Wheat lipids to discriminate species, varieties, geographical origins and crop years

Abstract

The knowledge of lipid components of wheat finds a precious information in order to differentiate between Triticum durum (TD) and Triticum aestivum (TA). The determination of the percentages of methyl esters of the differently unsaturated fatty acids with 18-carbon atoms (C 18), of sterol fraction and of the other components is of particular weight. In this paper, the classification methods of linear discriminant analysis (LDA) and quadratic discriminant analysis (QDA) were applied in order to measure the classification and prediction abilities of the determined (percentages of the) components of the lipid fraction of wheat in differentiating among species, origins, varieties and crops. By univariate feature selection method (Fisher weights (FW)) and linear discriminant analysis, it was found that the only oleate is able to distinguish between the two species with a prediction rate of 100%. Inside the species Triticum durum, it was obtained a prediction rate of 83.9% while discriminating between the different origins, a prediction rate of 82.2% while discriminating among varieties and a prediction rate of 94.3% among crop years.