In contrast, G8, G6, and G9 were among the genotypes with the lowest stability and with higher (G8) and lower (G6 and G9) mean yield performances than the overall mean. The yield, stability and yield–stability ranks for 20 tested genotypes in 24 environments based on each of the statistical methods mentioned above are given in Table 3. Comparison of the statistical methods PI3K targets based on the yield ranks showed that the methods generally gave similar results in the ranking of genotypes. For example, the five top-ranked

genotypes based on AMMI were G4, followed by G10, G19, G1, and G17; based on the GGE biplot were G4 followed by G10, G1, G19, and G17; based on JRA were G8, G4, G1 = G12, and G10; and based on the YSi statistic were G4, G10, G19, G1, and G17. With respect to stability ranks, genotypes G2, G15, G12, G11, and G17 were found to be stable based http://www.selleckchem.com/products/r428.html on AMMI distance, whereas the five top-ranked genotypes based on the GGE biplot were G18 = G12 = G2, G14, and G3, showing that AMMI and the GGE biplot gave similar results in identifying two of the five top-ranking genotypes as stable. According to JRA the most desirable genotypes based on stability ranks were G2, G17, G10,

G16, and G3, and based on the YSi statistic the most stable genotypes were G2, G17, G3, G16, and G18. Similar stability ranks were assigned by the JRA method and YSi statistic, as they identified four of the five top-ranking genotypes as stable. For yield–stability, the AMMI analysis identified G10 followed by G17, G3, G15, and G12 as the top-ranking high-yielding and stable genotypes; whereas G18 followed by G17 = G12 and G4 = G10 were characterized

by the GGE biplot as high-yielding and stable. According to JRA, the top-ranking high-yielding and stable genotypes were G10, followed by G4, G12, G17, and G3, and based on the almost YSi statistic the highest-ranking genotypes were G4 = G10, G17, G19, and G1 = G18. All four methods identified G10 and G17 as among the five top-ranking high-yielding and stable genotypes. Significant rank correlations were found between the statistical methods in the ranking of genotypes for yield, stability and yield–stability (Table 4). With respect to yield, the statistical methods were significantly correlated (P < 0.01) in the ranking of genotypes. The correlations varied from 0.72 (JRA–YSi; P < 0.01) to 0.99 (GGE–AMMI; P < 0.01) indicating that AMMI and the GGE biplot agreed most closely in ranking genotypes for yield. The statistical methods were positively correlated in identifying stable genotypes. The Spearman’s rank correlations for stability indices ranged from 0.53 (GGE–YSi; P < 0.05) to 0.97 (JRA–YSi). The AMMI distance (AMMID) was highly correlated with the stability indices in JRA (r = 0.83; P < 0.01) and YSi (r = 0.86; P < 0.01).