Use of REML/BLUP for the selection of sugarcane families specialized in biomass production

Abstract

One hundred and fourteen full-sib families were evaluated at two sites in Minas Gerais in experiments in augmented blocks and the data analyzed by the REML/BLUP method. Results showed a low genetic variability among families, low dominance variance, and absence of families x sites interaction. The mean components estimated via BLUP permitted the selection of superior families and parents. The use of the augmented block design without replication in family selection experiments did not prove adequate due to the low estimates of selective accuracy and family mean heritability, The adoption of other designs that use a higher number of replications is desirable to increase the efficiency of family selection for this population. Owing to the large number of families and the use of the REML/BLUP procedure, using or not using the checks in the augmented block design did not alter the estimated genetic parameters.