Identification Of Food-borne Pathogens Based On Fourier Transform Near-infrared Spectroscopy

The rapid detection technique of food-borne pathogen is one of an effective method for solving food safety problems.Because of nondestructive, efficient, speediness, convenient and environmental,near infrared spectrum technology is widely used in food safety area. On the basis of research achievements about E.coli,this work took listeria monocytogenes (LM) and salmonella which in logarithmic growth as object and obtained the optimal parameters of sample pretreatment.Using projection discriminant method and clustering analysis based on principal component analysis,the best method of Near infrared spectra pretreatment was determined.It was proved that there were specific absorption of the near infrared spectral among the whole cell,the cell wall and the cytoplasm of these three pathogenic.At last, the method of identifying food-borne bacteria by Near-infrared spectroscopy was determined.This result was meaningful to clarify mechanism of detecting food-borne bacteria by Near-infrared spectroscopy and laided the foundation for the development of detecting food pathogenic method by near infrared spectral analysis technology.The main results:1. In this study, logarithmic growth were determined through the growth curves of LM and salmonella by turbidimetry. At last, the time of enrichment culture were identified respectively 24h and 16h.2.Took the rate of cell disruption of LM and salmonella with index, the optimal parameters of ultrasonic cell disruption were determined by single factor experiments and orthogonal tests.The best parameters for ultrasonic cell disruption of LM were: output power was 600W, interval time was 5s, total time was 70min.And the best parameters of salmonella were: output power was 600W, interval time was 5s, total time was 60min.3. Took the whole cell of E.coli,LM and salmonella which in logarithmic growth for example and used projection discriminant method based on principal component analysis, we obtained that the first two principal component were optimal to identify these three food-borne pathogenic bacteria.4. After the spectral data were processed by vector normalization method and multiple scattering method, the resolution of the projection discriminant analysis was 100%. In this study,the Ward 's method was used for clustering analysis.The results of identification were the most accurate after the spectral data of bacteria processed by vector normalization method,but not expected. There are all LM in category 2; E.coli and salmonella distributed in category 3 which each focused on a region; category 1 only contained two error samples.5. The results of near-infrared used model identification method indicated that there were specific absorption among different bacterial which could be characterized by infrared analysis technique and chemometrics methods. In the three kinds of sample,the result of the cell wall identification model was optimum that the resolution was as high as 100%.