ABSTRACT
Effective identification of multiple cariogenic bacteria in saliva samples is important for oral disease prevention and treatment. Here, a simple colorimetric sensor array is developed for the identification of cariogenic bacteria using single-atom nanozymes (SANs) assisted by machine learning. Interestingly, cariogenic bacteria can increase oxidase-like activity of iron (Fe)ânitrogen (N)âcarbon (C) SANs by accelerating electron transfer, and inversely reduce the activity of FeâNâC further reconstruction with urea. Through machine-learning-assisted sensor array, colorimetric responses are developed as "fingerprints" of cariogenic bacteria. Multiple cariogenic bacteria can be well distinguished by linear discriminant analysis and bacteria at different genera can also be distinguished by hierarchical cluster analysis. Furthermore, colorimetric sensor array has demonstrated excellent performance for the identification of mixed cariogenic bacteria in artificial saliva samples. In view of convenience, precise, and high-throughput discrimination, the developed colorimetric sensor array based on SANs assisted by machine learning, has great potential for the identification of oral cariogenic bacteria so as to serve for oral disease prevention and treatment.