Melhoramento da soja visando produtividade e teores de óleo e proteína nos grãos

Abstract

The importance of soybean for Brazilian agribusiness is notably. In terms of area, it is the most cultivated crop in this South America country. The oil and protein content, 20 and 40%

respectively, have a significant importance for soybean cultivation. However, the genotype- by-environment (GE) interaction has influence on grains chemical composition, such as oil

and protein content. Due to its importance, two scientific works were performed in order to increase the knowledge about GE interaction effects in soybean oil and protein content. On the first work, soybean progenies were evaluated for their agronomic performance and grain quality. The collected data were analyzed by using mixed model approach in order to estimate genetic and phenotypic parameters associated to grain yield, grain oil and protein content. The expected and realized selection gains, correlation and correlation-response between these traits, and BLUPs frequency distribution were also estimated. The significance of the estimate of genetic variance denotes variability between soybean progenies, allowing selection of good agronomic performance and high levels of oil and protein content genotypes. The GE interaction influence was noticed on all three evaluated traits. In addition, a second study was

performed aiming to determine and decompound the GE interaction effects into genotype-by- crop years and genotype-by-location interactions, as well as the relative contribution of source

of variation in grain yield, grain oil and protein content of soybean progenies cultivated in different environments. For this purpose, joint analyzes by location, by crop year, and by all environments (location x crop year) were performed. The relative contribution of each source of variation as well as the heritability for these traits were estimated. The relative contribution of the genetic effect was greater than the effect of the GE interaction for all traits. Therefore, genetic factors perform greatly in phenotypic expression. For protein content trait, it is preferable to evaluate a vast number of locations over a larger number of crop seasons. On the other hand, for oil content trait, aiming experiment optimization, it is recommended to evaluate the progenies over several crop seasons in detriment of a larger number of locations.