Eficiência da seleção de genótipos de arroz de terras altas via modelos mistos

Abstract

It is necessary, within breeding programs, to obtain estimates of genetic and phenotypic parameters, genetic gains and progress over the years, in order to evaluate the efficiency of the program in the conduction and selection of populations. When working with different years of selection, it is desired that the program be dynamic, with exclusion and inclusion of genotypes each year. In the end, the data become unbalanced and, in these cases, the mixed models become tools of paramount importance in the analysis of the data. The objective of this study was to estimate genotypic and phenotypic parameters in order to verify the efficiency of genotypes selection in the preliminary trials of UFLA highland rice breeding program, using mixed models. The experiments were conducted in five environments during three harvests and, at each harvest, two common controls and 34 genotypes were used, totaling 91 genotypes. The evaluated characteristics were: grain yield (kg.ha -1 ), plant height (cm), days until flowering and resistance to leaf blast (scale of notes). In addition to the genetic and phenotypic parameters, expected gains were obtained with the selection, genetic progress, renewal rate (TR%) and dynamism of the preliminary tests. It was verified efficiency in the selection of genotypes superior to each year, with genetic progress for grain yield of 643 kg (ha.year -1), and reduction of plant size of 1.14 cm.year-1.