DeepLiver Activity

Contents

DeepLiver Activity#

The DeepLiver Activity is a fine-tuned hybrid CNN-RNN model starting from the trained DeepLiver Accessibility model to predict hepatocyte MPRA activity in vivo.

This transfer-learning strategy overcomes the limited number of regions available for activity (4215).

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

Citation

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

Usage#

 1import crested
 2import keras
 3
 4# download model
 5model_path, output_names = crested.get_model("DeepLiver_activity")
 6
 7# load model
 8model = keras.models.load_model(model_path)
 9
10# make predictions
11sequence = "A" * 500
12predictions = crested.tl.predict(sequence, model)
13print(predictions.shape)

Warning

DeepLiver_Activity 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.

Contents