• data/ contains all simulated alignments, and the 512-taxon balanced trees (with branch lengths of either 0.5 or 0.01) used during simulation. Alignments are named with this format: <dataset_name>_bl<0.5/0.01>.phy, where bl indicates the branch lengths of the tree used for simulation.

• simulation/ contains all code used for simulating sequences, as well as simulating parameters for use in sequence simulation.

  • ramsey2011_alignments contains all sequence alignments from Ramsey et al. (2011).
  • derive_natural_simulation_parameters.py derives parameters for simulating natural sequences
  • derive_dms_simulation_parameters.py derives parameters for simulating DMS sequences. Note that experimental preferences are in the directory true_simulation_parameters
  • true_simulation_parameters contains all true parameters for simulation, including true dN/dS and entropy, amino-acid fitness, codon frequencies, and selection coefficients
  • simulate_alignments.py simulates a sequence alignment, specifically on UT's (now defunct..) PhyloCluster.

• inference/ contains all code used for mutation-selection model inference. All scripts named *.sh and *.qsub are used for submitting jobs to UT's Phylocluster, and all *.py scripts conduct and process inferences.

• results/ contains all inference results.

  • swmutsel/ contains all inference results with swMutSel for a variety of penalizations, indicated in file name. The script ./results/extract_sw_fitness.py extracts fitness values from the MLE inferences from swMutSel into separate text files for later use
  • phylobayes/ contains all inference results with pbMutSel

• postprocessing/ contains all code used to process, analyze, and plot data (mostly R scripts). All generated plots are also in this directory. All scripts should be executed from this directory! Note: R code requires the packages (and their dependencies) cowplot, ggrepel, dplyr, tidyr, readr, grid, lme4, multcomp, and lmerTest.

  • calculate_inferred_quantities.py calculates dN/dS, entropy, selection coefficient distributions, and JSD for inferences in results/. Resulting quantities are in the subdirectory dataframes.
  • process_results.R processes inference results in dataframes to create the final csv file inference_results.csv
  • plot_figures.R makes all the figures in the manuscript. Figures are saved in either in the subdirectory maintext_figures/ or SI_figures/

• universal_functions.py is a python module containing various functions used throughout the repository.

*This repository is maintained by SJS. Please file any questions/comments in Issues, or contact me at stephanie.spielman@gmail.com.