FTIR-based Metabolomics Approach to Discriminate Soybeans from Different Geographical Origins of China

Abstract

Abstract Origin traceability of soybeans using infrared spectroscopy is bound by data mining, which can be solved by metabolomics analysis. In this study, a novel infrared spectroscopy-based metabolomics approach via seeking ‘wave number markers’ was developed to achieve the discrimination of soybeans from ten different cities of China. Multivariate analytical procedures including principal component analysis (PCA), cluster analysis and orthogonal partial least squares discriminant analysis (OPLS-DA) were designed for separation of all soybean groups, which provides a possibility to discern ‘markers’ among groups. S-plot, permutation test and variable importance in projection (VIP) embedded in OPLS-DA model took on the screening of ‘markers’, which were further verified by pairwise t-test in univariate analysis. There are 27 ~ 330 ‘markers’ picked out in ten soybean groups, with the wave number range to be 761.882 ~ 956.693, 2430.308 ~ 2789.068, 974.052 ~ 1068.564, 1504.476 ~ 1554.626, 2796.783 ~ 3431.364, 3890.422 ~ 4000.364, 3805.554 ~ 4000.364, 761.882 ~ 819.747, 457.129 ~ 530.424 and 460.987 ~ 514.994 cm− 1, during which significantly high absorbance can be observed for No. 2 ~ No. 7 soybeans, but for No. 1 and No. 8 ~ No. 10 soybeans, we can observe significantly low absorbance. The results indicate that infrared spectroscopy coupled with metabolomics analysis is equal to origin traceability of soybeans, thus, it provides a novel and viable approach for the accurate and rapid discrimination of soybeans from different geographical origins.