Indices for assessing the stability of wheat under the genotype x environment interaction
Nikolay Tsenov, Todor Gubatov, Ivan Yanchev
Abstract: Objects and purpose: The object of the study is a group of 24 varieties of winter common wheat, which are included in a multi factorial field experiment (MET), consisting of five growing locations during a four-year test period. The way of ranking (rank) of the varieties according to their stability under the conditions of strong genotype x environment interaction (GEI) is analysed, according to different statistical parameters. The aim is to compare the effectiveness of different types of statistical parameters and indices to determine the stability of the genotype in combination with its yield.
Methods: The comparison between the stability parameters was made using routinely recommended statistical analyze for such cases, such as: correlation, principal component and cluster analyzes. Through them, the studied parameters are grouped, according to their relationship with yield, into three main groups. In this way, information is collected showing the extent to which the combination of magnitude and stability of yield, each of the parameters could be assessed correctly if applied alone or in combination with others like it. Key results: The field trail analysed here is characterized by a strong “cross-over” interaction of genotype by environments (GEI). It causes a huge change in yield, amounting to about 80% of the total variation. Genotype, as a factor in the study, causes only about 5% of the variation, but in interaction with the environment, this share reaches almost 1/5 of the total variation. The study locations were selected randomly but proved to be a good prerequisite for estimating yield, because the impact they have on it is different in size and direction. This change in each of them can be predicted with a probability of more than 0.75%, with the exception of Plovdiv location (H2= 0.42). There is no fundamental difference between the popular and the little-known parameters in the way the varieties are arranged according to the set goal.
Conclusions: Significant differences between the effects of the locations are a prerequisite for the establishment of varieties with specific stability, which is important for their proper zoning. The indices Pi, YSi, KR, S(3), S(6), GAI, I, R and WAASBY alone or in combination with each other are sufficiently informative to determine the varieties with the most optimal combination between yield and stability. Many of the little-known indices Di, Ii, SF, R2, PA are fully applicable with the proviso that several of them be used, rather than alone. The lack of correlation with the yield is a reason for each of them to arrange varieties in a unique way. Their combined use with others would increase the accuracy of the information about the combination of yield + stability that each variety has. Objective evaluation of the variety from any group is recommended to be done through statistical programs that could assess the characteristics of GEI.
Methods: The comparison between the stability parameters was made using routinely recommended statistical analyze for such cases, such as: correlation, principal component and cluster analyzes. Through them, the studied parameters are grouped, according to their relationship with yield, into three main groups. In this way, information is collected showing the extent to which the combination of magnitude and stability of yield, each of the parameters could be assessed correctly if applied alone or in combination with others like it. Key results: The field trail analysed here is characterized by a strong “cross-over” interaction of genotype by environments (GEI). It causes a huge change in yield, amounting to about 80% of the total variation. Genotype, as a factor in the study, causes only about 5% of the variation, but in interaction with the environment, this share reaches almost 1/5 of the total variation. The study locations were selected randomly but proved to be a good prerequisite for estimating yield, because the impact they have on it is different in size and direction. This change in each of them can be predicted with a probability of more than 0.75%, with the exception of Plovdiv location (H2= 0.42). There is no fundamental difference between the popular and the little-known parameters in the way the varieties are arranged according to the set goal.
Conclusions: Significant differences between the effects of the locations are a prerequisite for the establishment of varieties with specific stability, which is important for their proper zoning. The indices Pi, YSi, KR, S(3), S(6), GAI, I, R and WAASBY alone or in combination with each other are sufficiently informative to determine the varieties with the most optimal combination between yield and stability. Many of the little-known indices Di, Ii, SF, R2, PA are fully applicable with the proviso that several of them be used, rather than alone. The lack of correlation with the yield is a reason for each of them to arrange varieties in a unique way. Their combined use with others would increase the accuracy of the information about the combination of yield + stability that each variety has. Objective evaluation of the variety from any group is recommended to be done through statistical programs that could assess the characteristics of GEI.
Keywords: genotype x environment; parametric and non parametric stability statistics; stability; wheat
Date published: 2022-04-21
Download full text