Using multivariate statistical methods to estimate the model maize population

Abstract

In order to objectively assess the degree of similarity of varieties and hybrids by a set of indicators and to interpret a large amount of data in breeding practice, methods of multidimensional statistics are used, including factor and cluster analyzes. The study included 42 maize hybrids (FAO 100-499) of the selection of various scientific institutions of Russia, which were assessed according to the following characteristics: the length of the interphase stage of seedlings - flowering of panicles and ears, plant height, ear attachment height, grain yield, number of grains per ear, crude protein and crude fat content. Factor analysis made it possible to calculate the correlation coefficients between the variables under consideration and to reveal high positive correlations between the interfacial stages of a panicle seedlings–flowering and seedlings–-flowering of ears (r = 0.99), plant height and ear attachment height (r = 0, 81), the length of the interphase periods and the ear height (r = 0.77). A significant correlation coefficient was determined between the traits, the length of the interphase periods, seedlings – flowering of panicles and ears with grain yield (r = 0.34 and 0.36, respectively), grain yield and number of grains per ear (r = 0.37). When calculating the weights of the variables for the components, it was found out that the largest contribution (variance 54.814%) to the first hypothetical factor was made by the following signs: interphase periods of panicle seedlings–flowering and of ear seedlings–flowering, plant height, ear attachment height, grain yield, 1000 grain weight. The use of cluster analysis made it possible to group the studied hybrids at 32 iteration steps (Euclidean distance - 7.340) into 10 clusters with different breeding values. The hybrids that make up the clusters are characterized by a similar set of features within the clusters and significant differences between the clusters, confirmed by the method of unorganized repetitions of one-way ANOVA.