Machine learning assisted CRCA multi-color fluorescent droplet imaging for high-sensitivity simultaneous detection of Maillard hazards.

Acrylamide and advanced glycation end products are two typical hazards produced simultaneously through the Maillard reaction of food. Traditional instrumental detection methods require specialized instrument operation, resulting in low detection efficiency. Here, a red-green dual fluorescence detection system combined with Nt.BbvCI-assisted cyclic rolling circle amplification (CRCA) technology was designed to reduce the limit of detection to the pg/mL level. Furthermore, CRCA is encapsulated in emulsion droplets for the reaction. Especially when the target concentration is low, the fluorescence is concentrated in a small number of droplets, making the fluorescence signal easier to detect and greatly improving the detection sensitivity. The fluorescence imaging pictures were further analyzed by machine learning, and the analysis time was reduced from 3 min/picture to 1.4 s/picture, improving the analysis efficiency and accuracy, and the detection limit was lowered to 10 fg/mL, which ultimately achieved the simultaneous detection of Maillard hazards in food with high sensitivity.