Clustering Environments to Minimize Change in Rank of Cultivars

Abstract

Crossover interactions occur in evaluation trials when ranks of cultivars change across environments. Determining groups of environments within which crossover interactions are minimized may facilitate making cultivar recommendations. The goal of this research was to test a new approach for determining these environmental groups in which crossover interaction between a pair of cultivars was defined across all environments. The number of groups was based both on reduction in crossover interaction and repeatability of cultivar means within groups. The validity of this procedure was tested on three simulated data sets with known crossover interactions. For each data set, the approach divided the environments into the two groups that minimized crossover interactions. The approach also was applied to yield data from a maize (Zea mays L.) trial in which 59 environments previously had been clustered by a different measure of crossover interaction. Three groups of 12 environments and one group of 23 environments were defined. The previous clustering had identified six clusters. The amount of crossover interaction within the four environmental groups was reduced by 53% from the total crossover interaction in all 59 environments. The results of the clustering depended on whether all pairwise comparisons among cultivars or only the significant crossover interactions among the higher yielding cultivars were used. The latter method was deemed more appropriate when the goal is to recommend specific cultivars for specific groups of environments. Regardless of the approach used, clustering based on crossover interactions only has practical significance if these interactions are repeatable.

Understanding grain yield: it is a journey, not a destination

“…If a breeder was interested in improving grain yield in low yielding environments, WI(CNN1D) increased grain yield compared to WI in low yielding environments, but reduced grain yield when compared to WI in high yielding environments (a classic crossover interaction; Haldane 1946;Russell et al 2003). The mean over environments of WI (2.72 mg/ha) was very similar and not statistically different from WI(CNN1D) (2.71 mg/ha), but the response of the two lines to the environment was very different.…”

Isoyield Analysis of Barley Cultivar Trials in the Canadian Prairies

“…2002), but has not been explicitly assessed. Some bilinear models have been used to develop the tests for negligible crossover GEI within a cluster (Crossa and Cornelius 1997, Russell et al. 2003) for determining different clusters when balanced data are available.…”

“…2002, Truberg and Hühn 2002). The tests for negligible crossover GEI within a cluster based on bilinear models such as the shifted multiplicative model (SHMM) or site regression model (SREG) (Crossa and Cornelius 1997, Russell et al. 2003) have been suggested for determining different clusters but these models developed for the analysis of balanced data are generally not easily extended for the analysis of unbalanced data, an unavoidable reality for almost all cultivar trials.…”

Improving Lives: 50 Years of Crop Breeding, Genetics, and Cytology (C‐1)

“…The usefulness of these different approaches depends on the complexity of the G × E, with multivariate approaches being more appropriate for interactions of high dimensionality. In the past two decades, there has been an increased focus on using crossover interactions (a type of G × E in which significant changes in genotype ranks occur across environments) to classify environments (Baker, 1988; Crossa et al, 1993; Russell et al, 2003). Genotype × environment interactions also have been described by measuring actual attributes of the environment (Wood, 1976; Beckett, 1982; Baker, 1990).…”