This folder contains all scripts needed to 
  1) compute the silencing dynamics for X chromsomal genes based on undifferentiated PRO-seq and mRNA-seq data as well as differentiated mRNA-seq data: scripts/fit_halftimes.R
     - input data for teh script can be found in the subfolder data/silencing_halftimes/raw_data/ , which contains the mapped PRO-seq and mRNA-seq data that can also be downloaded from the GEO submission (GSE121144)
     - the output of this script, the computed half-times on mm10, can be found in data/silencing_halftimes/fitted_data/
  2) run the Random Forest pipeline, which trains a Random Forest model: model.R and model_functions.R
     - model.R is the main script where all files and parameters are specified while model_functions.R contains all functions that are executed in model.R
     - main input is the pre-computed feature matrix (data/feature_matrix/promoter_matrix_reannotated_normRAdjusted_pro_seq_genes.RData) that also contains corresponding half-times
     - the threshold ranges used for both model are listed in the script as well as the training parameters for the Random Forest model and parameters for the forest-guided clustering
     - predictions are made based on a second input feature matrix (data/feature_matrix/promoter_matrix_reannotated_normRAdjusted_all_chrX_genes.RData) which contains all features for all X chromosomal genes with active TSSs
     - additional input data needed for predictions with the Random Forest model are in the subfolder data/annotation_files/