ifDEEPre: large protein language-based deep learning enables interpretable and fast predictions of enzyme commission numbers
This is the tensorflow implementation of ifDEEPre.
The ifDEEPre package is built under the Linux system with the popular softwares Anaconda and Tensorflow. The versions of the software dependencies that both packages use are provided in the environment.yml.
The versions of the software dependencies and data-analysis packages that ifDEEPre has been tested on are given in the environment.yml. Users can conveniently create the same environment by running the command:
conda env create -f environment.yml
The ifDEEPre package does not require any non-standard hardware.
The environment that we use is given in environment.yml. You can create the same environment by running the command:
conda env create -f environment.yml
You can use the trained ifDEEPre models to predict enzyme commission numbers by navigating to the ./src_v6_Final_server folder and running the command:
python code_4_ifdeepre_inputExample.py
The online server of this package is available from this link, ifDEEPre, which is freely available without any registration requirement.
Users can directly upload their protein sequences and get accurate enzyme commission number prediction results conveniently after a short time of waiting. Furthermore, the prediction results can be directly downloaded for convenient future usage.
If you found this library useful in your research, please consider citing
@article{tan2024ifdeepre,
title={ifDEEPre: large protein language-based deep learning enables interpretable and fast predictions of enzyme commission numbers},
author={Tan, Qingxiong and Xiao, Jin and Chen, Jiayang and Wang, Yixuan and Zhang, Zeliang and Zhao, Tiancheng and Li, Yu},
journal={Briefings in Bioinformatics},
volume={25},
number={4},
year={2024},
publisher={Oxford Academic}
}