GENOTYPES OR VARIETIES EVALUATED ACROSS YEARS AND SITES. CROSSED vs NESTED FACTORS

Sahagún-Castellanos, Jaime; Rodríguez-Pérez, Juan Enrique; Peña-Lomelí, Aureliano

doi:10.5154/r.rchsh.2005.11.056

ISSN e: 2007-4018 / ISSN print: 2007-4018

English | Español

Search

Advanced search

Vol. 11, issue 2 July - December 2005

Share

Previous article |

GENOTYPES OR VARIETIES EVALUATED ACROSS YEARS AND SITES. CROSSED vs NESTED FACTORS

EVALUACIÓN DE GENOTIPOS O VARIEDADES A TRAVÉS DE AÑOS Y LOCALIDADES. FACTORES CRUZADOS vs ANIDADOS

Jaime Sahagún-Castellanos; Juan Enrique Rodríguez-Pérez; Aureliano Peña-Lomelí

Keywords: Variance component estimation, annual crops, heritability, response to selection, environmental interaction

10.5154/r.rchsh.2005.11.056

Received: 2005-11-12
Accepted: 2004-04-14
Available online: 2016-07-01
Pages:317-321

Abstract Highlights

In series of experiments conducted in several locations (L) during a years (A) to evaluate genotypes (G) associated with annual crops in several levels of the main plot factor B the definition of the model should consider the true relationship between A and L. Considering that choice of the wrong model affects quality of analysis, the objective of this paper was to study the effect on the estimators of variance components, heritability, and response to selection, of the use of a model based on the common consideration that A and L are crossed factors (model AxL) when in reality A is nested in L (model A/L). Two cases were studied: a) all factors were random, and b) only B was a fixed effects factor. In both cases the estimators of the variance components for G and GB from the model AxL showed a bias which was the negative value of the variance of GA and GAB divided by a, respectively; and their variances were larger than the ones of the corresponding estimators generated in model A/L, the excess being the variance of the unbiased estimator of their respective bias divided by a2. The underestimation of the variance of G in turn must produce an underestimation of the heritability and response to selection because the phenotypic variance would be equally estimated in both models.

....

Help us, the file don´t exist in the server, please send me a email a chapingo.horticultura@gmail.com, Thank you.