Movatterモバイル変換

rgm is an R package that implements state-of-the-artRandom Graphical Models (RGMs) for the analysis of complex multivariatedata. It is able to handle heterogeneous data across variousenvironments, offering a powerful tool for exploring intricate networkinteractions and structural relationships.

Key Features

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

Installation

Install the latest version ofrgm from GitHub using thefollowing commands in R:

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

Usage

For detailed instructions on usingrgm for dataanalysis, refer to the package vignette and documentation:

library(rgm)vignette("rgm")

Note: While initially designed for microbiomeanalysis,rgm is broadly applicable across various fieldsrequiring advanced graphical modeling of multivariate data from multipleenvironments.

Principal Reference

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