Abstract

Most sugarcane breeding programs tend to evaluate low heritability characteristics during the initial stages of genotype selection. Thus, family selection has been recently preferred. In this context, the aim of the present study was to select the best family among 78 sugarcane families, as well as estimate genetic values through the mixed models of restricted-maximum likelihood and best non-bias predictor (REML/BLUP) methodology, originating from the República Brasil 2005 (RB05) series. This strategy was deemed efficient, and 34 to 38 families were chosen from four evaluated characteristics underexplored by genetic researchers such as total plot mass (MTT), mean mass of one tiller in the plot (M1C), stature (EST), and mean number of canes per square meter (NCM). The family increments ranging from 6.02 to 82.11%, in the next genetic culture improvement program selection phases.

This work is licensed under a Creative Commons Attribution 4.0 License.