Aptamer-based sensing array coupled with multiple pattern recognition algorithms for accurate β-lactoglobulin...

The β-Lactoglobulin (β-LG), a major milk allergen, contributes to rising health concerns, especially in infants and children. Although rapid detection methods have been reported, the complex matrix effects in food samples pose a crucial challenge to their practical application, demanding more available and effective β-LG detection strategies.

In this study, we introduced a novel strategy of aptamer-based sensing array coupled with multiple pattern recognition algorithms for accurate β-LG detection by circumventing the challenge of the complex matrix effects. Four aptamers were selected as the recognition probes, while fifteen samples were chosen as the pattern recognition model, with the color change of AuNPs serving as the response signal. In the AuNP colorimetric array, β-LG-spiked sample matrices induced red-to-blue color transitions by modulating salt-induced AuNP aggregation, due to variable protective effects from aptamers and potential cross-reactivity with matrix. The difference in colorimetric responses between spiked and unspiked samples was considered as the unique fingerprint pattern for β-LG in each matrix, followed by multiple pattern recognition algorithms, for better orthogonality to complete β-LG identification and prediction of the unknown samples. Based on 1500 responses (4 aptamers × 15 matrices × 5 replicates × 5 concentrations), the three algorithms collectively achieved over 95% classification accuracy and established a minimum discriminable concentration of 20 nM for reliable pattern-recognition-based differentiation of β-LG across diverse food matrices. Linear discriminant analysis achieved 95.3% accuracy in blind testing of 150 samples.

In contrast to conventional single "lock-and-key" aptasensors, the proposed strategy offers a platform with high sensitivity and robust anti-interference performance, enabling reliable allergen detection in complex food matrices.