IMPROVE - SDCNet: Drug Synergy Prediction

---

This is the IMPROVE implementation of the original model with original data. This is the implementation of the script with Loewe synergy values. The code has been restructured so that instead of the original 5-fold cross validation, it is one fold when all three scripts are run.

Dependencies and Installation

Conda Environment

conda create --name sdcnet_IMPROVE python=3.8 pandas=1.3.5 numpy=1.21.2 tensorflow-gpu=2.4.1 scikit-learn=1.2.2
conda activate sdcnet_IMPROVE
pip install git+https://github.com/ECP-CANDLE/candle_lib@develop

Clone this repository

git clone https://github.com/JDACS4C-IMPROVE/SDCNet
cd SDCNet
git checkout IMPROVE-original
cd ..

Clone IMPROVE repository

git clone https://github.com/JDACS4C-IMPROVE/IMPROVE
cd IMPROVE
git checkout develop
cd ..

Download Original Data

The original data is in this repo in /data.

Running the Model

Activate the conda environment:

conda activate sdcnet_IMPROVE

Set environment variables:

export IMPROVE_DATA_DIR="./"
export PYTHONPATH=$PYTHONPATH:/your/path/to/IMPROVE

Run preprocess, train, and infer scripts:

python sdcnet_preprocess_improve.py
python sdcnet_train_improve.py
python sdcnet_infer_improve.py

References

Original GitHub: https://github.com/yushenshashen/SDCNet

Original paper: https://pubmed.ncbi.nlm.nih.gov/36136353/