Near infrared spectroscopy integrated with chemometrics for rapid detection of E. coli ATCC 25922 and E. coli K12

A rapid and accurate technique based on near infrared (NIR) spectroscopy integrated with chemometrics to identify and quantify two strains of Escherichia coli including E. coli ATCC 25922 and E. coli K12 grown in liquid media was developed. The samples were analyzed using aerobic count and recorded as colony forming units (CFU). The NIR spectra of both E. coli strains at different growth stages were recorded from 750 to 1350 nm. Principal component analysis (PCA) was performed for a data reduction purpose. Discriminant factor analysis (DFA) was then used to visualize sample classification. From DFA, it was possible to differentiate between E. coli ATCC 25922 and E. coli K12 as well as different cell concentrations. The cell concentrations were simultaneously predicted from the NIR spectral data using a supervised artificial neural network (ANN) based on multilayer perceptrons with back propagation algorithm. The training function was a Levenberg–Marquardt algorithm. The ANN yielded good predictions as measured by a regression coefficient (R2 = 0.98) between actual and predicted data. Compared to the colony counting method, the developed technique could be accurately performed in far less time and eliminate the need for intensive sample preparation.