Pep2TCR: accurate prediction of CD4 T cell receptor binding specificity through transfer learning and ensemble approach

Overview

Pep2TCR can serve as a valuable tool for CD4 TCR specificity prediction and biology applications. This Github repository comprises the codes of Pep2TCR, providing comprehensive guidance for researchers interested in Pep2TCR. Also, we provide a docker image at Docker Hub and a website at http://pep2tcr.liuxslab.com for convenient usage.

Contents

Users can download Pep2TCR package with git clone https://github.com/XSLiuLab/Pep2TCR.git, there are some contents:

• data contains collected data used in the project.
• model contains Pep2TCR model codes.
• TCR_web contains Pep2TCR website codes.
• Pep2TCR.py is a startup interface.
• environment.yaml is a user-friendly yaml file for creating a conda environment.

Environment

We used Pytorch to train and validate Pep2TCR, so users should install the following packages:

• python == 3.8.13
• pytorch == 1.12.0
• pandas == 1.5.3
• numpy == 1.23.5
• scikit-learn == 1.2.2

If the user's system is equipped with a GPU, they can install cudatoolkit == 11.3.1, which will result in an acceleration of prediction speed.

Users also can setting up Conda environment through conda install -c conda-forge mamba,mamba env create -f environment.yaml, this might take a bit of time, but it's convenient. In addition, please rewrite your .condarc file as following:

channels:
  - conda-forge
  - bioconda
  - menpo
  - main
  - r
  - msys2
  - pytorch
  - pytorch-lts
  - simpleitk
show_channel_urls: true

Usage

Please modify the ab_path paramemter of the paras.py file in model\code folder to /path/to/model_dir as the first time use.

Pep2TCR has two modes: Single mode and Batch mode. The help page of Pep2TCR is as follows:

usage: Pep2TCR.py [-h] [--mode {single,batch}] [--cdr3 CDR3] [--pep PEP] [--data_path DATA_PATH] [--outdir OUTDIR]

optional arguments:
  -h, --help            show this help message and exit
  --mode {single,batch} default is single mode
  --cdr3 CDR3
  --pep PEP
  --data_path DATA_PATH csv format, first column is CDR3, second column is Epitope
  --outdir OUTDIR       default is . (current directory)

For Single mode, the command line is python Pep2TCR.py --mode single --cdr3 xxx --pep xxx

For Batch mode, the command line is python Pep2TCR.py --mode batch --data_path /file path --outdir .

Citation

Waiting ... ...

Acknowledgement

We acknowledge the computing services provided by ShanghaiTech University High Performance Computing Public Service Platform. This work received supports from the Shanghai Science and Technology Commission (21ZR1442400), the National Natural Science Foundation of China (31771373), and startup funding from ShanghaiTech University.

License

---

Cancer Biology Group @ShanghaiTech

Research group led by Xue-Song Liu in ShanghaiTech University