Detection and Classification of Bacteria using Raman Spectroscopy Combined with Multivariate Analysis

Vibrational spectroscopic techniques have advantages over conventional microbiological approaches towards identification & detection of pathogens. Since unique spectral fingerprint is obtained, one can identify very closely related bacteria using such methods. In this study Raman microspectroscopy in combination with chemometric method has been used to classify four strains of E. coli (two pathogenic & two non-pathogenic). Different multivariate approaches such as hierarchical cluster analysis, principal component analysis & linear discriminant analysis were explored to obtain efficient classification of the Raman signals obtained from the four strains of E.coli. It was observed that multivariate analysis was able to classify the bacteria at strain level. Linear discrimination analysis using PC scores (PC-LDA) was found to give very good result with as high as 100 per cent accuracy. This hybrid technique (Raman spectroscopy & multivariate analysis) has tremendous potential to be developed as a tool for bacterial identification.