Ensemble Epistasis

This repository contains scripts and jupyter notebooks used for the analysis presented in Ensemble epistasis: thermodynamic origins of non-addivity between mutations by Anneliese J. Morrison, Daria R. Wonderlick, and Michael J. Harms. There are three jupyter notebooks (Fig-3.ipynb, Fig-4.ipynb, and Fig-5-6-SI.ipynb), corresponding to the display figures in the manuscript. The preprint directory contains jupyter notebooks corresponding to the figures in the preprint. The rosetta directory contains the ROSETTA ddG calculations needed for figures 5, 6, and SI.

ROSETTA calculations

ROSETTA calculations were run using linux build 2018.33.60351. python scripts were run using python 3.7.3. Scripts use pandas and numpy as dependencies.

• 00_initial-structures/: cleaned up and standardized sturctures for the apo, ca, and ca-pep states.
• 01_relaxed-structures/: relaxed structures fed into the ddG calculation. These were relaxed using the rosetta/rosetta-scripts/relax-structure.sh script.
• 02_mut-files/: ROSETTA-formatted files for introducing the point mutants. The mut files for each state ( apo, ca, and ca-pep) are stored in individual .zip files.
• 03_ddg-results/: ROSETTA output for ddG of each mutant. The results for each state ( apo, ca, and ca-pep) are stored in individual .zip files. These were generated using the rosetta/rosetta-scripts/mutate.sh script. The extract.py script was used to extract thes reuslts into ddg-summary.csv.
• 04_cycles/: script for calculating thermodynamic cycles from mutants. Script takes a couple of hours to run and dumps out a large file. This file is output-cycles.txt as read in by Fig-5-6-SI.ipynb. To generate, navigate into the 04_cycles directory and run: python extract-from-rosetta.py ../03_ddg-results/ddg-summary.csv.

This repo does not contain the cluster-specific bash scripts we used to loop over structure/mut-file combinations to run mutate.sh.