Welcome to the GitHub repository for the following publication: MoCHI: neural networks to fit interpretable models and quantify energies, energetic couplings, epistasis and allostery from deep mutational scanning data (Faure AJ & Lehner B, 2024)
Here you'll find an R package with all scripts to reproduce the figures and results from the computational analyses described in the paper.
- 1. Required Software
- 2. Required Data
- 3. Installation Instructions
- 4. Usage
To run the mochims pipeline you will need the following software and associated packages:
- R (Cairo, bio3d, data.table, ggplot2, ggrepel, hexbin, plot3D, reshape2)
Fitness scores, inferred free energy changes and required miscellaneous files should be downloaded from here and unzipped in your project directory (see 'base_dir' option) i.e. where output files should be written.
Make sure you have git and conda installed and then run (expected install time <5min):
# Install dependencies manually (preferably in a fresh conda environment)
conda install -c conda-forge cairo r-base>4.0.0 r-bio3d r-cairo r-data.table r-devtools r-ggplot2 r-ggrepel r-hexbin r-plot3d r-reshape2 r-roxygen2
# Open an R session and install the mochims R package
devtools::install_github("lehner-lab/mochims")
The top-level function mochims() is the recommended entry point to the pipeline and by default reproduces the figures and results from the computational analyses described in the following publication: MoCHI: neural networks to fit interpretable models and quantify energies, energetic couplings, epistasis and allostery from deep mutational scanning data (Faure AJ & Lehner B, 2024). See Required Data for instructions on how to obtain all required data and miscellaneous files before running the pipeline. Expected run time <20min.
library(mochims)
mochims(base_dir = "MY_PROJECT_DIRECTORY")