Partitioning of Genotype-Environment Interaction in Sugarcane (Saccharum spp.): Variance Component Analysis

Genotype-environment (GE) interaction in sugarcane (Saccharum spp.) is very important under Ethiopian agroecological and climatic conditions. However, yield trials conducted in Ethiopia are customarily handled using fixed models which consider the crop year and location effects as fixed. The study was conducted to partition and compare the relative importance of the components GE interaction and estimate the repeatability of yield trials using both fixed and mixed models. Forty-three genotypes along with the six commercial varieties were evaluated over four crop years and across five locations using simple lattice design. Data of yield, yield components and yield qualities were collected and subjected to analysis of variance (ANOVA) and variance component analysis. Variance component analysis indicated large and highly significant estimates of genotypic and Genotype-Location-Crop year (GLC) interaction effects for all yield traits thereby suggesting strong genotypic effect and excess crossover interaction, respectively. Regarding to the relative importance of the components of the genotype-environment (GE) interaction, genotype effect was more important for cane yield components and yield quality traits than others while GLC interaction was more important for cane and sugar yields than other components. Generally, the variance component and broad sense heritability estimates revealed the effect of GLC interaction was prevalent. It implicated a substantial selection gain could be attained from conducting variety trials across locations and over crop years under Ethiopian agro-climatic conditions. Moreover, the heritability estimates obtained from GLC and GE models indicated better repeatability of the experiments in which the later model showed overestimation of broad sense heritability values. Thus, we recommend the use of GLC model in sugarcane METs conducted across locations and repeated over crop years