DeepDDI2

#Project This project is to develop a framework that systematically predicts drug-drug interactions (DDIs).

#Features

• DeepDDI2 predicts DDIs using names of drug-drug pairs and their structural information (SMILES) as inputs.
• DeepDDI2 predicts 113 DDI types in human-readable sentences as output.

###Procedure Note: This source code was developed in Linux, and has been tested in Ubuntu 16.04

1. Clone the repository

    git clone https://bitbucket.org/kaistsystemsbiology/deepddi2.git
   

2. Create and activate a conda environment

    conda env create -f environment.yml
    conda activate deepddi
   

##Input files for the DeepDDI2## Following working input files can be found in: ./data/. These files were used for the data presented in the manuscript.

#Example#

python run_DeepDDI.py -i ./DDI_input.txt -o ./output

Further details can be found in Supplementary Information.