Machine Learning-Assisted Portable Dual-Readout Biosensor for Visual Detection of Milk Allergen

Abstract

Beta-lactoglobulin (β-LG), the primary allergen in cow’s milk, makes developing a rapid, sensitive, and convenient detection method essential for individuals with allergies. In this study, a graphdiyne-based self-powered electrochemical biosensor has been cleverly integrated into the corresponding test strip. This biosensor uses glucose as fuel and correlates the β-LG concentration with the glucose value displayed on a mobile phone application, enabling real-time and quantitative detection. Additionally, an electrochromic substance reacts with the byproduct (H₂O₂) of glucose oxidation by biological enzymes. A quantitative relationship between color and β-LG concentration has been established using mobile phone software. Dual detection of electrochemical and colorimetric signals in the 0.01–10,000 ng/mL range, with detection limits as low as 0.0033 and 0.0081 ng/mL, is possible. Machine learning is finally employed to analyze its performance. Our dual-readout biosensor demonstrates significant potential for rapid food allergy detection.