Chemometrics approaches of near-infrared spectroscopy for chemical marker compounds in food.

摘要

Near-infrared (700-2500nm) is the region in which absorptions are due to the overtone and combination bands of the molecular stretching and bending vibrations of the fundamental absorptions of -CH, -NH, and -OH groups. As a result, near-infrared absorption bands are complex bands which are characteristically weak, highly overlapped, and very broad. Chemometrics approaches would be the only way to extract the useful information from near-infrared complex spectra for calibration and finally establish rapid, simple, nondestructive near-infrared spectroscopic method. It can be used the determination of chemical marker compounds in food.; Chemometrics approaches include mathematical treatment of the raw near-infrared spectral data, regression for the calibration and statistical evaluation of the calibration and prediction. In this research, smoothing and derivatives were used to treat the near-infrared raw spectral data, and Multiple Linear Regression (MLR), Stepwise Regression and Partial Least Squares regression (PLS) were used for development of the calibration models.; There were four food systems selected in this research and chemometrics approaches were used for their near-infrared spectra to develop the calibration models. We designed the silica gel absorption model system so that the near-infrared spectroscopic method can directly determine the volatile compounds adsorbed and carried by silica gel. This silica gel adsorption/near-infrared spectroscopic method presented a much broader application of near-infrared spectroscopy. We also designed gravy model by adding different amount of beef flavor seasoning with variation of starch contents so that we could use multivariate calibration techniques to obtain the near-infrared calibration model. For the first time, a near-infrared transmission spectroscopic method has been developed to directly determine the starch content in a liquid gravy model system. The developed near-infrared spectroscopic method for the determination of moisture, protein and fat in fish meal can be used at the processing site over the chemical analysis method which would save extensive time and money for the fish meal industries. Near-infrared spectroscopic method with chemometrics approaches has also been developed in this research to directly determine free fatty acids as quality marker compounds in fish oil and mackerel. The developed method can be used for the assessment of mackerel quality.