Use of Factor Analysis in Grain Sorghum Breeding

Abstract

The article presents the results of calculating factor loads for 2015-2018 obtained by the method of principal components of grain sorghum varieties. Field studies were laid and carried out in the fields of the Federal State Budgetary Scientific Institution RosNIISK Rossorgo. The objects of the study were grain sorghum plants of the Institute's selection, as well as samples presented by the VIR world collection. 25 economically valuable indicators in 32 samples were analyzed. The main role in selection of indicators for inclusion in the correlation model is assigned to the expert approach in assessing breeding material. Effective and combinational groupings of correlation coefficients by hypothetical factors are very important tools in solving the issue of selecting indicators in a correlation model of interrelationships of selectively valuable indicators of various plant populations. Grouping the correlation coefficients makes it possible to assess the general nature of the interaction of the indicators included in the experimental data matrix. Calculations of the loads on hypothetical factors in the experiment made it possible to identify factors whose contribution to the accumulated variance is more than 5 %.