| Near infrared Raman spectroscopy is a non-invasive non-destructive technique that provides molecular structural and compositional information about medical and biological specimens. It has been developed for identification of micro-organisms in the past decade. In this thesis, we present a new application of this method on identification and quantification of major oral bacterial species including the most cariogenic species, Streptococcus mutans, for the eventual purpose of monitoring and preventing dental caries.; A home-built confocal Raman microscope system designed for this purpose is described in this thesis. This system is capable of acquiring a Raman spectrum from about a thousand bacterial cells in a few minutes with 100 mW of optical fiber-delivered 830 nm laser light. Experiments using this system are presented.; The performed experiments have shown that with the Raman spectroscopic method various oral streptococci species, including S. mutans, can be identified via principal component analysis (PCA) and hierarchical clustering analysis (HCA) of the Raman data. Identifications have been performed successfully with bacteria cultured in different types of environment, including agar colonies, liquid culture and biofilms; the identification can be performed regardless of sub-species differences and, particularly important, regardless of bacterial growth phase.; For the first time, we demonstrate the ability of confocal Raman spectroscopy as a novel approach to quantify bacteria concentration directly in mixtures. The relative populations of multiple bacterial species, including S. mutans, can be measured directly from mixtures with accuracy of a few percent. Concentration prediction models for measurement built with partial least-squares (PLS) analysis have been proven for robustness by cross-validation as well as validation with separate samples. The prediction accuracy has relevance for routine S. mutans concentration monitoring; the data acquisition time is 100-200 seconds per spectrum; total time spent to get a result from sample collection is about 5-20 minutes, which is much shorter than the 48 hour time scale required by conventional plating methods. No external chemical agent is required. The amount of sample required can be easily collected from clinical dental sites, so no sample incubation time is needed. Inclusion of salivary proteins did not affect the accuracy of the predictions. Preliminary calibration transfer results are presented. Continued study of biofilm samples, a model for dental plaque, is proposed as well. |
