README.md

DGDTA

Dynamic graph attention network for predicting drug-target binding affinity. It is built with *Pytorch and Python 3.

Installation

Requirements

• python 3.9, numpy, scipy, pandas, pytorch, pyg

1. Create a virtual environment

# create
conda create -n DGDTA python=3.9
# activate
conda activate DGDTA
# deactivate
conda deactivate

2. clone DGDTA

• After creating and activating the DGDTA virtual environment, download DGDTA from github:

git clone https://github.com/luojunwei/DGDTA.git
cd DGDTA

3. Install

conda activate DGDTA
conda install numpy, scipy, pandas, Pytorch, pyg

Tested data

The example data can be downloaded from

Davis and KIBA

https://github.com/thinng/GraphDTA/tree/master/data

Usage

Train Model

1. Create Dataset

python data_creation.py

First, divide the data into training and test sets and create data files in pytorch format.

2. Train model

Run the following script to train the model.

python training2.py

The default values of the parameter parser are the DGDTA-CL version and the KIBA dataset.

3. Validate the training prediction model

Run the following script to test the model.

python training_validation.py 

This returns the best MSE model for the validation dataset during the training process.