Advanced Inference with Random Graphical Models.

rgm: Random Graphical Models for data from multiple environments

rgm is an R package that implements state-of-the-art Random Graphical Models (RGMs) for the analysis of complex multivariate data. It is able to handle heterogeneous data across various environments, offering a powerful tool for exploring intricate network interactions and structural relationships.

Key Features

• Joint Inference Across Multiple Environments: rgm enables simultaneous analysis of multivariate data from diverse environments, providing a comprehensive understanding of complex network interactions.
• Random Graphical Modeling: The package includes a generative model of graphs across environments to handle heterogeneity and quantify structural relationships across environments.
• Integration of External Covariates: Users can incorporate external covariates at both node and interaction levels, allowing for a more complete analysis of network data.
• Bayesian Framework: rgm uses a Bayesian approach to quantify parameter uncertainty, including uncertainty on the inferred graphs.

Installation

Install the latest version of rgm from GitHub using the following commands in R:

install.packages("devtools")
devtools::install_github("franciscorichter/rgm", build_vignette=TRUE)

Usage

For detailed instructions on using rgm for data analysis, refer to the package vignette and documentation:

library(rgm)
vignette("rgm")

Note: While initially designed for microbiome analysis, rgm is broadly applicable across various fields requiring advanced graphical modeling of multivariate data from multiple environments.

Principal Reference

The methodologies implemented in the rgm package are principally derived from the work described in Vinciotti, V., Wit, E., & Richter, F. (2023). "Random Graphical Model of Microbiome Interactions in Related Environments." arXiv preprint arXiv:2304.01956.