Precision Discrimination of Microbial Phenotypes by Integrating Biosynthetic AgNPs and Machine Learning Algorithms

Accurate and rapid identification of the microbial phenotype is pivotal for public safety and health. However, the traditional methods for microbial taxonomic discrimination are limited by their time-consuming nature and the requirement for specialized personnel. Here, we developed a biosynthetic strategy capable of identifying microorganisms at multiple taxonomic levels (i.e., kingdoms, phyla, orders, genera, and species) without biochemical assays or genetic amplification. Silver nanoparticles (AgNPs) biosynthesized by different microorganisms display distinct localized surface plasmon resonance (LSPR) spectra, zeta potential, and nanoparticle diameter. The correlation between the characteristics of AgNPs and microbial phenotypes is constructed by using machine learning algorithms. Notably, external stimuli were exploited to amplify differences among undifferentiated microorganisms, thus enabling the accurate differentiation of species-level microbes. This study not only offers a universal sensing tactic for accurate discrimination of microorganisms at different taxonomic levels but also opens an avenue for the application of biosynthetic nanomaterials in point-of-care screening of biomarkers.