micInt is an R package designed for analyzing microbial co-occurences. It takes in OTU tables was either pure dataFrames or as experiment-level phyloseq objects. Roughly, the package is composed of three major parts:

• CoNet-based analysis of pairwise co-occurences. Primary functions: runAnalysis (high level) and ccrepe (low level)
• Lotka-Volterra modelling of time dynamics. Primary functions: integralSystem, cv.LV and ridge_fit
• Utility functions, includes refine_data, subset_by_environment, similarity_measures and scale_by_column.

This package have the following dependencies which must be installed prior to installing this package:

From CRAN:

• infotheo
• matrixStats
• magrittr
• deSolve
• igraph
• dplyr
• rlang
• glue
• viridis
• RhpcBLASctl
• ggplot2
• ggfortify

From Bioconductor:

• phyloseq

If all dependencies are satisfied, the package can we installed from GitHub as follows:

install.packages("devtools")
library(devtools)
install_github("AlmaasLab/micInt")

For a quit start automatically installing all dependencies, consider:

install.packages("BiocManager","devtools")
devtools::install_github("AlmaasLab/micInt",repos=BiocManager::BiocManager::repositories())

Please read the article "Robust bacterial co-occurence community structures are independent of r- and K-selection history" (https://www.nature.com/articles/s41598-021-03018-z) where this package is used. The source code for the article is available on https://github.com/yaccos/Microbial-co-occurence.

Unfortunately, I have not yet taken the effort to write a proper vignette. However, if you are interested in using the package, don`t hasitate asking me for help, either as a github issue or directly ([email protected]).