Fluorescent Dual-Signal Immunosensor Combined with Machine Learning Algorithm for Accuracy Discrimination of Natural Brucella Infection and Vaccination

Brucellosis is a worldwide zoonoses posing a significant threat to humans and animals. The difficulty in distinguishing between Brucella infection and vaccination has led to substantial health risks and economic losses. In this study, we developed a fluorescent dual-signal immunosensor, integrating it with machine learning to solve this problem. Immunogens were selected through structural analysis, followed by the screening of monoclonal antibodies. The antibodies were conjugated with fluorescent microspheres, serving as immunoprobes. Employing a dual-antibody sandwich strategy, we prepared the immunosensor and utilized machine learning to analyze the correlation between multibiomarkers and the status of Brucella infection or vaccine-induced immunity. The immunosensor achieved detection limits of 7.56 pg/mL for IL-1β and 76.61 pg/mL for IL-1Ra, without cross-reactivity with common inflammatory factors. Utilizing the Random Forest algorithm, the combination of the IL-1Ra/IL-1β ratio and RBPT achieved a F1 score of 1.0. Overall, the strategy effectively distinguishes between Brucella-infected and vaccinated subjects, offering valuable applications for brucellosis monitoring.