Duktova N.A., Grib S.I., Ivanova E.V., Andronik E.L., Vinogradov D.V.

Application of multidimensional analysis for assessing the breeding value of the initial material of durum spring wheat

The paper presents the use of multivariate statistical analysis for the selection of source material from 263 collection samples of durum wheat. As a result of factor analysis, 10 studied characteristics were reduced to 3 significant factors. The informative features, which should first of all be paid attention to when assessing this set of genotypes, are highlighted: the number of spikelets of the main spike, the number of grains of the main spike, grain weight of the main spike, grain weight per plant, average grain weight of an ear (factor 1); the length of the main spike and the density of the main spike (factor 2); productive bushiness (factor 3). The set of studied collection samples of durum wheat, based on informative characters, were divided into 3 reliably distinguishable clusters by the k-means method. Each cluster unites in itself from 56 to 123 samples of the most similar genotypes in terms of the complex of studied traits. The selected clusters are characterized as independent groups for subsequent breeding studies.

Keywords: TRITICUM DURUM, SELECTION, FACTOR ANALYSIS, CLUSTER ANALYSIS, INFORMATIVE FEATURES, SOURCE MATERIAL