featuremap

featuremap is an R package for identifying and interpreting functional metabolic modules from untargeted metabolomics data. It connects metabolite annotation, metabolic network extraction, module identification, and optional literature- and LLM-assisted functional interpretation in one workflow.

Installation

Install BiocManager and remotes first:

if (!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}

if (!requireNamespace("remotes", quietly = TRUE)) {
  install.packages("remotes")
}

Install the upstream packages used by featuremap:

remotes::install_github("tidymass/masstools")
remotes::install_github("tidymass/massdataset")
remotes::install_github("tidymass/metid")
remotes::install_github("tidymass/pcsf2featuremap")

Then install featuremap from GitHub:

remotes::install_github(
  "tidymass/featuremap",
  dependencies = TRUE,
  repos = BiocManager::repositories(),
  upgrade = "ask"
)

library(featuremap)

Check the installed version:

packageVersion("featuremap")

Workflow

The main analysis follows four functions:

res <- annotate_feature(...)
res <- extract_subnetwork(res = res, ...)
res <- identify_module(res = res, ...)
res <- annotate_module(res = res, ...)

The final annotation step can optionally use PubMed and an LLM provider. API keys are not required for the first three functions.

Tutorial

For input formats, complete examples, output interpretation, and the Shiny workflow, see the FeatureMAP tutorial.

Getting help

Open the package help page in R with:

?featuremap

Issues and feature requests can be submitted through the FeatureMAP GitHub repository.