Disk Diffusion Assay plate. (Photo credit: Dr. Pierangeli Vital)MANILA — Three researchers from the University of the Philippines (UP) have published a study on the use of artificial intelligence in "analyzing antimicrobial resistance," which they believed may be utilized in certain agricultural environment.

E.coli found in some Metro Manila vegetables: UP study

The Food and Agriculture Organization of the United Nations defines antimicrobial resistance (AMR) as "the ability of microorganisms to persist or grow in the presence of drugs designed to inhibit or kill them." 

The Australian government explained that AMR can be a threat to agriculture and food production since "infections in animals can be a risk to humans through contact and in the food chain."

"Traditional laboratory methods for analyzing antimicrobial resistance are often time-consuming and labor-intensive, making them impractical for large-scale monitoring. As a result, researchers are exploring faster approaches using whole-genome sequencing (WGS) and predictive modeling," according to Eunice Jean Patron of UP Diliman College of Science.

This dilemma prompted Marco Christopher Lopez and Dr. Pierangeli Vital of the UPD-CS Natural Sciences Research Institute, along with Dr. Joseph Ryan Lansangan of the UPD School of Statistics, to embark on a study, titled "Prediction models for antimicrobial resistance of Escherichia coli in an agricultural setting around Metro Manila, Philippines."

The trio tested various AI prediction models "to determine the antimicrobial resistance of E. coli using genetic data and laboratory test results from the National Center for Biotechnology Information (NCBI) database."

The AI models used were the following: 

• Random Forest (RF) — well-suited for high-dimensional data

• Support Vector Machine (SVM) — excels in classification tasks, particularly when dealing with complex decision boundaries

• Two ensemble methods: Adaptive Boosting (AB) and Extreme Gradient Boosting (XGB), which enhance accuracy by focusing on hard-to-classify samples

"These AI prediction models most accurately predicted resistance to streptomycin and tetracycline, showing high accuracy and reliably distinguishing resistant strains from susceptible ones. On the other hand, ciprofloxacin was the most challenging to predict due to the limited number of resistant samples in the data," they said.

"Among the models, AB and XGB consistently delivered good results, even when tested on imbalanced antimicrobial resistance data," they added.

Vidal said their strategy "has great potential for real-time monitoring of antimicrobial resistance, particularly in agriculture."

“As DNA sequencing becomes faster and cheaper, prediction models such as ours can pick up resistant bacteria early—before they lead to outbreaks. This can facilitate better decision-making in food safety, agriculture, and public health programs," the researcher added.

Their study was published in the Malaysian Journal of Microbiology, an open-access, peer-reviewed journal.

RELATED VIDEO