MAARDTI: A multi-perspective attention aggregating model for predicting drug-target interaction

Requirments

• pytorch >=1.2
• numpy
• sklearn
• tqdm

Environment

Try the following command for installation.

# Install Python Environment
conda env create -f environment.yml
conda activate MAARDTI

Installation

• You can install the required libraries environment.yml
• If you encounter any installation errors, please don't hesitate to reach out to us for assistance.

Download datasets

Datasets (DrugBank, Davis, KIBA) are provided by project MCANet. Cold Drug/Target/Binding are provided by project DLM-DTI.

Sample test

We provided sample scripts for easily training by MAARDTI.

python start.py data=Sample c_p=256 c_d=8 outpath='samle_p256_d8'

Training and Testing

We provided scripts for easily training by MAARDTI.

python train.py ds=Davis outpath='sample' epoch=300 c_p=16 c_d=8'