山西省南部花生品种产量稳定性的模型分析及评价

Analysis and Evaluation of Different Models for Yield Stability of Peanut Cultivars in Southern Shanxi

为了探究在不同统计模型下花生品种多性状的表现规律,以山西省南部花生主产区6个试点中的7个花生品种为研究对象,通过Eberhart-Russell、AMMI、DTOPSIS和PCA等4种模型分析方法,综合对比、评价不同花生品种及统计模型的适应性。结果表明,4种模型因统计方法的不同结果存在差异,Eberhart-Russell模型分析中品种晋花8号、花育911和临花9号遗传稳定性更好,而AMMI双标图及参数Di显示花育911和张户坡试点分别在产量变异中有更好的品种稳定性和试点分辨力,DTOPSIS和PCA分析分别以临花9号、花育911和临花9号、晋花8号为综合多性状最优品种。结合品种田间产量表现,AMMI分析能更科学地反映加性遗传模型中的GEI交互效应,而DTOPSIS分析在多品种多性状比较中更有优势。

In order to explore the characteristic expression rules of the peanut cultivars under different statistic models, 6 sites and 7 breeds were surveyed in southern Shanxi. The data were used in 4 models which were Eberhart-Russell, AMMI, DTOPSIS and PCA, to comparatively evaluate different peanut cultivars and the adaptability of the statistic models. The results of all the four models differed because of the different statistic calculation. Jinhua8, Huayu911 and Linhua9 cultivars showed better genetic stability in Eberhart-Russell model, while Huayu911 and Zhanghupo showed better cultivar stability and site discrimination in yield variation for AMMI biplot and Di parameter. Linhua9 and Huayn911 were optimal cultivars with synthetically multiple characters in DTOPSIS model, while Linhua9 and Jinhua8 were optimal in PCA model. With the actual field yield performance, we thought the effect of GEI interaction was more scientifically reflected by additive genetic model in AMMI analysis, the DTOPSIS analysis was more advantageous in multiple cultivars and traits comparison.