DeepSEA is a deep learning-based algorithmic framework for predicting the chromatin effects of sequence alterations with single nucleotide sensitivity. This implements by PyTorch again.

Jian Zhou, Olga G. Troyanskaya. Predicting the Effects of Noncoding Variants with Deep learning-based Sequence Model. Nature Methods (2015).

Considering your ease of use, I have included the most recent version numbers of the software packages for the configuration that worked for me. For the record, I am using Ubuntu Linux 16.04 LTS with an NVIDIA Titan 1080Ti GPU and 32GB RAM.

• [Python] (https://www.python.org) (3.6.8). The easiest way to install Python and all of the necessary dependencies is to download and install [Anaconda] (https://www.anaconda.com/download/) (4.6.14).

• [PyTorch] (https://pytorch.org/) (1.0.1).

• [CUDA] (https://developer.nvidia.com/cuda-80-download-archive) (8.0)

• [cuDNN] (https://developer.nvidia.com/rdp/cudnn-download) (7.1.3)

You need to first download the training, validation, and testing sets from DeepSEA. You can download the datasets from [here] (http://deepsea.princeton.edu/media/code/deepsea_train_bundle.v0.9.tar.gz). After you have extracted the contents of the tar.gz file, move the 3 .mat files into the data/ folder.

Because of my RAM limited, I firstly transform the train.mat file to 10 .pt files. If you don't worry about this problem, you can fix the train-part code according to the valid-part code in DeepSEA_train.py file.

Then you can train the model DeepSEA_train.py initially. Don't forget to install visdom and fix the save_model_time parameter according to your needs. Due to safety concerns, I set many model-saving checkpoints, you can fix it flexibility.(For your convenience, I've already uploaded the my bestmodel in the hyperlink, and I am grateful that if you can update it.)

When you have trained successfully, you can use **DeepSEA_test.ipynb ** to evaluate the model.Because of the flexibility of the jupyter notebook, I integrate the pred, ROC/PR curve and aus file together.

For convenience，you can download my trained [bestmodel] (https://pan.baidu.com/s/1h_LJIwP5ozUoBfiXFSCSnQ) with the password 'u0t6' .

[DeepSEA] (http://deepsea.princeton.edu/job/analysis/create/)