en
Why is AI being used in microbiology?
Diverse technological advances, including genomic sequencing, metagenomics, digital imaging and automated clinical data collection have greatly increased the availability of microbiological data. From studying the human gut microbiome to improving clinical diagnostics, scientists now have unprecedented opportunities to better understand microorganisms [1].
However, this wealth of information presents a new challenge: how can these vast, complex datasets be analysed and interpreted at scale?
To address this, the use of artificial intelligence (AI) is being evaluated in a range of microbiology applications. Machine learning (ML) models can help interpret sequencing data, recognise patterns, classify microorganisms, predict outcomes, and generate new research hypotheses [1]. While the field is still evolving and the role of the laboratory scientist remains essential for validating findings and interpreting results, AI technologies are beginning to change how microbiological data is analysed, helping researchers focus more of their time on interpretation, experimental design and higher-level clinical decision-making.
Drawing on recent published studies, we explore how AI is currently used across microbiology research, from the gut microbiome to clinical microbiology, and consider what the future may hold for this rapidly evolving field.
How is AI helping researchers understand the gut microbiome?
Whilst AI is already being applied to areas in gut microbiome research such as microbiome profiling, analysing DNA sequencing data, and classifying unknown organisms [1], another exciting area is understanding microbe-host interactions: the relationship between microbial communities and their impact on human health.
One example of this is the use of ML in differentiating inflammatory bowel disease (IBS) subtypes, including ulcerative colitis and Crohn’s disease [2]. Huang et al. (2025) conducted a systematic review of studies that used ML models to support the classification of these conditions. After identifying 31 suitable studies, the review found that approaches including random forest and support vector machine models have the potential to improve diagnostic accuracy when distinguishing between ulcerative colitis and Crohn’s disease. The findings highlight the potential of ML approaches to support more personalised treatment strategies and improve patient outcomes. However, further validation using larger and more diverse datasets remains essential before these approaches enter routine clinical practice.
As well as analysing relationships between microbial communities and disease, AI is increasingly being explored for precision nutrition, helping researchers understand why individuals can respond differently to the same foods. By combining data from microbiome composition, blood markers, dietary information and metabolic responses, ML models can identify patterns that may support more personalised nutritional recommendations [3].
For example, the PREDICT programme, a series of clinical trials involving researchers from Massachusetts General Hospital, King’s College London, Stanford Medicine and Harvard T.H. Chan School of Public Health, has used ML approaches to investigate individual variations in responses to food and develop a better understanding of personalised nutrition [3]. The PREDICT-1 study explored differences in post-meal glucose, insulin and triglyceride responses following standardised meals, considering factors including genetics, gut microbiome composition, meal composition, age, sex and BMI. The study demonstrated that genetic factors alone could not fully explain variation in metabolic responses, with the gut microbiome and other lifestyle factors providing additional insights. By combining these multiple data sources, ML models were able to predict individual post-meal glycaemic and triglyceride responses more accurately, demonstrating the potential of AI to support future personalised nutrition approaches.
How is AI being used in clinical microbiology and diagnostics?
AI is also helping to advance the field of clinical microbiology as laboratories increasingly explore the potential of digital microbiology. Digital microbiology combines traditional microbiological methods with automation, imaging technologies, artificial intelligence and data analysis to improve efficiency, support interpretation and reduce the time required for certain manual processes [1].
One emerging application is the use of AI to support faster pathogen identification. AI approaches are being explored alongside technologies such as MALDI-TOF mass spectrometry and whole genome sequencing (WGS) [1], helping researchers analyse complex datasets to improve microbial classification, identify unusual organisms and better understand transmission patterns during outbreaks.
By integrating with existing microbiology workflows, AI has the potential to enhance interpretation and support faster decision-making, rather than replacing the expertise of microbiologists. This approach is also central to COPAN’s development of PhenoMATRIX®, an AI-powered software platform designed to support the interpretation of bacterial culture plates within clinical laboratories. PhenoMATRIX® combined AI image analysis with laboratory information system (LIS) data to automatically analyse and sort digital images of culture media plates, helping microbiologists identify relevant growth patterns and prioritise samples requiring further investigation [4]. The software can detect microbial growth, estimate colony counts and differentiate isolates based on phenotypic colony characteristics before applying laboratory-defined classification rules. The technology is now FDA 510(k) cleared (K251511) as a Class II medical device [4], representing an important step in the adoption of AI-supported digital microbiology workflows. By reducing the burden of repetitive screening tasks, AI tools such as PhenoMATRIX® aim to allow microbiologists to dedicate more time to complex interpretation, clinical decision-making and applying their specialist expertise where it adds the greatest value.
Another promising application of AI in clinical microbiology is antimicrobial resistance (AMR) prediction. Traditionally, determining whether a bacterium is resistant to a particular antibiotic requires culturing the organism and performing susceptibility testing, a process that can take several days. AI models are being explored to analyse a combination of genetic data, phenotypic characteristics and clinical information to predict resistance profiles more rapidly [5]. While these approaches could support earlier targeted treatment decisions and improved antibiotic stewardship, laboratory testing and microbiological expertise remain essential to validate AI-generated predictions.
What are the challenges and future possibilities of AI in microbiology?
As AI continues to develop, these uses are likely to become increasingly integrated into laboratories, while new potential applications across microbiology are likely to emerge. From improving our understanding of microbial communities to supporting faster diagnostics and enabling new approaches to disease prediction and treatment, AI could become an invaluable tool across research and clinical workflows.
Whilst the possibilities are exciting, challenges such as inconsistent datasets, model transparency and the need for further validation highlight why AI should be viewed as a tool to support, rather than replace, microbiological expertise. Laboratory validation, experimental research and expert interpretation remain essential to ensure AI-generated insights within studies are accurate, reliable and meaningful when applied to microbiology studies.
References:
- Wang, XW, Wang, T, Liu, YY. Artificial Intelligence for Microbiology and Microbiome Research. Cell Systems. 2026 Feb;17(2).
- Huang, J, Zhu, X, Ma, Y, et al. Machine learning in the differential diagnosis of ulcerative colitis and Crohn’s disease: a systematic review. Transl Gastroenterol Hepatol. 2025 Jul; 10(56).
- Berry, SE, Valdes, AM, Drew, DA. Human Postprandial Responses to Food and Potential for Precision Nutrition. Nat Med. 2020 June;26(6).
- COPAN. PhenoMatrix. [Online]. Murrieta: Copan Diagnostics Inc; [Accessed 21 July 2026]. Available from: https://www.copanusa.com/laboratory-automation/microbiology-laboratory-automation-ai/phenomatrix/
- Ren, Y, Chakraborty, T, Doijad, S, et al. Prediction of antimicrobial resistance based on whole-genome sequencing and machine learning. Bioinformatics. 2022 Jan;38(2).
England