Applications of multi-trait selection in common bean using real and simulated experiments

Title 
Applications of multi-trait selection in common bean using real and simulated experiments 
Publication Type 
Journal Article 
Authors 
Balestre M, Torga PP, Pinho RG, Santos JB 
Year of Publication 
2013 
Volume 
189 
Journal 
Euphytica 
Issue 
Pagination 
225 - 238 
Date Published 
1/2013 
ISSN 
1573-5060 
URL 
http://link.springer.com/article/10.1007/s10681-012-0790-1 
Keywords 
Bayesian analysis, Beans, Multi-trait heritability, Supplemented EM algorithm, Yield components 
DOI 
10.1007/s10681-012-0790-1 
Abstract 

The analysis of multiples traits has been largely employed in animal and forestry crop breeding; however, in annual crops such as common bean this approach is practically lacking. This study examined the selective accuracy and estimates of genetic parameters in common bean using single- and multi-trait analyses in several designs of genetic correlations and heritabilities by means of simulated- and real-traits. We also compared the efficiency of the Bayesian and REML methods as regarding its selective accuracy and estimates of genetic parameters. In real data experiments, 100 endogamic families of common bean were assessed for grain yield (GY) and grain type (GT) in three environments using a 10 × 10 triple lattice design. Single- and multi-trait analyses were applied for these traits using REML-based BLUP analysis and Bayesian methods. In simulated experiments, traits corresponding to GY and yield components in common bean were simulated under several levels of genetic correlations and evaluated under different levels of heritability in order to verify in which genetic/experimental design it is advantageous to apply single- or multi-trait analyses. The results obtained in this study indicated that GY and GT present low genetic correlation and, therefore, multi-trait analysis may be unjustifiable for these traits. The Bayesian and REML multi-trait analyses were equivalent in point and interval estimates of genetic parameters and predictive accuracy, both for the field and simulated experiments. In addition, it was found that multi-trait analysis is justified in situations where the traits present low/moderate heritability and medium/high genetic correlations among them. It also was found that GY in common bean may be outstandingly favored in accuracy when evaluated jointly with its components. Thus, one can conclude that the simultaneous analysis of GY and yield components in common bean may provide superior gains accuracy and selection efficiency, being the response to selection better obtained by use of the multi-trait heritability.

 
AttachmentSize
Applications of multi-trait selection in common bean using real and simulated experiments582.07 KB
Back to top