GuiltyTargets Results

This repository contains the results of [1]:

[1]	Muslu, Ö., Hoyt, C. T., Hofmann-Apitius, M., & Fröhlich, H. (2019). GuiltyTargets: Prioritization of Novel Therapeutic Targets with Deep Network Representation Learning. bioRxiv, 1–14.

Due to licensing reasons, analyses that use TTD drug targets and Alzheimer's disease data sets have been removed from this reproduction.

Installation

You will need Python 3.7+ and R 3.6.0+ to run the program.

On mac, install the latest version of R with:

$ brew install R

Install BioConductor with the instructions from https://www.bioconductor.org/install:

$ R -e 'install.packages("BiocManager")'
$ R -e 'BiocManager::install()'
$ R -e 'BiocManager::install(c("limma", "GEOquery", "Biobase"))'

To install the required Python libraries, you can run:

$ git clone https://github.com/GuiltyTargets/reproduction.git guiltytargets-results
$ cd guiltytargets-results
$ pip install -e .

To run the code:

$ source run.sh

You can find the output under reproduction/data. The results.csv file gives an overview of all AUROC values under different settings.