Artificial intelligence-assisted microsphere immunoassay for ultrasensitive Salmonella detection through microfluidic translocation event counting.

Salmonella pose significant threats to public health, necessitating rapid and sensitive detection strategies for effective food safety monitoring. Here, we present an artificial intelligence-assisted microsphere immunoassay (AIMI) that converts pathogen recognition into an ultrasensitive event level quantitative readout through microsphere-based microfluidic analysis. In this system, antibody-coated magnetic nanoparticles (MNPs) and polystyrene beads (PS) form sandwich immunocomplexes with Salmonella. These complexes are selectively depleted by magnetic separation, leaving residual unbound PS beads whose number inversely correlates with target concentration. When introduced into a microfluidic channel, these detection microspheres generate flow regulated signals that can be resolved at the event level and characterized by microsphere number and translocation dynamics. In AIMI, pulsed low-velocity transport refers to transient microsphere movement generated by briefly activating the syringe pump for 1 s under a preset flow rate of 1.5 μL min-1. Motion based correction further compensates for flow induced displacement variations, corrects motion related counting bias, and improves quantitative accuracy. With displacement informed correction, the linearity of detection improves from 0.983 to 0.99. The assay achieves a broad linear detection range of 10-106 CFU mL-1 and a limit of detection of 2.51 CFU mL-1. As a result, this AIMI platform achieves accurate translocation event counting at the single microsphere level, providing an effective strategy for ultrasensitive detection of Salmonella contamination.