A interação dos genótipos x ambientes na cultura da soja e o risco na recomendação de novas linhagens

Abstract

The greatest challenge of the main breeding programs for all cultivated plants is using information obtained from previous experiments, recommending new cultivars to be used in different management locations and future years with the lowest error possible. When concerning soybean culture, this fact has even greater importance, given that the seed market is very competitive, and the culture is cultivated in more than 33 million ha in Brazil, encompassing nearly all states. Because of this, it is expected that the decision made by breeders involve an expressive participation of lines x environment interaction. Thus, the quantification of the lines x environment interaction, at the final phase of recommendation, is important. However, it is not enough to simply estimate the interaction. It is necessary to provide breeders with easily interpreted alternatives to reduce the risk in recommending new strains. In this context, the objectives of this work were to quantify genotype x environment interaction and verify the participation of lines x locations interaction in relation to lines x years, in order to contribute to the future planning of Value for Cultivation and Use (VCU) trials, and propose a new risk evaluation method for adopting a strain. For this, data from the soybean-breeding program of the company Tropical Melhoramento & Genética (TMG) were used. We analyzed data corresponding to grain productivity of soybean strains, grouped according to maturation (Group 1 - 6.5; 6.8 and 7.1, Group 2 – 7.4; 7.7 and 8.0 and Group 3 – 8.3; 8.6 and 8.9), obtained from the VCU trials involving 71 environments in different years and locations. Experimental details from conducting the experiments met the recommendations of the Ministério da Agricultura, Pecuária e Abastecimento – MAPA (Ministry of Agriculture, Livestock and Supply). We detected a significant difference between strains in 98% of the experiments. For all maturation groups, the participation of lines x environment interaction for total variation was expressive. The contribution of the strain x location interaction was nearly always greater than the lines x years interaction. One of the reasons for this difference was the small number of strains evaluated in two or more successive years. The proposed method for estimating the adoption risk in the polar graphic method classified the strains coincidentally to the Annichiarico methodology (1992).