Premium
Visual Scorings of Drought Stress Intensity as Covariates for Improved Variety Trial Analysis
Author(s) -
Mühleisen J.,
Reif J. C.,
Maurer H. P.,
Möhring J.,
Piepho H.P.
Publication year - 2013
Publication title -
journal of agronomy and crop science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.095
H-Index - 74
eISSN - 1439-037X
pISSN - 0931-2250
DOI - 10.1111/jac.12025
Subject(s) - analysis of covariance , covariate , statistics , triticale , heritability , interaction , mathematics , agronomy , biology , genetics
Abiotic stress can enhance the heterogeneity of fields, which leads to imprecise estimates of genotypic effects in variety trials. Our study is based on multilocation field trials of triticale (× T riticosecale W ittm.). Two of the six locations were affected by drought stress and showed increased field heterogeneity. At W illstätt, drought stress intensity was visually scored twice, and at I ssoudun, overall impression of a plot was scored once. We investigated if the visual scorings can be used as covariates in an analysis of covariance ( ancova ) to eliminate the influence of different drought stress intensity on plot yields. For evaluation of the ancova models, we examined if the covariates were independent from the genotypic effects and linearly associated with grain yield. In addition, ancova models were compared with baseline and spatial models based on AIC and phenotypic correlation between genotype means estimated with the model under investigation in a drought stress location with genotype means calculated across the remaining locations. We found that both scorings in W illstätt fulfilled the requirements of an ancova and led to an increase in broad‐sense heritability ( h 2 ) and efficiency. ancova with the second scoring increased h 2 from 0.03 for the baseline model to 0.60, whereas the best spatial model increased h 2 only up to 0.50. The scoring at I ssoudun was not independent from the genotypic effects and reduced phenotypic correlations. We concluded that environmental factors causing spatially differing yield potential can be scored or measured and used as covariates to obtain more precise genotypic estimates.