Geographical traceability of soybean based on elemental fingerprinting and multivariate analysis

Abstract

In this study, elemental composition differences of soybeans were analysed from four regions of the Chinese Heilongjiang Province (Daqing, Suihua, Heihe, and Jiamusi), and the characteristic fingerprints representative of the producing areas were screened. The contents of 25 elements in soybeans from the four producing areas were determined using an inductively coupled plasma mass spectrometer (ICP-MS), and the geographical sources of soybean were identified using difference, correlation, cluster heat map, and orthogonal partial least squares discriminant analyses (OPLS-DA). The results of the difference and correlation analyses showed that the elemental compositions were significantly affected by the soil environment for growth, and there were significant differences in element content among the four soybean-producing areas with regional characteristics. Heat map clustering showed the aggregation of the element content among different producing areas, distinguished the samples, and allowed classification of all elements. A discriminant model was established for the samples in the training set using the indices of 17 screened elements, including Mg, Al, K, Mn, Mo, p, Cu, Cr, Rb, Ni, Ca, Fe, Se, Pd, Zn, Ga, and Pb, and was used for the prediction and analysis of soybean samples in the testing set. Overall, the correct discrimination rate of the four soybean samples was 93.33%, which indicated these 17 elements contained sufficient information representative of the soybean-producing areas. Furthermore, they could be used as stable and effective traceability indicators to identify the production area of soybean samples from the four producing areas in Heilongjiang Province.