Work

Application of mixed model and spatial analysis methods in multi-environmental and agricultural field trials.

Year: 2015
Type: dissertation
Abstract: Agricultural experimentation involves selection of experimental materials, \nselection of experimental units, planning of experiments, and collection of \nrelevant information, analysis and interpretation of the results. An overall \nwork of this thesis is on the importance, improvement and efficiency of variety \ncontrast by using linear mixed mode with spatial-variance covariance compare \nto the usual ANOVA methods of analysis. A need of some considerations on the \nrecently widely usage of a bi-plot analysis of genotype plus genotype by \nenvironment interaction (GEE) on the analysis of multi-environmental crop \ntrials. An application of some parametric bootstrap method for testing and \nselecting multiplicative terms in GGE and AMMI models and to show some \nstatistical methods for handling missing data using multiple imputations \nprincipal component and other deterministic approaches. \nMulti-environment agricultural experiments are unbalanced because several \ngenotypes are not tested in some environments or missing of a \nmeasurement from some plot during the experimental stage. A need for \nimputation of the missing values sometimes is necessary. Multiple \nimputation of missing data using the cross-validation by eigenvector method \nand PCA methods are applied. We can see the advantage of these methods \nhaving easy computational implementation, no need of any distributional or \nstructural assumptions and do not have any restrictions regarding the pattern \nor mechanism of missing data in experiments. \nGenotype by environment (G×E) interaction is associated with the differential \nperformance of genotypes tested at different locations and in different years, \nand influences selection and recommendation of cultivars. Wheat genotypes \nwere evaluated in six environments to determine the G×E interactions and \nstability of the genotypes. Additive main effects and multiplicative interactions \n(AMMI) was conducted for grain yield of both year and it showed that grain \nyield variation due to environments, genotypes and (G×E) were highly \nsignificant. Stability for grain yield was determined using genotype plus \ngenotype by environment interaction (GGE) biplot analysis. The first two \nprincipal components (PC1 and PC2) were used to create a 2-dimensional GGE \nbiplot. Which-won where pattern was based on six locations in the first and five \nlocations in the second year for all the twenty genotypes? The resulting pattern \nis one realization among many possible outcomes, and its repeatability in the \nsecond was different and a future year is quite unknown. A repeatability of \nwhich won-where pattern over years is the necessary and sufficient condition \nfor mega-environment delineations and genotype recommendation. \nThe advantages of mixed models with spatial variance-covariance structures, \nand direct implications of model choice on the inference of varietal \nperformance, ranking and testing based on two multi-environmental data sets \nfrom realistic national trials. A model comparison with a ᵪ2-test for the trials in \nthe two data sets (wheat and barley data) suggested that selected spatial \nvariance-covariance structures fitted the data significantly better than the \nANOVA model. The forms of optimally-fitted spatial variance-covariance, \nranking and consistency ratio test were not the same from one trial (location) to \nthe other. Linear mixed models with single stage analysis including spatial \nvariance-covariance structure with a group factor of location on the random \nmodel also improved the real genotype effect estimation and their ranking. The \nmodel also improved varietal performance estimation because of its capacity to \nhandle additional sources of variation, location and genotype by location \n(environment) interaction variation and accommodating of local stationary \ntrend. The knowledge and understanding of statistical methods for analysis of \nmulti-environmental data analysis is particularly important for plant breeders \nand those who are working on the improvement of plant variety for proper \nselection and decision making of the next level of improvement for country \nagricultural development.
Language: English
Subfield: Plant Science
Open Access status: green
OpenAlex ID: W2554253903
Cites: 133
Cited by: 1
Related to: 10