Gauch [10] and Gauch et al. [12] reviewed the AMMI and GGE literature, favoring AMMI. Yan et al.

[11] responded to those articles, favoring GGE. Several studies have also been performed comparing GGE biplots and YSi in bean [13], maize [14], and durum wheat [15]; GGE biplots and JRA in maize [16] and triticale [17]; and JRA and AMMI models in cereal crops [18] for stability analysis. However, little is known about rank correlation selleck chemicals among the four statistical methods (AMMI analysis, GGE biplot, JRA, and YSi statistic) applied in a single study. The main objectives of the present study were to (i) compare the statistical methods (AMMI analysis, GGE biplot, JRA, and the YSi statistic) in the ranking of 20 winter wheat genotypes for yield, stability, and yield–stability

and (ii) evaluate rank correlations among the statistical methods on the basis of yield ranks, stability ranks, and yield–stability ranks. Grain yield data obtained from 20 winter wheat genotypes, consisting of 18 breeding lines Stem Cell Compound Library high throughput (G1–G18) and two check cultivars (G19 and G20, representing the landrace “Sardari” and the released cultivar “Azar-2”, respectively), grown in eight test locations representative of winter wheat growing areas in Iran for three consecutive cropping seasons (2003–2005), were subjected to analysis of rank correlation among the four statistical procedures (AMMI, GGE biplot, JRA, and YSi statistic) in the rankings of genotypes. In each environment (location–year combination), the

experimental layout was a randomized complete block design with four replicates. The plot size was 7.2 m2 (6 rows, 6 m long, 20 cm row spacing). The fertilizer rate was 50 kg N ha− 1 and 50 kg P2O5 ha− 1 applied at planting stage. Combined analysis of variance (ANOVA) for grain yield data was performed to determine the effects of environment, genotype, and GE interaction. Four statistical methods were applied to evaluate GE interaction in the wheat MET data. Regression analysis was performed for each of the 20 wheat genotypes based on the method of Eberhart and Russell [5]. The performance of each genotype in each environment was regressed on the means of all genotypes in each environment. Genotypes with regression coefficient (b) of unity and variance of regression Protein kinase N1 deviations (S2di) equal to zero will be highly stable. The yield stability (YSi) statistic was generated as described by Kang [19] and applied for selecting high-yielding and stable genotypes. Ranks were assigned for mean yield, with the genotype with the highest yield given a rank of 20. Similarly, ranks were assigned for the stability parameter with the lowest estimated value receiving the rank of 1. Stability ratings were computed as follows: − 8, − 4, and − 2 for stability measures significant at P < 0.01, 0.05, and 0.10, respectively; and 0 for the non-significant stability measure.