Valuations of Variables as Selection Index for Improving Grain Yield in Bread Wheat

Identifying selection indices is the most important step of a breeding project that aims to improve grain yield. The definition of the selection index is usually done by evaluating the variables in multivariate statistical methods. In the present study, the relationship between grain yield and its components in bread wheat genotypes was determined by multivariate statistical methods. The experiment was conducted in a randomized complete block design with 3 replications in the research farm of Gorgan University of Agricultural Sciences and Natural Resources in the 2018-19 crop years. Ten commercial cultivars of bread wheat along with their offspring from direct and inverse crosses in a dialysis arrangement were evaluated for morphological and phenological traits, especially grain yield and its components. The results of genotypic and phenotypic correlation coefficients showed a positive and significant correlation (at 1% probability level) between grain yield and spike length, spike weight, number of fertile tillers, number of seeds per spike, number of spikes per spike, 1000-seed weight, biological yield and harvest index. Based on the results of stepwise regression analysis, biological yield, harvest index, number of grains per main spike and main spike weight were entered into the regression model, respectively, and explained a total of 98% of the variation in grain yield. Based on the results of path analysis, biological yield had the highest direct effect on grain yield. After biological yield, the most direct effect on grain yield was related to the weight of main spike. Also, by considering eigenvalues greater than one in factor analysis, 8 hidden factors were identified that explained a total of 75.18% of the data changes. In general, it can be concluded that biological yield, harvest index, number of seeds per spike and weight of spike compared to other traits can be used as appropriate indicators in breeding programs to select high-yield genotypes in field conditions. Genotypes Alo, Ehsan♂ × Gonbad♀ and Ehsan had the highest value for the studied traits, which can be used in future breeding researches.