Estimation Of Genetic Parameter From Trails Designed As Augmented Block Design.

Abstract

The aim of this study was to obtain unbiased estimates of genetic parameters in the application of trials designed in augmented block design. In addition to this, it was planned to compare the variance and covariance components of the intra-block analysis of an augmented experiments and the estimators of the components associated to genetic effect. For analysis and discussion; the data used consisted four local checks and one hundred eighty eight Durum wheat genotypes evaluated using augmented block design at Sinana Agricultural research center in 2013 cropping season. A model that have been contained check treatments and block as fixed factors and Test treatments as random factor have been selected for the data set based on Akaike’s Information Createria evidence. The ratios of standard errors of GLM to that of mixed model using trial designed as ABD were about 2.7 for both ML and REML. These results tell us that a mixed model is more valuable than a GLM to remove the down ward bias of variance of the response variable and the boosted variance of the error terms of GLM. The inter-class correlation coefficient result showed that estimation based on REML techniques best to estimate variance component in linear mixed model for trial designed as augmented block design. Finally, the findings of this study showed that about one hundred one genotypes of durum wheat have the highest mean yield effects than standard check genotypes. Therefore about 101Durum wheat genotypes materials have been recommended for next selection program in similar ecology to Sinana.