Simplify Gene-Set Enrichment Analysis • geneSetSimplifyR

Gene-set enrichment analysis often returns hundreds or thousands of significant gene sets, far too many to read one by one. geneSetSimplifyR groups them into a small number of clusters of overlapping gene sets and gives each cluster an informative two-word label.

Clustering. Gene sets are compared by the genes they share, using latent semantic indexing and a shared nearest neighbor graph. Several community-detection methods are available, and you can plug in your own.
Labeling. Each cluster is named after the word pair that best describes its gene-set names (TF-IDF weighted by how central the gene set is).
Visualization. UMAP, bar chart, lollipop, violin, heatmap, dendrogram and clustering-tree plots, with support for results from several conditions or analyses.

It works on the output of any enrichment tool, with helpers for fgsea and pairedGSEA results.

Installation

if (!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")
}
BiocManager::install("remotes") # needed to install from GitHub
BiocManager::install("IsoformAnalysisGroup/geneSetSimplifyR")

BiocManager installs the dependencies from both CRAN and Bioconductor.

Platform notes

Linux: igraph needs the GLPK library. On Debian/Ubuntu, run sudo apt-get install libglpk40 first. For other missing system libraries, pak::pkg_install("IsoformAnalysisGroup/geneSetSimplifyR") installs the package together with the system libraries it needs.
macOS: if installation stops with ERROR: sub-architecture '...' is not installed, add INSTALL_opts = "--no-multiarch" to the BiocManager::install() call.

Quick start

library(geneSetSimplifyR)

# A named list of gene sets and, optionally, a table of enrichment results
# with the columns 'pathway', 'enrichment_score' (or 'NES'), and optionally
# 'pval' and 'source'.
gsList <- geneSetSimplifyR(
  geneSetsList = exampleGeneSets,
  geneSetsDF = exampleEnrichment
)

gsList
plotUMAP(gsList)
plotEnrichments(gsList)

# One row per gene set with its cluster label, for supplementary tables
extractTable(gsList, summarize = FALSE)

Documentation

The full documentation, including the vignette and all help pages, is online at https://isoformanalysisgroup.github.io/geneSetSimplifyR/. The vignette walks through the whole workflow, from fgsea results to renaming clusters: Get started.

GitHub installs skip the vignette by default. To also read it offline in R, install with build_vignettes = TRUE. This needs a few extra packages and pandoc, which comes with RStudio:

BiocManager::install(c("knitr", "rmarkdown", "BiocStyle"))
BiocManager::install(
  "IsoformAnalysisGroup/geneSetSimplifyR",
  build_vignettes = TRUE,
  force = TRUE # reinstall if the package is already installed
)
vignette("geneSetSimplifyR", package = "geneSetSimplifyR")

Report bugs and questions at https://github.com/IsoformAnalysisGroup/geneSetSimplifyR/issues.