Selection of barley advanced lines at rainfed conditions using regression and cluster analyses

Abstract

Due to the low heritability, direct selection for yield has usually low efficiency at rainfed conditions. Therefore, plant breeders may improve yield indirectly through selection of yield contributing traits. The purpose of this study was to suggest a method for indirect selection of promising lines through yield contributing traits. This method was tested on a set of 18 advanced lines of winter barley cultivated at Gachsaran Agricultural Rainfed Research Station, Iran, for two years from 2009-10 to 2010-2011. The experimental design was a randomized complete block with four replications. Grain yield along with eight yield-related traits were assayed. In this method, initially, yield contributing traits were detected using a multiple linear regression analysis. Then, the obtained predictors were weighted by the corresponding regression coefficients. Selection of lines was carried out based on cluster analysis. Estimation of correlation coefficients revealed that, grain yield (GY) was positively correlated with plant height, weight of hectoliter and the number of grains per spike (GS). The highest and lowest broad-sense heritability belonged to GS and GY, respectively. Cluster analysis determined that, genotypes 7-B-Gachsaran, 8-BNYT-Gachsaran and 12B-Gonbad had the best level of yield predictors in both years and therefore, were the most stable genotypes.