huabei/miRPVS

Equivalent Graph Neural Network-based Accurate and Ultra-fast Virtual Screening of Small Molecules Targeting miRNA-Protein Complex

Contents

• Equivalent Graph Neural Network-based Accurate and Ultra-fast Virtual Screening of Small Molecules Targeting miRNA-Protein Complex
  • Contents
  • Software Requirements
    • OS Requirements
    • Python Dependencies
  • Installation Guide
    • download this repo
    • install env
  • ligand docking
  • train model
  • eval
  • predict

If you find it useful, please cite:

Equivalent Graph Neural Network-based Virtual Screening of Ultra-large chemical libraries Targeting miRNA-protein complex Huabei Wang; Zhimin Zhang; Guangyang Zhang, Ming Wen* and Hongmei Lu*. Will Published in: DOI:

Software Requirements

autodock vina

python

OS Requirements

The package development version is tested on Linux: Ubuntu 22.04 operating systems.

Python Dependencies

Dependencies for SMTarRNA:

pytorch
pyg
rdkit=2022.09.1

Installation Guide

download this repo

git clone https://github.com/huabei/miRPVS.git

install env

you can install the env via yaml file

cd miRPVS
conda env create -f requirements.yaml

this project use ashleve/lightning-hydra-template as the base project.

ligand docking

train model

This template is suitable for multi-platform operation, please note that the config/local is configured specifically for different platforms.

You just need to configure your own hyperparameters in config/experiment and then run：

python src/train.py experiment=exp_name

The configuration used for this job is also stored in the config/experiment directory and can be used directly.

eval

The config/eval.yaml file needs to be configured with your data locations, model parameter paths, etc. And run:

python src/eval.py

predict

The config/predict.yaml file needs to be configured with your data locations, model parameter paths, etc. And run:

python src/predict.py