This repository includes the scripts that were used in the paper called [The ability to classify patients based on gene-expression data varies by algorithm and performance metric](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1009926) in PLOS Computational Biology. This repository has several components: * `Biomarker_Benchmark_Analysis.tar.gz` is an archive with the scripts and data that we used for the analysis. *Executing these steps required thousands of hours of computer time.* * The `RawResults` folder contains archives with the raw ShinyLearner output of our analyses. * `Summarized_Results.tar` is an archive with the results in summarized form. * The scripts in the `FigureAndTableScripts` folder were used to generate the figures and tables for the manuscript. Biomarker_Benchmark_Figures.R is the primary script (it loads the other scripts). It also downloads Summarized_Results.tar from this repository and extracts its contents.