Grigson, S.R.Bouras, G.Dutilh, B.E.Olson, R.D.Edwards, R.A.van der Meer, J.R.2025-10-222025-10-222025Microbiology and Molecular Biology Reviews, 2025; 89(3):e0002225-1-e0002225-301092-21721098-5557https://hdl.handle.net/2440/147928Understanding protein functions is crucial for interpreting microbial life; however, reliable function annotation remains a major challenge in computational biology. Despite significant advances in bioinformatics methods, ~30% of all bacterial and ~65% of all bacteriophage (phage) protein sequences cannot be confidently annotated. In this review, we examine state-of-the-art bioinformatics tools and methodologies for annotating bacterial and phage proteins, particularly those of unknown or poorly characterized function. We describe the process of identifying protein-coding regions and the systems to classify protein functionalities. Additionally, we explore a range of protein annotation methods, from traditional homology-based methods to cutting-edge machine learning models. In doing so, we provide a toolbox for confidently annotating previously unknown bacterial and phage proteins, advancing the discovery of novel functions and our understanding of microbial systems.en© 2025 American Society for Microbiology. All Rights Reserved.bioinformatics; function prediction; microbial proteins; machine learningBacteriaBacteriophagesBacterial ProteinsViral ProteinsComputational BiologyMolecular Sequence AnnotationMachine LearningComputational function prediction of bacteria and phage proteinsJournal article10.1128/mmbr.00022-25858542Bouras, G. [0000-0002-5885-4186]