DeepBrain

⚠️ We recommend using our latest enhancer modeling software package, CREsted, which provides well documented code for training and evaluating enhancer models. The models in this repository (DeepMouseBrain1&2, DeepHumanCortex1&2 and DeepChickenBrain1) and the CREsted models from the Hecker & Kempynck et al. paper (DeepMouseBrain3, DeepHumanBrain and DeepChickenBrain2) are also available in CREsted. ⚠️

DeepBrain contains an example of using our enhancer models to score and understand any region in any genome for the cell types in our datasets.

Installation and usage

1. Clone the repo

git clone https://github.com/aertslab/DeepBrain.git

2. Install libraries

Create and activate conda environment:

conda env create -f environment.yml
conda activate DeepBrain

If the installation fails for some reason, another option is to run the steps in the following script for a manual installation of all packages and the environment:

./install.sh

If you are using a GPU (recommended), and it is not found after installation, a potential fix may be to link an installed libcusolver.so.11 to the correct path:

#Define CUDA_INSTALL_PATH depending on where it is installed on the local machine
ln -s $CUDA_INSTALL_PATH/CUDA/11.3.1/lib64/libcusolver.so.11 $(python -c "import tensorflow.python as x; print(x.__path__[0])")/libcusolver.so.10

3. Download the DeepBrain models

The weights of the models are stored using Git Large File Storage (LFS). To download them, you will need to have installed Git LFS (https://git-lfs.com/). On Linux, you can install Git LFS with the following command if it was not installed yet:

sudo apt-get install git-lfs

Then the following commands are required after installation to retrieve the model weights:

git lfs install
git lfs pull

If Git LFS does not work, you can also download the model weights from Zenodo: https://zenodo.org/records/10868679

4. Usage

Run the notebook DeepBrain_example.ipynb for example usage for predicting on genomic regions, getting contribution scores and calculating correlation between cell types. If you are running JupyterLab, you can make the environment visible by running:

ipython kernel install --user --name DeepBrain --display-name "DeepBrain"

Citation

If the models or accompanying files are helpful for your research please cite the following publication:

Enhancer-driven cell type comparison reveals similarities between the mammalian and bird pallium

Nikolai Hecker*, Niklas Kempynck*, David Mauduit, Darina Abaffyová, Roel Vandepoel, Sam Dieltiens, Ioannis Sarropoulus, Carmen Bravo González-Blas, Elke Leysen, Rani Moors, Gert Hulselmans, Lynette Lim, Joris De Wit, Valerie Christiaens, Suresh Poovathingal, Stein Aerts

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!