The regime files are located in the regime_files/ folder. These files define the HA-R, HA-S, and LA-S regimes for EvoGeneX. Each node in the tree is labelled by either "chosen" or "background."

The tree file used is in the tree_files/ folder. The file is in Newick format.

The data used for our simulation files are available in the Supplementary Files. See below on how to generate new data under the same parameter settings.

First, you will need to download the 24-subline data from the Trisicell API and save it locally in a readable format. This is done in prep_data/get_multigene_data.py. By default, this script will use the tpm-normalized data, output files with up to 1000 genes for all 55,401 genes, include only genes at least one non-zero value for each species, and log-transform the values. You can change the maxiumum number of genes in each file using the --inc flag. If you do not want to save all the genes, you can specify the gene range you are interested in using the --start and --end flags. Genes are separated into multiple files for parallelization of running EvoGeneX. If you want a single file with all the genes, use --inc 55401.

The code to run EvoGeneX is in the run_scripts/ folder. There are three files for running either adaptive or constrained versions: _runner.sh, _driver.sbatch, and _evogenex.R. The _runner.sh file loops over all gene files in the specified folder and either calls the _driver.sbatch or the _evogenex.R scripts for each file individually, depending on if you specify using slurm or the cpu (by default slurm is used). The _driver.sbatch file submits a slurm job to run the _evogenex.R script for the specified file. If you want to submit a single slurm job rather than jobs for all the gene files, you can use this script directly. The _evogenex.R script runs EvoGeneX over all the genes in the specified file. If you want to run EvoGeneX on a single gene file on the cpu, you can call this script directly.

These scripts will save the results in specified file locations.

Evaluation scripts are in the eval_scripts/ folder. These files organize the results and add more statistics columns. The get_gene_results.py script saves files with only positive results and all the results and organizes the columns. The add_stats_cols.R script will add columns for adjusted p-values, and log_2 fold change. The get_all_gene_results.sh will loop through all regimes and run get_gene_results.py and add_stats_cols.R for all results. The output of these files will provide files like the 24 subline results files in the Supplementary Files.

These scripts assume that the results are saved in results/tpm/adpt_{regime}/ and are saved in results/tpm/gene_lists/adpt_{regime}/. Edit the paths as needed to fit your file structure.

The scripts in kmeans/ provides code to cluster the results and make the heatmap in Figure 2. kmeans.py runs kmeans clustering on the results (with a defined random seed for reproducibility). heatmap.py plots the clusters as a heatmap.

These scripts assume that the results are saved in results/tpm/adpt_{regime}/ and are saved in results/tpm/. Edit the paths as needed to fit your file structure.

Scripts to run dN/dS methods are in the dNdS/ folder. To run dN/dS using the counting method, you can run dNdS_by_counts.py. To run dNdScv, first you will need to download the GRCm38.p2 mouse reference genome (NCBI RefSeq assembly GCF_000001635.22; https://www.ncbi.nlm.nih.gov/datasets/genome/GCF_000001635.22/). The script assumes this file is in the highest level folder and named RefCDS_mouse_GRCm38.p2.rda, however you can modify the code to point to the file in your directories. Second, you will also need a file with the mutation information in the required format. These files are pregenerated in dNdS/mut_tables, or they can be regenerated by running get_mut_table.py. Then run dndscv.R.

Scripts to evaluate the genes with adaptive expression within the differentially expressed genes between post-treatment responders and non-responders are in the eval_scripts/treatment_eval/ folder. The deseq2_eval.py script will calculate the number of genes with adaptive expression that are also differentially expressed and save the information about the adaptivity and differential expression to a file.

This script assumes that the adaptivity results are saved in results/tpm/gene_lists/adpt_{regime}/ and evaluation results are saved in results/tpm/gene_lists/adpt_{regime}/enrichment/. Edit the paths as needed to fit your file structure.

Scripts to simulate data can be found in the simulation/ folder. simulate_neutral.R will simulate neutral data using a Brownian motion process. simuate_adaptive.R will simulate constrained data (using theta ratio parameter = 0) or adaptive data (theta ratio parameter != 0). You can control all parameter values using the parameters. You can run multiple types of simulations using the _driver.sh files, where you can specify multiple parameter options and it will run simulation scripts for the cross product of all the options. If you want to run this on a cluster, the _runner.sh files start the _driver.sh jobs using slurm.

You can run EvoGeneX on the simulated data using the same scripts used to run it on the 24 subline data in the run_scripts/ folder. See above for details.

You can run differential expression on the simulated data using the simulation/differential_expression.py script.

Scripts to organize and evaluate the simulation results are in the eval_scripts/simulation/ folder. get_sim_results.sbatch will run eval_scripts/get_results/get_gene_results.py and eval_scripts/get_results/add_stats_cols.R and generate results files. Then, get_csv.sh will collate the results for all the simulations into a single file. The get_de_results.py will collate the results of differential expression on all simulations into a single file.

Scripts used to generate figures are in the figures/ folder. See readme in that folder for further information and instructions.

In all R files, make sure that the library paths are set correctly. If not using the default install location, set the library path using .libPaths("path_to_library"). If using the default, delete this line.