Identification of peach accessions stability and adaptability in non-balanced trials through years

Abstract

Identification of genotypes with acceptable yield and yield stability in different environments is an important issue in plant breeding. Genotype-by-environment interaction (GEI) can alter genotypes performances making the selection of superior material a tedious task for breeders. Consequently, it is necessary to assess the usefulness of different available methods and identify the most suitable for understanding GEI. The objectives of this work were to compare three methods to study genotype stability considering incomplete data sets: (i) Di Rienzo, Guzmán and Casanoves' test (DGC), (ii) relative yield (RY) and (iii) Piepho's method. In addition, AMMI (additive main effect and multiplicative interaction) analysis and eight AMMI stability measures SIPC, EV, ASV, Da, FP, B, FA and Za were computed to explore their advantages and disadvantages to select stable entries. The usefulness of the genotype selection index (GSI) and the rank-sum (RS) procedures to identify stable and high-yielding genotypes were evaluated and then compared with the superiority (P) and reliability indexes (I). The association between yield variation and climatic factors as frosts, chilling, heat, rainfall and the interactions among them were also analyzed. 29 peach entries were assessed in four to seven seasons in a completely randomized design with three replications. DGC and RY tests agreed on classifying Fireprince as a stable and high-yielding peach, RY classified 25 entries as stable, while Piepho's method did not separate the tested genotypes as DGC and RY did. The results of AMMI indicated that 25.06% of total variability was justified by genotypes, 9.76% by environments and 58.97% by GEI. The first five interaction principal components could explain 94.82% of GEI and showed the efficiency of AMMI model to study and understand GEI. The AMMI parameters showed no association with fruit yield, therefore, they could be useful to indicate stable entries but they would not be appropriate to select stable and high-yielding genotypes. The EV and Za indicated static stability while ASV, SIPC, Da, FA and FP pointed out the dynamic stability concept. The performance of the best entries selected by GSI, RS, P and I procedures were not different, therefore, any of them can be used to select superior peach genotypes. Rainfall during endodormancy, rainfall from floral bud endo- to ecodormancy - and heat accumulation during fruit development period showed significant correlation with yield variation across seasons.