Improvement of the metod of protein content prediction in wheat kernels

The study was conducted in order to expand the possibilities of using the developed multiple nonlinear regression equation, re ecting the dependence of the protein content in wheat grain (Y, %) on the content of raw gluten (X1, %) and the mass of 1000 grains (X2, g). The possibility of practical use of the developed and three improved equations for predicting the protein content of wheat grain is shown. An algorithm and results of checking the prognostic capabilities and accuracy of predict equations from independent data are presented. Generalization of experimental data from 288 literature sources of domestic and foreign authors with a total number of observations n = 4604 on more than three hundred varieties of winter and spring soft and durum wheat grown in the period from 1959 to 2020 in Russia and abroad (Albania, Belarus, Bulgaria, Egypt, Kazakhstan, Lithuania, Poland, Slovakia and Ukraine) showed that the number of values that go beyond to limits regulated by GOST 10846-91 «Protein determination method» was 755 or 16.4 %. At the same time, the accuracy of prediction the protein content reached 83.6 %. The developed equations can be used to predict of protein content in almost all cases: when the results of the analysis of protein content and raw gluten in wheat kernel, as well the 1000-kernels weight are given at actual or fxed moisture content, or in terms of dry matter

wheat (Triticum aestivum L.), protein, raw gluten, 1000-kernels weight, multiple regression analysis, protein content prediction