The Gene Ontology knowledgebase in 2026
Suzi A Aleksander, James P Balhoff, Seth Carbon, J Michael Cherry, Dustin Ebert, Marc Feuermann et al. · 2025 · 168 citationsRead the paper
The Gene Ontology (GO) knowledgebase (https://geneontology.org) is a comprehensive resource describing the functions of genes. The GO knowledgebase is regularly updated and improved. We describe here the major updates that have been made in the past 3 years. The ontology and annotations have been expanded and revised, particularly in several areas of biology: cellular metabolism, multi-organism interactions (e.g. host-pathogen), extracellular matrix proteins, chromatin remodeling (e.g. the "histone code"), and noncoding RNA functions. We have released version 2 of a comprehensive set of integrated, reviewed annotations for human genes, which we call the "functionome." We have also dramatically increased the number of GO-CAM models, with over 1500 models of metabolic and signaling pathways, primarily in human, mouse, budding and fission yeast, and fruit fly. Finally, we discuss our current recommendations and future prospects of AI in the use and development of GO.
What the full text shows
Deployed in practice- Net increase in experimentally supported annotations. — 12% increase (1,135,364 total vs 996,903 in Nov 2022) (As of July 2025)
- Expansion of GO-CAM causal pathway models. — 1,571 pathways (nearly five-fold increase) (As of July 2025; primarily human, mouse, yeast, and fruit fly)
- Human protein-coding gene coverage in the PAN-GO functionome (v2.0). — 84% of human genes (Version 2.0 of the human functionome)
- Human genes with annotations in all three GO aspects (MF, BP, CC). — 47% of genes (5% increase over v1.0) (Version 2.0 of the human functionome)
- Ontology term updates. — 768 terms added (~2% increase) and 4,173 terms removed (~10% decrease) (Over the past 3 years)
- Alignment of catalytic activity branch with external resources. — Over 9,400 RHEA and EC terms mapped to GO (Current version)
- Setup
- Knowledgebase covering 144 organisms in PANTHER gene families; data compared between releases 2022-11-03 and 2025-07-22.
- How
- The GO Consortium performed top-down refactoring of the ontology, manual curation of experimental data from ~15,000 publications, and phylogenetic annotation using the PAN-GO approach. Causal pathways were modeled using the Noctua Visual Pathway Editor to create GO-CAMs.
- Limitations
- PAN-GO coverage is substantially higher for vertebrate genomes than other genomes. · Non-human functionomes lack a dedicated browsing website and must be downloaded. · Reactome-derived GO-CAM models are not yet publicly available.
Summarized by Gemma 4 from the open-access full text · 9 Oct 2026. Read it