New strains of microbes have been regularly discovered and brought into a laboratory for study in many different directions. When a new microbial genome is sequenced and assembled, the homologues in other microbes are generally used to annotate the new genome. However, some genes or open reading frames may not have sequence similarity to any other microbial sequences. Knowledge of their protein structure might suggest the function of the new unknown gene. Until recently, two homology modelings, Swiss-Model and Modeller, were taught in our Bioinformatics course. Remarkably breakthrough of protein structure prediction with AI called AlphaFold2 from DeepMind has advanced the field, even though there are some limitations. Therefore, study protein structure modeling for postgraduate study in Microbiology has changed to focus on automatic homology modeling with Swiss-Model webserver and protein structure model predicted with AlphaFold2. Students will learn how to obtain structure models from an amino acid sequence and visualize their secondary structures, parts with hydrophobicity and electrical charge, and conserved regions, which can suggest their function. These two practical labo