“GGE BIPLOT” ASSESSMENT OF PHENOTYPIC STABILITY OF SPRING BARLEY VARIETIES

The article presents GGE biplot analysis of results of environmental trials in 17 varieties of spring barley bred at the Plant Production Institute nd. a V.Ya. Yuriev of NAAS. The study results discriminate genotypes with stable realization of their genetic potential in a number of environments as well as genotypes combining a high level of a trait with its stable expression. The varieties Kozvan, Perl, Agrariy and Kosar were chosen as valuable source material for spring barley breeding. We think that GGE biplot can be a comprehensive alternative to the most conventional methods of assessment of adaptive features in genotypes.

Introduction.Barley (Hordeum vulgare L.) is a strategic export-oriented agricultural plant in Ukraine.Increase in gross output of barley grain is impossible without implementation of high-yielding barley varieties that are resistant to biotic and abiotic factors.
Analysis of publications, pose the problem.Environmental variety trials are an important tool for selection of genotypes with specific (narrow) or wide adaptation to a certain environment or to a range of environments, which enables predicting yield capacity of genotypes under these conditions and ultimately increases farmers' labor efficiency [1,2].Nevertheless, capabilities of environmental trials are not always used to the full: usually yield capacity of genotypes is only analyzed, but information on other traits remains unstudied [3].
The observed phenotypic variance (P) of traits consists of environment variance (E), genotype variance (G) and genotype × environment interaction (GE): P = G + GE + E or P -E = G + GE [4].W. Yan [5] points out that E effect forms the major part of the total phenotype variability, and contributions of G and GE are generally small.However, G and GE effects must be taken into account in the process of selection of high-yielding genotypes.
The term «GGE» emphasizes understanding that G and GE are two sources of variation that are pertinent to genotype assessment and must be considered simultaneously, when genotype × environment interactions are investigated.
With time, GGE biplot analysis has turned into a complex analysis system, as a result of which the majority of environmental trial datum patterns can be displayed graphically [6][7][8][9].
The aim and tasks of the study.The study purpose was evaluation of adaptive features of spring barley varieties in terms of performance and its elements using GGE biplot and discrimination of valuable source material for breeding of this plant.
Material and methods.The source material was 17 varieties of spring barley bred at the Plant Production Institute nd. a V. Ya.Yuriev of NAAS.To determine their adaptive potential, in 2013 environmental trials were conducted in three locations with different soilclimatic conditions: Plant Production Institute nd. a V. Ya.Yuriev of NAAS (Eastern Forest- Steppe)environment E1, Donetsk Experiment Station of NAAS (Southern Steppe)environment E2 and Research Station of Bast Crops of the Institute of Agriculture of Northern -East NAAS (North-Eastern Forest-Steppe)environment E3.In addition to yield capacity, variability of performance elements was evaluated: grain weight per plant, productive tillering, grain number per spike and 1000-grain weight.The environmental trial data were analyzed by GGE biplot.
GGE biplot graphs were constructed using the first two principal components PC1 и PC2 derived from subjecting the data to singular-value decomposition.Only two principal components (PC1 and PC2) are retained in the model because such a model tends to be the best model for visualizing interaction between each genotype and test environments.
Results and discussion.The results of the environmental trials showed a significant differentiation of the studied varieties in terms of plant performance and its elements (table 1).Analysis of variance demonstrated strong significant differences between the gen otypes, environments and their interactions by all the estimated traits as well as differences in influence of these factors on formation of trait level (Table 2).Environment (E) was the dominant factor in productive tillering and grain weitht per plant variances (50.% and 49.7 %, respectively), but this factor is considered as of no importance upon genotype assessment, which allows focusing on the investigation of genotype (G) and genotype × environment interaction (GE) effects [8,10].
Environmental variety trial results are always a large conglomeration of data, which are rather difficult to analyze without visualization.GGE biplot is an ideal tool to solve this problem, enabling discrimination of genotypes realizing their potentials in specific soil-climatic conditions or genotypes with wide adaptation to a variety of test environments.
In Fig. 1 the polygon vertices are genotype markers that are maximally remote from the biplot center, so all the genotype markers are inside the polygon.The lines dividing the biplot into sectors represent a set of hypothetical environments.The genotype forming the polygon angle for each sector dividing the biplot has the highest yield capacity in environments falling within this sector.Thus, the genotype of Kozvan variety (G12) had the maximum productive tillering in all the three environments, suggesting its wide adaptation by this trait.Modern variety (G14) was the winner by grain number per spike in environment E3, and Vitrazh (G6)in environments E1 and E2.Vektor (G3) in environments E1 and E2 and Perl (G16) in environment E3 were noticeable for 100-grain weight.In environment E3 Agrariy (G1) variety had the highest performance, and in environments E1 and E2 Kozvan and Vitrazh varieties, which were similar by their parameters, showed the highest performance.
GGE biplot ranks genotypes by their performance and stability in a number of environments.In Fig. 2 the average tester coordinate (ATC) (X-axis) or the performance line passes through the biplot origin with an arrow indicating the positive end of the axis.The ATC Y-axis (stability axis) passes through the biplot origin and is perpendicular to the ATC X-axis.Thus, the mean value of a trait of a genotype is estimated by the projection of its marker to the ATC X-axis, and stability -by the projections to the ATC Y-axis.
Vektor (G3) and Perl (G16) varieties were distinguished for 1000-grain weight; Vektor (G3) and Parnas (G15) were the most stable.Genotypes selected by level and stability of traits are valuable as source material for breeding.
In AV Kilchevskyy, LV Khotyleva and VV Khangildin methods there is a very important integral parameter -breeding value of genotype‖, which provides a comprehensive assessment of genotypes in terms of yield capacity and its stability.GGE biplot also ranks genotypes by -breeding value‖.The center of concentric circles (Fig. 3) represents the position of a genotype with maximum -breeding value‖ or so-called -ideal‖ genotype.The closer a genotype to the ideal one is, the more valuable it is.In our studies Kozvan (G12) variety was of the greatest breeding value in terms of productive tillering; Kosar (G13) varietyin terms of grain number; the awnless variety of Vektor (G3)in terms of 1000-grain weight; Perl (G16) varietyin terms of performance, because it was much more stable than Kozvan variety, which exceeded Perl by performance (see Fig. 3).
The results of GGE biplot analysis of adaptive features of spring barley varieties very closely correlate with the results that we obtained by AV Kilchevskyy, LV Khotyleva method [11,12], but GGE biplot has a number of advantages over the latter, in particular, it does not require heavy calculations.Conclusions.Use of GGE biplot enabled analyzing the environmental trial data and discriminating the most valuable genotypes.Among the studied varieties of spring barley, Kozvan variety was the most valuable by productive tillering, Kosar varietyby grain number, Vektor varietyby 1000-grain weight, Perl and Kozvan varieties -by performance.
Thus, GGE biplot can be a comprehensive alternative to the most conventional methods of assessment of adaptive features in genotypes.

3 .
GGE Biplot Based on Genotype-Centered Scaling for Comparison of Genotypes with the -Ideal‖ Genotype by Productive Tillering (A), Grain Number per Spike (B), 1000-Grain Weight (C), Plant Performance (D)