Predict lncRNA-drug associations based on graph neural networks

1. Introduction

This reository contains source code and data for the paper "Predict lncRNA-drug associations based on graph neural networks".

2. Installation

This python code depends on the following packages. You must have them installed before running this code.

conda install pytorch==1.8.1 torchvision==0.16.0 -c pytorch
pip install pandas pip install numpy
pip install scipy
pip install torch_geometric
pip install networkx
pip install pickle
pip install scikit-learn

3. Usage

3.1 Data

The all datasets are freely downloaded from https://www.jianglab.cn/D-lnc/index.jsp

3.2 Training model

Run Main.py, and set train parameter to TRUE and use-feature parameter to TRUE, specifying the dataset to train the model.

3.3 Predicting LDAs

Run Main.py and set the parameter train to False and the parameter use-features to TRUE, using the model.pth file generated in 3.2 to specify the data set to predict potential LDAs.