Prediction of genetic value in F3 populations of Avena sativa L. using Reml/Blup
Jefferson Luís Meirelles Coimbra, Mauricio Marini Kopp, Velci Queiroz de Souza, Giovani Benin, Volmir Sergio Marchioro, Fernando Iraja Felix de Carvalho, and Antonio Costa de Oliveira
In genetics and breeding studies it is common to conduct experiments of fixed (sowing method) and random (populations) factors. Therefore, the most appropriate statistic analyses would use mixed linear models. In this sense, objectives of this work were i) to estimate the variance components for the random factor effects and interaction population x sowing method; ii) to perform selection among populations and among populations in each sowing method; iii) to compare the effect of the fixed factor. Both the effect of components of the population variance as well as the effect of the single interaction population x sowing methods revealed low contributions to the total variance. It can therefore be concluded that segregating populations present narrow genotypic variability. The Reml/Blup procedure is indicated for estimation/prediction in oat improvement where experiments generally produce unbalanced data.