Genetic variability and stability analysis of multi environments trials for durum wheat grain yield in Ethiopia

.


INTRODUCTION
Plant breeders usually evaluate a series of genotypes across environments before a new improved variety is released for production (Assefa et al., 2020).Ethiopia is the largest producers in the sub-Saharan countries.It ranks third and second in terms of area coverage and productivity in respective order.Both bread and durum wheat is the two species grown in the country.According to CSA 2021 the national wheat productivity estimated about 3 tons /ha that is below the world average.Durum wheat is among the major cereal crops grown in Ethiopia since antiquity because of its wide adaptation to the different agro-ecologies of the country.There are a number of factors associated to the low productivity of wheat in Ethiopia.Genotype x Environment interactions and production of grain crops under rain fed environment is one that can be mentioned among the most important challenges in almost all wheat growing environment of Ethiopia.Genotype x Environment interactions and Despite its importance, the productivity of durum wheat is very low.The use of local cultivars and biotic and a biotic stresses are partially attributed to the low productivity of durum wheat.To improve the productivity, thousands of different local crosses locally adapted landraces and exotic materials are screened in preliminary screening stages.
Those that show outstanding performances are usually selected to be tested in more number of replications and locations under national variety trial.On the other hand, identification of high yielding and stable genotypes are one of the major objectives of any breeding program and need understanding of interaction among genotypes and environment.Stability analysis using AMMI can be an effective and appropriate for reliable yield estimates (Zobel et al., 1988) and it shows the association of genotype and environment (Crossa, 1990).Therefore, the identification of genotypes that are high yielding with better stability across environments under high moisture wheat growing environment have become the major concerns in the national durum wheat research program.The aim of the study is (i) to identify mega-environments in major wheat growing regions of Ethiopia (ii) to identify high yielding and stable performing genotypes

MATERIALS AND METHODS
Forty-nine durum wheat genotypes promoted from the preliminary stages and two standard checks were evaluated across six testing locations from 2020 to 2021 cropping seasons (Table 1).A row column design with three replications was used in all environments.Each plot consisted of six rows, 2.5 m long, and 20 cm apart.Fertilizer was applied all at planting uniformly at the rate of 100 kg/ha NPS and split application of urea, one third (50 kg /ha) at planting and the remaining 2/3 (100kg/ha).Seeding was done manually in rows at the rate of 125 kg/ha per entry as per the recommended time at each respective environment and under similar conditions.Weeds were also pulled out as required in the same manner, at the respective environment.
The following data were recorded on the central four rows of each plot: days to 50% heading, days to maturity, days to maturity, plant height, hectoliter weight, thousand kernel weight, and grain yield.Grain yield adjusted at 12.5% moisture content.Broad-sense heritability (H b ) was calculated using the formula presented in Allard (1960): Broad sense heritability=Hb= (σ2g/σ2p) × 100 Data were subjected to analysis of variance (ANOVA) for each environment separately; and combined analysis of variance was conducted to determine the effect of environment (E), genotype (G) and G x E interaction on the expression of traits.Individual and combined data as well as AMMI analysis was conducted for grain yield using SAS-9.1.andR-statistical software procedure (R 4.3.1, 2018).
were ranged from low to moderate.It varied from 49 % (CD) to 72 % (Sinana) suggesting that both genetic and non-genetic actions had significant role for the control of grain yield at Chefe-Donas where broad-sense heritability was relatively low and indirect selection using high heritable traits would be more effective to improve grain yield (Fetemeh et al., 2023).
The grain yield and other agronomic data for top yielding genotypes and a standard check tested at 12 environments, are shown in Table 3.Based on comparison made, there was a tendency for test genotypes to produce better grain yield than the best standard check variety, Utuba.The average grain yield advantages of the top five lines over Utuba estimated about 7.7% indicating the existence of potential of genotypes that showed better performance over the control checks despite high stem rust incidence (Table 3).The stem rust incidences were high, ranging between 15 MR/MS and 60s.In general, Protein content of the promising lines genotypes was also higher than the standard check, Utuba.
Comparable performances were also observed between the lines and the standard check variety in days to heading, plant height and thousand kernels weight (Table 3).

Analysis of grain yields in different environment
The combined analysis of variance revealed differences in grain yield (p<0.01) among environment (E), genotypes (G) and G x E interaction (Table 4).Significant variations of the G x E indicates the differential responses of genotypes in different environments.Therefore, the differences in average grain yield of durum wheat varieties across environments requires further analysis improve selection and provide better varietal recommendation.As a strategy in order to identify stable and more adapted variety in the presence of G X E interaction, AMMI analysis is an efficient tool to partition the factor involved in crop breeding.

AMMI analysis of the G x E interaction
The AMMI analysis indicated that the environmental effect was the largest source of variation which captured 77.3%, followed by G x E interaction, (19%) and genotypes (3.7%) to the total variation.The higher environmental variation for sum of squares suggests the significant differences between some of the environments that resulted in variations in grain yield of durum wheat.Similar findings reported in bread wheat, in that most of the variations (85.4%) was associated to environmental factors and G x E interaction, indicating that identifying the right genotypes is very difficult (Mohammadi et al., 2023;Fetemeh et al., 2023;Bishwas et al., 2021;Ferhat et al., 2019;Sisay and Sharma, 2015).
The sum of squares of G x E interactions was greater compared to the genotypes main effects, suggested the suitability of using AMMI analysis.The first two principal components explained 57.5% of the interactions sum of AMMI graph in mega-environment analysis and genotype evaluation because it explains more G+GE and has the inner-product property of the biplot.In the current study, GGE biplot of 51 durum wheat genotypes evaluated at 12 environments are provided in Figure 2. Lines (G-25, G7 and G11 were identified as the highest yielding ones since they were categorized in one sector of the polygon.The line G-22 and G-47 were the poorest in grain yield with low stability as they were found farthest from all environments.Similar to this finding (Reza et.al 2010) reported that appropriate and efficient discrimination of durum wheat genotypes based on grain yield using GGE biplot.The findings of the study showed DZ-20 (E9), Adet (E7) and Enwary (E10) and DZ-21(E3) could be considered as one of ideal mega environment and would be more suitable to G-25, G7 and G11 lines.

Conclusion
The multi environment analysis on durum wheat revealed the existence of highly significant differences on environment, genotypes and their interactions.Environment was found the largest sources of variations compared to G x E and main effects of genotypes.Genotypes 25, 7 and 11 DW 193667, DW184042 and were identified as the most stable based on AMMI model and GGE biplot analysis.The GGE biplot indicated DZ (E9) Enwary to be an ideal environment followed by Adet (E7) that could be recommended for future wheat program to wards identifying the most adapted genotypes.

Figure 2 .
Figure 2. GGE biplot showing the mean performance and stability of 51 wheat line and environments.

Table 1 .
Description of the study environments for durum wheat lines.
at Kulumsa and DW-183808 (4530.1) at Sinana.Out of 51 genotypes, only two genotypes (DW-184042 and Dw-193667) stood first at two locations each.DW-184042 was found the best yielder at Adet and Debre-Zeit whereas DW-193667 was the highest at Chefe-Donsa estimates for grain yield

Table 3 .
Summary of top five high yielding lines and a standard check with associated traits at 6 locations over two years.