Detailed evaluation results and web prediction tools are available at: http://drugreposition.goodluckcome.com.
Matlab 2020a
Download by
git clone https://github.com/yyq83/DR-method-evaluation.git
and install dependencies by
conda env create -f environment.yaml
Go to the Snakemake folder
cd Snakemake
If you only want to get the prediction results and scalability (time and memory peak consumption) of the method on the specified dataset, run the command(rule run_method_pre):
snakemake -j 1 {outdir}/{method}_{dataset}.csv
Example : BNNR method on Fdataset
Command : snakemake -j 1 Evaluation/BNNR_Fdataset.csv
Following the completion of the run, you'll discover the prediction results within the Evaluation folder. Additionally, you can access logs detailing time and memory consumption in their respective Benchmark and Log folders.
If you want to get the full results of the method on the specified dataset, including prediction results, scalability,performance,auc curve and aupr curve,run the command(rule evaluate):
snakemake -j 1 {outdir}/{method}/{dataset}/Plot/{method}_{dataset}_auc.png
or
snakemake -j 1 {outdir}/{method}/{dataset}/Plot/{method}_{dataset}_aupr.png
Example : BNNR method on Cdataset
Command : snakemake -j 1 Evaluation/BNNR/Cdataset/Plot/BNNR_Cdataset_auc.png
After completing the run, you can locate comprehensive results in the Evaluation folder. This encompasses prediction results, AUC curves, and AUPR curves found in the Plot folder, as well as AUC, AUPR, and F1 values stored in either the Benchmark folder or the Log folder.
Users can add the command line parameter --use-conda, and snakemake will automatically create a running environment.
The following datasets were used in our study：
Fdataset, Cdataset, Ydataset, DNdataset, HDVD, LAGCN, LRSSL, SCMFDD_L, deepDR, iDrug, TLHGBI, which is available at: https://zenodo.org/record/8357512.
