Ontologizer 3：跨平台桌面应用，支持频率学派与贝叶斯 GO 富集分析
Ontologizer 3: a cross-platform desktop application for frequentist and Bayesian GO enrichment analysis
Ontologizer 3 是一款跨平台桌面应用，用于 Gene Ontology（GO）过表征分析，提供频率学派与贝叶斯两类互补方法。频率学派方法用单侧 Fisher 精确检验逐个评估 GO 术语，给出术语水平的显著性；贝叶斯方法用基于模型的基因集分析（MGSA）联合评估术语，给出术语水平的后验概率。由于 GO 层级结构导致注释基因集大量重叠，两种方法结果不同：频率学派倾向输出一长串注释高度重叠的术语，MGSA 则给出解释观测基因活性的简约术语集。在已知真值的模拟数据上，两种方法都能可靠识别因果术语，但 MGSA 精度明显更高。
来源：NAR Genomics and Bioinformatics · doi.org
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 Fisher’s 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 GO’s 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.