Detection limit of FT-IR-based bacterial typing based on optimized sample preparation and typing model

Accurate and efficient bacterial typing methods are crucial to microbiology. Fourier transform infrared (FT-IR) spectroscopy enables highly distinguishable fingerprint identification of closely related bacterial strains by producing highly specific fingerprints of bacteria, which is increasingly being considered as an alternative to genotypic methods, such as pulsed field gel electrophoresis (PFGE) and whole genome sequencing (WGS), for bacterial typing. Compared with genotypic methods, FT-IR has significant advantages of convenient operation, fast speed, and low cost. Fundamental research into the detection limit based on optimized analytical conditions for FT-IR bacterial typing, which can avoid excessive bacterial culture time or sampling volume, is particularly important, especially in clinical practice. However, the corresponding parameters have not been fully investigated. In this study, we developed a simplified and reliable procedure for sample preparation, optimized the data analysis procedure, and evaluated the FT-IR detection limit based on the above conditions. In particular, we combined the film mold and calcium fluoride plate for sample preparation. We evaluated the detection limit (about 108 CFU/mL) after parameter optimization using hierarchical cluster analysis (HCA) and artificial neural network (ANN). The optimization and evaluation of these key fundamentals will better promote future application of FT-IR-based bacterial typing.