BACKGROUND: Maize (Zea mays L.) is a staple cereal crop and feed crop throughout the world. In this article, a mineral element fingerprinting technique was applied to single out suitable element indicators to determine the geographical origin of maize. A total of 90 fresh maize samples were collected in 2107 from Jilin, Gansu, and Shandong provinces in China. The contents of 25 mineral elements in all maize samples were measured by inductively coupled plasma mass spectrometry (ICP-MS). The composition of mineral elements was analyzed by multivariate statistical analysis, including one-way analysis of variance (one-way ANOVA), principal component analysis (PCA), k-nearest neighbor (KNN) analysis, and stepwise linear discriminant analysis (SLDA). RESULTS: As compared by one-way ANOVA, the contents of 19 mineral elements in maize samples were significantly different among three provinces. Principal component analysis based on these 19 elements could obtain preliminary visual classification groups of maize samples. K-nearest neighbor analysis produced a total correct classification rate of 83.9% on the training set, and 82.2% on the prediction set. The SLDA model, based on eight indicative elements (Na, Cr, Rb, Sr, Mo, Cs, Ba, and Pb) obtained a total correct classification rate of 92.2% with cross-validation. CONCLUSION: The mineral element fingerprinting technique combined with multivariate statistical analysis could be a helpful method to identify the geographical origin of maize.
BACKGROUND:Maize (Zea mays L.) is a staple cereal crop and feed crop throughout the world. In this article, a mineral element fingerprinting technique was applied to single out suitable element indicators to determine the geographical origin of maize. A total of 90 fresh maize samples were collected in 2107 from Jilin, Gansu, and Shandong provinces in China. The contents of 25 mineral elements in all maize samples were measured by inductively coupled plasma mass spectrometry (ICP-MS). The composition of mineral elements was analyzed by multivariate statistical analysis, including one-way analysis of variance (one-way ANOVA), principal component analysis (PCA), k-nearest neighbor (KNN) analysis, and stepwise linear discriminant analysis (SLDA). RESULTS: As compared by one-way ANOVA, the contents of 19 mineral elements in maize samples were significantly different among three provinces. Principal component analysis based on these 19 elements could obtain preliminary visual classification groups of maize samples. K-nearest neighbor analysis produced a total correct classification rate of 83.9% on the training set, and 82.2% on the prediction set. The SLDA model, based on eight indicative elements (Na, Cr, Rb, Sr, Mo, Cs, Ba, and Pb) obtained a total correct classification rate of 92.2% with cross-validation. CONCLUSION: The mineral element fingerprinting technique combined with multivariate statistical analysis could be a helpful method to identify the geographical origin of maize.
Authors: Bhamidipati Kishore; Ali Yasar; Yavuz Selim Taspinar; Ramazan Kursun; Ilkay Cinar; Venkatesh Gauri Shankar; Murat Koklu; Isaac Ofori Journal: Comput Intell Neurosci Date: 2022-08-10