Team Tsinghua-A 2023 Software Tool

Antibody drugs, with their better specificity and lower side effects, account for a large proportion of newly approved human drugs. However, the development of drugs for various types of animals is a very slow process. Inspired by the humanization of mouse-derived antibodies, we hope to design antibodies suitable for more species. Leveraging deep learning technology and based on self-collected antibody data, we develop a comprehensive automated monoclonal antibody drug design system. Our system starts with the other sources of antibody sequences against target species antigen (such as mouse-derived antibodies), automatically generates antibody sequences that may have better efficacy and lower immunogenicity, and performs comparative scoring. It then provides users with improved antibody sequences based on their specific requirements. This system can accelerate the development of monoclonal antibody-based animal drugs. Our work helps animals avoid suffering from diseases, and contributes to creating a better living environment for both humans and animals.

Description

This project accomplishes all of the features in the project description, including:

1. construction of FR and CDR homology libraries for multiple species.
2. a rationalized method for scoring antibody immunogenicity.
3. a comprehensive set of automated tools for designing antibodies for different species, enabling multi-dimensional evaluation of sequences.

Installation

To show how our project is used, go through the commands first:

git clone git@gitlab.igem.org:2023/software-tools/tsinghua-a.git

Clone our project locally. Then, pass the command enter the folder:

cd '###"

Go to the project folder. Next, pass the command:

pip install -r requirements.txt

Install all the dependency packages required by the project.

Usage

Pass the command:

python demo.py

Run our project demo. If you wish to score antibody sequences via igfold, please first install igfold according to the igfold documentation and go to the igfold folder, via the command:

python structure_score.py

Generate the score.

If you want to know more algorithm details, please see our wiki page. In addition, if you have questions about using our software, you can check our wiki software page.

https://2023.igem.wiki/tsinghua-a

Details page is as follows:

https://2023.igem.wiki/tsinghua-a/model

https://2023.igem.wiki/tsinghua-a/engineering

https://2023.igem.wiki/tsinghua-a/experiments

https://2023.igem.wiki/tsinghua-a/results

https://2023.igem.wiki/tsinghua-a/software

Authors and acknowledgment

Thank you to all the team members involved in this program for their hard work, please visit our homepage: https://2023.igem.wiki/tsinghua-a/team Also, we refer to the following programs:

1. igfold: https://github.com/Graylab/IgFold

@article{ruffolo2023fast,
  title={Fast, accurate antibody structure prediction from deep learning on massive set of natural antibodies},
  author={Ruffolo, Jeffrey A and Chu, Lee-Shin and Mahajan, Sai Pooja and Gray, Jeffrey J},
  journal={Nature communications},
  volume={14},
  number={1},
  pages={2389},
  year={2023},
  publisher={Nature Publishing Group UK London}
}
@article{ruffolo2021deciphering,
    title = {Deciphering antibody affinity maturation with language models and weakly supervised learning},
    author = {Ruffolo, Jeffrey A and Gray, Jeffrey J and Sulam, Jeremias},
    journal = {arXiv},
    year= {2021}
}

2. Deep one class：https://github.com/PramuPerera/DeepOneClass

@ARTICLE{2018arXiv180105365P,
author = {{Perera}, P. and {Patel}, V.~M.},
title = "{Learning Deep Features for One-Class Classification}",
journal = {ArXiv e-prints},
archivePrefix = "arXiv",
eprint = {1801.05365},,
year = 2018,
month = jan,
}
}