Synthetic coevolution reveals adaptive mutational trajectories of neutralizing antibodies and SARS-CoV-2

Implementation of the paper [Synthetic coevolution reveals adaptive mutational trajectories of neutralizing antibodies and SARS-CoV-2], by Roy Ehling^* , Mason Minot^*, Max Overath, Daniel Sheward, Jiami Han, Beichen Gao, Joseph Taft, Margarita Pertseva, Cedric Weber, Lester Frei, Thomas Bikias, Ben Murrell, and Sai Reddy.

Table of contents

1. Environment
2. Reproducing Study Results
3. Citation

Environment

Note: analysis run with torch 2.1.2+cu121, but environment contains torch 2.1.2. Find the correct torch 2.1.2 for your OS/torch/cuda combination here https://pytorch.org/get-started/previous-versions/#v212

Conda

cd envs
conda env create -f syn_coev.yml
conda activate syn_coev

venv

1. python -m venv syn_coev
2. Windows: syn_coev\Scripts\activate.bat Unix / MacOS: source syn_coev/bin/activate
3. in envs/, run pip install -r requirements.txt

Reproducing Study Results

1. Preprocessing.
2. Model training and evaluation.
3. Plot results.

Step 1 - Preprocessing

Note: Data will be made available following publication. In src/scripts/ run preprocessing.sh.

Step 2 - Model Training and Evaluation

In src/scripts/ run train.sh. This will populate the folder results/ with .csv files in the appropriate format for plotting in Step 3.

Note: to execute DDP training runs via slurm, an example is provided in src/example_slurm_submission.slurm.

Setp 3 - Plot Results

In src/scripts/ run plot.sh.

Citation

@article {Ehling2024.03.28.587189,
	author = {Roy A. Ehling, Mason Minot, Max D. Overath, Daniel J. Sheward, Jiami Han, Beichen Gao, Joseph M. Taft, Margarita Pertseva, C{\'e}dric R. Weber, Lester Frei, Thomas Bikias, Ben Murrell, and Sai T. Reddy},
	title = {Synthetic coevolution reveals adaptive mutational trajectories of neutralizing antibodies and SARS-CoV-2},
	year = {2024},
	doi = {10.1101/2024.03.28.587189},
	publisher = {Cold Spring Harbor Laboratory},
	journal = {bioRxiv}
}