GuiltyTargets Results

This repository contains the results of¹:

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.

R Installation

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"))'

Python Installation

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 .

Running

To run the code:

$ source run.sh

Output

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

---

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.↩