Interaction genotype environment in rice to identify mega and ideal environments by the model of site regression and biplot

Abstract

The selection of cultivars based on genetic and environmental effects is insufficient for the plant breeder when they do not consider the interaction genotype by environment (GxE). The GxE studies allow appropriate identification of high performance materials for several or specific environments. The objectives of this study were to detect the GxE to identify mega-environments and their relationship with the genotypes, using the GGE biplot obtained from the analysis of the regression-by-site model (SREG) for yield in irrigated rice cultivars in Venezuela. In 12 environments, distributed in the main producing areas, we evaluated during 2010-2011 six genotypes of the most sown cultivars in that time and the present for the commercial production of rice in the country. A randomized complete blocks design with three repetitions was used with experimental units of 20 m². The analysis of variance of the SREG model detected significance in the first two main components that explained almost 73 % of the GxE variation. The GGE biplot identified two mega-environments with 'Soberana FL' and 'SD 20A' as the best cultivars. The GGE biplot "means and stability" showed the line 'AP06B041' as the most stable genotype and the cultivar Soberana FL as the "ideotype genotype" with high average productivity and stable phenotypically. The best localities of tests for trials of genetic improvement of rice in Venezuela resulted in the Guárico State Experimental Field of INIA in both periods (dry and rainy), and in the Portuguesa State Experimental Field of Sehiveca for the rainy period