Associação entre caracteres e seleção de híbridos de sorgo biomassa para produção de bioenergia

Abstract

Biomass sorghum, a promising energy crop for bioenergy production, is efficient for both electricity generation and second generation (2G) ethanol production. The efficiency of the selection of superior sorghum biomass genotypes depends on the industrial purpose and the knowledge of the interrelationship of the several traits that directly or indirectly influence the phenotypic expression. In addition, the selection of genotypes involving multiple traits using suitable indices such as FAI-BLUP can contribute to better targeting in crop improvement programs for genotype selection based on an ideal genotype. Thus, the objective of this study was, in the first chapter, to estimate the associations between sorghum biomass traits; estimate the direct and indirect effects and verify the existence of cause-and-effect relationship of the other traits with the theoretical heating power (THP) via path analysis; and provide a complementary technique for multi-trait analysis and targeting sorghum biomass improvement programs through Data Mining. The THP showed both null and positive associations with the evaluated traits. Highlighting the positive associations with flowering (FLOW); plant height (PH); green stem biomass productivity (GSBP); stem (SDM) and leaf (LDM) dry mass; leaf neutral detergent fiber (LNDF); lead acid detergent fiber (LADF); stem lignin (SL); leaf lignin (LL); and leaf hemicellulose (LH). The traits FLOW, PH, GSBP, LDM and SL were considered the main determinants of THP variation, and can be used in the indirect selection of this target trait. In addition, the traits FLOW, PH and GSBP directly or indirectly influenced the expression of the traits LDM and SL, and can be used in the indirect selection of these traits. The data mining technique, regression tree, allowed a better visualization of the relationships between the traits, helping in some explanations about the influence of the traits on the THP. In the second chapter, the objective was to identify and select superior sorghum biomass genotypes to produce second-generation ethanol considering the selection for multiple traits, as well as to verify which strategy would be the most appropriate for the selection of these genotypes by the FAI-BLUP index considering experiments in multiple locations. The strategies used to calculate the FAI-BLUP index were: first strategy (E1) use of the genetic values predicted in each location; second strategy (E2) use of predicted genetic values free of genotype by environment (GxE) interaction considering all locations; third strategy (E3) use of predicted genetic values capitalizing on the GxE interaction; fourth strategy (E4) use of genetic values predicted in E1 and creating new traits by location; fifth strategy (E5) using predicted values in E3 and creating new traits per location. Genotypes 4, 13, 17 and 18 stood out in terms of superiority, predictability and genotypic adaptability, and their use is interesting for second-generation ethanol production. The use of E2 and E3 were the most suitable for the selection of superior genotypes by the FAI-BLUP index considering experiments in multiple locations and aiming at the recommendation, respectively, for locations that present and/or not the same pattern of GxE interaction of the evaluated experimental network.