AMMI Analysis for Stability and Genotype by Environment Interaction on Common Bean (Phaseolus vulgaris L.) Genotypes in Mbeya Region, Tanzania

Abstract

A significant Genotype by Environment Interaction (GEI) makes selection of stable genotypes difficult. This study was conducted to establish the effect of GEI on yield of Common bean genotypes and reduce complaints on the under performances. Eighteen (18) Common bean genotypes were assessed for variation in gene expression linked to yield and yield predictors on three different districts in Mbeya region (Mbarali, Mbozi and Mbeya districts).  Regression, pooled ANOVA and AMMI biplot models were used to evaluate the data. Variety performance showed significant variations in yield between the districts.  A similar scenario was observed in regard to yield predictors. Regression analysis showed that in Mbarali 50% was the significant yield predictor (P = 0.027) while pods/ plant was the trait mostly linked to yield in Mbozi. (GEI) analysis using the AMMI model revealed that best variety performance by location based on yield. Interaction principle component (IPC1) was highly significant (P = 0.0001) and contributed about 69.1% of GEI variation.  The genotypes SER 83 and RCB 266 where highly adaptable in Mbarali site.  The genotypes SER 45 and KG 521 showed specific interaction with the environment of Mbozi district. A total of five genotypes proved to be superior in Mbeya district.  The most adapted stable variety with highest grand mean yield across all three mega environments was RCB233 (IPC1= 0.07, yield = 1073 t/ha). The environment in Mbarali was found to be most predictable for evaluation of Common bean genotypes.