Bibliometric analysis of the application of artificial intelligence techniques in bacteriology: a PRISMA-guided research agenda.

Introduction: The integration of artificial intelligence (AI) into bacteriology has marked a pivotal advancement by enabling the analysis of large-scale microbiological datasets. Despite growing adoption, significant research gaps persist, hindering the full exploitation of AI's potential in bacterial research and diagnostics. Objective: To analyze global scientific production on the application of AI techniques in bacteriology and propose a future research agenda based on bibliometric trends. Methods: This study conducts a bibliometric analysis of artificial intelligence (AI) applications in bacteriology, explicitly guided by the PRISMA 2020 framework. Unlike traditional reviews, this approach combines PRISMA's methodological rigor with bibliometric techniques to map scientific production. Metadata were retrieved from Scopus and Web of Science using predefined search strategies. Quantitative indicators, co-occurrence networks, and thematic mapping were applied to examine the field's temporal evolution and conceptual structure. The findings provide an evidence-based overview of research trends and gaps, supporting the design of a future research agenda on AI integration in bacteriology. Results: The findings reveal exponential growth in scientific output, especially between 2022 and 2024. Leading authors include Singh and Waegeman, with high-impact journals such as Frontiers in Microbiology and MSystems. The United States and China are the most productive countries. Thematic evolution shows a shift from early topics like microbial spoilage toward advanced applications including bacterial classification and diagnostic modeling. Key conceptual clusters were identified around microbiomes, classification, and bioinformatics. Emerging terms such as "diagnosis," "metagenomics," and "transfer learning" indicate future research directions. Conclusion: AI applications in bacteriology are expanding rapidly yet still rely heavily on traditional machine learning methods. There is a need to incorporate advanced approaches such as deep learning and transformer-based models. The findings support a strategic agenda for promoting interdisciplinary collaboration and technological innovation in bacteriological research.