The use of big data and machine learning (ML) for antimicrobial resistance (AMR) surveillance in livestock production shows promise for informing AMR mitigation efforts, according to a recent study led by University of Nottingham researchers.

Working with academic, state, and industry collaborators from China and the UK, University of Nottingham researchers collected samples and analyzed the microbiomes from live chickens, chicken carcasses, and environments across ten large-scale commercial poultry farms in three Chinese provinces. China was chosen as the location for the study due to the nation being the largest global consumer of antimicrobials, and the level of antimicrobial use (AMU) in poultry production being five times higher in China than the international average.

Factors such as changes in diet, temperature, and stress may result in the colonization of new resident bacterial species or AMR transfer between species. Temperature, humidity, and both bacterial species abundance and the presence of antibiotic resistance genes (ARGs) can also influence bacterial infection in broilers. In many countries, such as China, chickens are housed in sheds that do not have an effective climate control system, and therefore experience substantial temperature and humidity variations.

During the study, a total of 461 microbiomes were analyzed using a data mining approach based on ML, through which 145 potentially mobile ARGs, shared between chickens and environments across all farms, were identified. The researchers found a core subset of the chicken gut microbiome, featuring clinically relevant bacteria and antibiotic resistance genes that correlate with AMR profiles of Escherichia coli, colonizing the gut. This core, which contains clinically high transmissible ARGs shared by chickens and environments, is influenced by environmental temperature and humidity, and correlates with AMU.

The associations between environmental variables, and the species and genes associated with AMR, present opportunities for the development of novel AMR monitoring solutions, especially in low- and middle-income countries where these variables are not always controlled.

The next step is to consider all relevant and interconnected AMR datasets in a 360° approach to deepen understanding and control of AMR spread. The study’s authors believe that investment in new artificial intelligence (AI) –powered AMR integrated surveillance approaches to identify the drivers and the mechanisms underlying the rise and spread of AMR, as well as new genetic variants of resistant pathogens across animals, environment, humans, and food, is highly promising.