Skip to content

DeepAVC is a deep learning framework for highly accurate broad-spectrum antiviral compound prediction

License

Notifications You must be signed in to change notification settings

KangBoming/DeepAVC

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

38 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepAVC

DeepAVC is a deep learning framework for highly accurate broad-spectrum antiviral compound prediction.

Abstract

Lethal viruses pose a significant threat to human life, with each pandemic causing millions of fatalities globally. Small-molecule antiviral drugs provide an efficient and convenient approach to antiviral therapy by either inhibiting viral activity or activating the host immune system. However, conventional antiviral drug discovery is often labor-intensive and time-consuming due to the vast chemical space. Although some existing computational models mitigate this problem, there remains a lack of rapid and accurate method specifically designed for antiviral drug discovery. Here, we propose DeepAVC, a universal framework based on pre-trained large language models, for highly accurate broad-spectrum antiviral compound discovery, including DeepPAVC for phenotype-based prediction and DeepTAVC for target-based prediction. We demonstrate the power of DeepAVC in antiviral compound discovery through a series of in silico and in vitro experiments, identifying MNS and NVP-BVU972 as two novel potential antiviral compounds with promising broad-spectrum antiviral activities.

DeepPAVC: Phenotype-based antiviral compounds prediction

DeepPAVC model only takes compound information as input, utilizes a pre-trained molecular encoder to extract compound features, and outputs the antiviral activity score of the input compound Overview

DeepTAVC: Target-based antiviral compounds prediction

DeepTAVC model requires input from two modalities: compounds and proteins. It employs two distinct pre-trained encoders to extract features from each modality separately, then captures intra- and inter- modality interaction patterns through self-attention and cross-attention mechanisms, and outputs the interaction score between the compound and the protein Overview

Publication

Highly accurate prediction of broad-spectrum antiviral compounds with DeepAVC

Main requirements

  • python=3.7.13
  • pytorch=1.10.0
  • cudatoolkit=11.3.1
  • scikit-learn=1.0.2
  • pandas=1.3.5
  • numpy=1.21.5
  • fair-esm=2.0.0
  • rdkit=2021.09.2

Quick start

Step1: clone the repo

mkdir ./DeepPAVC
cd DeepPAVC
git clone https://github.com/KangBoming/DeepAVC.git

Step2: create and activate the environment

cd DeepPAVC
conda env create -f environment.yml
conda activate DeepPAVC

Step3: model training

cd DeepPAVC

Please follow DeepPAVC_train.ipynb 

Please follow DeepTAVC_train.ipynb 

Step4: model infernece

cd DeepPAVC

Please follow DeepPAVC_inference.ipynb 

Please follow DeepTAVC_inference.ipynb 

License

This project is licensed under the MIT License - see the LICENSE.txt file for details

Contact

Please feel free to contact us for any further queations

Boming Kang kangbm@bjmu.edu.cn

Qinghua Cui cuiqinghua@bjmu.edu.cn

About

DeepAVC is a deep learning framework for highly accurate broad-spectrum antiviral compound prediction

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published