Ontologizer 3: a cross-platform desktop application for frequentist and Bayesian GO enrichment analysis

Abstract

We present Ontologizer 3, an easy-to-use cross-platform desktop application for Gene Ontology (GO) overrepresentation analysis. Ontologizer 3 offers two complementary methods. The first is a frequentist approach that evaluates GO terms individually using a one-sided Fishers exact test, yielding term-level significance values. The second is a Bayesian approach that jointly evaluates terms using model-based gene set analysis (MGSA), yielding term-level posterior probabilities. Due to substantial overlap among annotated gene sets resulting from the GOs hierarchical structure, the two methods produce different results. The frequentist approach tends to report a lengthy list of terms with strong annotation overlap, whereas, MGSA yields a parsimonious set of terms that explain the observed gene activity. Using simulated data with a known ground truth, we demonstrate that both methods reliably identify the causal term, but MGSA achieves substantially higher precision. Ontologizer 3 is implemented as a Tauri application with a Rust backend and an Angular frontend, presenting enriched terms in tabular and graphical form. The software is freely available under the MIT licence at https://github.com/P2GX/ontologizer-gui. Installation packages for Macintosh, Windows and Debian-based Linux are available on the GitHub Releases page.