names = crested.get_model("DeepLiver_zonation")

    # load model
    model = crested.utils.load_model(model_path)

    # make predictions
    sequence = "A" * 500
    predictions = crested.tl.predict(sequence, model)
    print(predictions.shape)

-------------------

.. warning::

    DeepLiver_Zonation was originally trained using Tensorflow 1 as the backend.
    Even though the model architecture and weights are exactly the same, there will be slight differences in the output compared to the original model due to backend changes between Tensorflow 1 and 2.
    Overall the correlation between the original and the Keras 3 model is very high (0.99+), but if you want the exact same outputs and contribution plots as in the original publication, you should use an older, compatible environment which you can find in `kipoi <https://kipoi.org/models/DeepLiver/>`_.