Abstract

BIASUTTI, C.A.  and  BALZARINI, M.. Estimation of maize hybrids performance using mixed models. Agriscientia [online]. 2012, vol.29, n.2, pp.59-68. ISSN 1668-298X.

The selection of superior cultivars is based on its performance in the so-named multienvironments trials. The problem of the correct choice of cultivars is increased when these trials give rise to heavily unbalanced data in addition to the presence of the genotype x environment interaction. To identify the best cultivars, the best unbiased prediction or BLUP has been proposed, using linear mixed models for the analysis of phenotypic data together with data of kinship. The objectives of this study were: to estimate the performance of maize hybrids from unbalanced data, with or without the inclusion of the genetic kinship, and to determine the predictive power of the years of assessment on the performance of the genotypes. Performance data for 49 hybrids of maize from 2006 to 2009 were used for the estimates of the BLUPs of each genotype. The consideration of relations of coancestry in the calculation of the BLUPs allowed a better discrimination between the genetic material. To improve the discrimination of hybrids according to their performance, the use of the tBLUP based on more than 2 years, including evaluation in environments which allow a better expression of the yield potential of experimental hybrids is recommended.

Keywords : Yield; BLUPs; Maize.