This repository contains the code to exactly reproduce all results and figures in our paper: AdaPT: An interactive procedure for multiple testing with side information. After the acceptance of the paper, we developed an R package adaptMT that significantly improves the implementation and enriches the flexibility of AdaPT. Unless the reader just aims at reproducing the results in the paper, we recommend checking out our new repository which re-implements all the examples in the paper, though producing slightly different yet improved results.

The folder R/ contains all R files:

• AdaPT.R implements AdaPT as well as its variants (using GLM, GAM and glmnet),
• AdaPT_mix_model/*.R implements the EM pipeline to estimate local FDR. AdaPT_mix_model.R sources all files in AdaPT_mix_model/,
• AdaPT_mix_summary.R and AdaPT_mix_plot.R provides auxiliary functions associated with AdaPT,
• useful_functions.R contains helper functions in developing AdaPT,
• *_expr.R contains the code to implement all examples: GEOquery_expr.R, GEOquery_random_expr.R, Bottomly_expr.R, airway_expr.R, pasilla_expr.R, proteomics_expr.R, simul1.R and simul2.R,
• AdaPT_expr_plot.R produces all figures in the paper,
• Other files provide auxiliary functions.

• GEOquery_expr.R, Bottomly_expr.R, airway_expr.R, pasilla_expr.R and proteomics_expr.R can be executed in a laptop,
• GEOquery_random_expr.R, simul1.R and simul2.R must be executed on a cluster. To do so, submit bash/gen_GEOquery_jobs.sh, bash/gen_jobs_simul1.sh and bash/gen_jobs_simul2.sh to a cluster, respectively.