Identification of plant pathogenic bacteria using Raman microspectroscopy coupled with a dielectrophoretic device

Phytopathogenic bacteria induce devastating effects on crops and their specific detection and identification are of utmost importance. Current diagnosis of bacteria is mainly based on serological and molecular methods, which are time-consuming and expensive. Rapid, accurate, reliable, and cheap diagnostic procedures for plant diseases and correct identification of etiological agents are needed, in order to guarantee high quality and quantity of agricultural products and food. Among these innovative technologies, Raman spectroscopy (RS) is gathering considerable attention. RS provides a direct and non-destructive analysis of the biochemical components present in a matrix, including bacterial cells . In this work, a dielectrophoretic (DEP) device was used to increase bacterial concentration and amplify the intensity of the Raman spectroscopic signal. Using dedicated Raman-DEP platform, a dataset of spectra from different isolates of harmful phytopathogenic bacteria belonging to the genera Pseudomonas, Xanthomonas, and Erwinia was obtained. With regards to Pseudomonas, successful differentiation of three different species and five different pathovars within the P. syringae species was possible. Machine learning approaches allowed to identify and discriminate isolates at the genus, species, and pathovar level, reaching in the latter case accuracies greater than 85%. This Raman-DEP strategy offers interesting advances in the identification of microorganisms and the evaluation of susceptibility to antibacterial substances. Moreover, the technique could be readily extended to environmental and food diagnostics.