Análise genética de Tachigali vulgaris para fins de melhoramento em condições experimentais de alta mortalidade das plantas

Abstract

Tachigali vulgaris is a native tree species that stands out for use in energy production. However, experiments in the silvicultural and breeding areas have indicated that the species has a high mortality rate. Thus, the hypothesis of this study was to verify whether the mortality of neighboring trees influenced the mean annual increment (IMA) of a focal tree, in the selection and estimation of genetic parameters in a test of provenance and progenies of T. vulgaris. The objectives were: 1) to evaluate whether considering mortality in the data analysis influenced the estimates of genetic parameters; 2) to analyze the influence of the adjustment with covariates related to competition and mortality of neighboring trees in the estimates of genetic parameters; 3) verify the possibility of carrying out the early selection of T. vulgaris individuals. Three different situations were simulated depending on the survival of the trees at three and 10 years of age and two independent mixed models were adjusted. One to determine which survival situation would be most appropriate to carry out the estimation of genetic parameters of T. vulgaris and the other model to verify the possibility of practicing early selection. Additive breeding values, total breeding values for both provenances and individuals and the population size of non-consanguineous individuals (effective size) were calculated. The hypothesis that the mortality of neighboring trees influences the average annual increment of a focal tree was confirmed, so that the greater the number of dead trees close to the analyzed tree, the greater the productivity of this tree. Considering the dead trees after three years of planting influenced the adjustment of the models to be able to estimate the genetic parameters of T. vulgaris, as well as using the adjustment with covariates related to the mortality of neighboring trees influenced the estimates of the genetic parameters to carry out the selection of the best genetic materials of T. vulgaris. We selected 480 individuals with an effective size of 107 non-consanguineous individuals and unrelated genotypes from the provenance and progeny test of T. vulgaris. Obtaining a selection gain of about 15% in the selection based on the total breeding values (EGV) and of about 10% in the selection of additive values (EBV). And the feasibility of carrying out the early selection was possible through the genetic correlation between the IMA at three years of age and the IMA at 10 years old, in which a value equal to 90.12% was obtained, indicating good feasibility to perform this technique.