mutualinfobio is an R package for computing mutual information (MI) and coevolutionary scores from protein multiple sequence alignments (MSAs), with built-in support for FASTA and plain-text input formats. The package includes visualization tools such as MI heatmaps.
To install the development version directly from GitHub:
install.packages("remotes") # if not already installed
remotes::install_github("TraxlerLab/mutualinfobio")library(mutualinfobio)
# Load an example alignment from the Cell 2008 paper by Weigt et al.
path <- system.file("extdata", "cell3925mmc4.fasta", package = "mutualinfobio")
# Construct an Alignment object
aln <- Alignment(alignment_path = path)
# Compute mutual information-based covariance scores
mi_scores <- getMICovarianceScore(aln, cores = 8)
# Visualize as a heatmap
# Visualize as a heatmap
draw_MI_heatmap(
data = mi_scores,
alignment_labels = c("HK", "RR"),
alignment_label_locations = c(50, 200),
alignment_domain_residues = c(100, 308),
heatmap_lines = FALSE
)✨ The file
cell3925mmc4.fastais included as example data under a Creative Commons Attribution (CC BY) license, from the supplementary material of Wkerker det al. (2008), Cell. DOI: 10.1016/j.cell.2008.04.040
- Compute MI-based covariation from aligned sequences
- Heatmap visualizations of MI scores
- Support for FASTA and plain text alignments
- Parallel computation support in UNIX systems(with CRAN-safe defaults)
- Cross-Platform (Parallel computation not supported in Windows)
This work was based on the methods described by Skerker et. al:
Skerker, J.M., Perchuk, B.S., Siryaporn, A., Lubin, E.A., Ashenberg, O., Goulian, M., & Laub, M.T. (2008).
Rewiring the specificity of two-component signal transduction systems. Cell, 133(6), 1043–1054.
https://doi.org/10.1016/j.cell.2008.04.040
This package is released under the MIT License. See LICENSE file for details.
The included alignment file (cell3925mmc4.fasta) is licensed separately under CC BY and is used with attribution.