Abstract
ABSTRACT
Rapid and accurate identification of bacterial pathogens is critical for effective disease control in pig farms. Conventional bacterial culture-based and molecular diagnostic methods are time-consuming and often impractical for clinical decision-making. Raman spectroscopy, combined with artificial intelligence, offers a label-free alternative and provides isolate-level identification within approximately 2 min. The aim of this study was to evaluate the performance of the commercial Artificial Intelligent Raman Microbial Analyzer for Biology (AIRMAB) for the rapid clinical identification of swine bacterial pathogens. Bacterial isolates were collected from various tissues of diseased pigs between March and November 2025 at the Animal Disease Diagnostic Center, National Pingtung University of Science and Technology, Taiwan. The species-level identification of all the isolates was confirmed using species-specific polymerase chain reaction assays prior to Raman spectral acquisition, ensuring the accuracy of the reference labels. Overall, the AIRMAB system exhibited high performance in the identification of all 14 clinically relevant pig bacterial pathogens. Notably, perfect recall (100%) was achieved for clinically common species in the test set, including
Pasteurella multocida
(
n
= 30) and
Salmonella enterica
serovar Choleraesuis (
n
= 20). In addition, high recall (>90%) was observed for other important bacteria in the test set, including
Escherichia coli
(97%) (
n
= 33),
Streptococcus suis
(94%) (
n
= 35), and
Glaesserella parasuis
(93%) (
n
= 30). This study reports that the AIRMAB system provides rapid, accurate, and label-free identification of swine bacterial pathogens, delivering results within 2 min per isolate, and demonstrating practical value for timely decision-making in routine veterinary diagnostics, especially for common, and clinically important bacterial species.
IMPORTANCE
Bacterial infections pose a major challenge to global swine production, leading to growth retardation, elevated mortality, and substantial production losses. As several swine pathogens are zoonotic, they also pose direct and indirect risks to human health. Traditional diagnostic methods such as polymerase chain reaction are reliable but time-consuming and require well-equipped laboratories. This study introduces the Artificial Intelligent Raman Microbial Analyzer for Biology system, which utilizes Raman spectroscopy and deep learning to identify 14 major swine pathogens in approximately 2 min. This technology offers a highly automated, cost-effective alternative to traditional methods and matrix-assisted laser desorption ionization-time of flight mass spectrometry, facilitating rapid clinical intervention and large-scale screening to safeguard both animal and public health.