Statistical methods for more precise selection of biofortified common bean lines

Abstract

Identifying statistical methods that provide greater experimental precision in the selection of biofortified common bean lines is an unprecedented practice, according to the literature. The present study was thus conducted to examine the experimental precision of mineral concentration in experiments with common bean by employing different statistical methods and to identify more precise methods to select mineral-biofortified common bean lines. Twenty-two experiments were carried out between 2004 and 2018 evaluating 682 common bean genotypes of the grain types most widely consumed in Brazil. The concentrations of potassium, phosphorus, calcium, iron, zinc and copper were determined by nitric-perchloric acid digestion, and the experimental precision of those traits was analyzed by eight statistical methods. For most experiments, a significant genotype effect was observed for mineral concentration, allowing the selection of biofortified common bean lines. The coefficient of experimental variation and significant minimum difference exhibited low and intermediate values, i.e., showed greater experimental precision, including the experiments in which the genotype mean square was not significant. Because these methods may result in error in genotypic differentiation, they are not adequate to evaluate the experimental precision of mineral concentration in common bean. In most experiments where the genotype effect was significant for the mineral concentration, high scores (greater experimental precision) were obtained in the F-test value for genotype, coefficient of genetic variation, coefficient of relative variation, heritability and selective accuracy. These statistical methods allow the selection of biofortified common bean lines with greater experimental precision and are thus recommended for breeding programs.