Genotype X Environment Interaction and Yield Stability of Bread Wheat Genotypes in South East Ethiopia

Transcription

1 World Journal of Agricultural Sciences (3): 2-27, 25 ISSN IDOSI Publications, 25 DOI:.5829/idosi.wjas Genotype X Environment Interaction and Yield Stability of Bread Wheat Genotypes in Sou East Eiopia Melkamu Temesgen, Sentayehu Alamerew, Firdissa Eticha and Muez Mehari Eiopia Commodity Exchange, P.O. Box 734, Addis Ababa, Eiopia 2 Colleges of Agriculture and Veterinary Medicine, Jimma University, Eiopia 3 Eiopian Institute of Agricultural Research, Kulumsa Research Center, P.O. Box 489, Asella, Eiopia 4 Tigray agricultural research institute, Alamata Agricultural Research center, P.O. Box 56 Alamata, Eiopia Abstract: Twenty bread wheat genotypes and two checks were evaluated at six test locations of Central and Sou Eastern Eiopia 23/4 during main growing season in a randomized complete block design using four replications. The objective of e study was to quantify e magnitude of genotype by environment interaction and yield stability of bread wheat genotypes. The additive main effect and multiplicative interaction effect model (AMMI) analysis revealed significant difference at (P=.) for genotype, location and genotype by location interaction for e response variable grain yield. Accordingly, e first and second IPCAs share 6.77 and 5.7% of e G x E SS captures 2.68 %.The AMMI model clearly indicate e genotype by environment interaction by partitioning into ree significant component taking 89.7% of genotype by environment variation 2.68% of e variation in e genotype by environment interaction. According to e Stability analysis of e additive main effect and multiplicative interaction, biplot analysis ranking biplot e genotype Digelu and ETBW6738 were e most stable coupled wi higher grain yield where as e genotypes Danda a and ETBW6732 were unstable.in each tested location. Based on e AMMI, environments were highly variable bo for main and interaction effects. Bokoji, Kulumsa, Sinana, Holeta and Asassa were favorable environments. Areka as low yielding (unfavorable) environments for majority of e traits. In general, e results of e analysis showed at bread wheat is highly sensitive to environmental changes and is necessitates screening of cultivars for wide and specific adaptation. Based e one year data e genotype Digelu were stable accompanied wi higher yield but in order to recommended appropriate genotypes to e sou Eastern Eiopia, repeating e trial is vital. Key words: AMMI Bread wheat Genotype-by- environment interaction Stability INTRODUCTION dominating e food habit and dietary practices and known to be a major source of energy and protein for e Bread wheat (Triticum aestivum L.), togeer wi highland population [4]. durum (Triticum turgidum L.) are e most widely The regions identified as highly suitable for distributed crops worldwide and of primary importance wheat production include Arsi, Sou - Shewa, for human nutrition []. Its high productivity contributed West- Shewa, Nor - Shewa, Ilubabor and Western - to e prevention of widespread food shortage and Harerge, Sidamo, Tigray, Nor - Gonder, Bale and stabilized e food security of countries like Asia, Latin Gojam. In some regions wheat covers an area of America and to some extent in Africa [2] ha wi production of t and ranks Eiopia is e second largest wheat producer, after four next to maize, sorghum and barely in terms of Sou Africa, in Sub- Saharan Africa. Wheat covers an productivity; four in terms of production next to area of ha wi production of t sorghum, teff, barley and ird in area production next to [3].Wheat has been one of e major cereals of choice, teff, sorghum [5]. Corresponding Auors: Sentayehu Alamerew, Colleges of Agriculture and Veterinary Medicine, Jimma University, Eiopia. 2

2 World J. Agric. Sci., (3): 2-27, 25 Average national grain yield of wheat in Eiopia is MATERIALS AND METHODS estimated to be around 2.2 t/ha, National average yield is.8 t/ha in e Arsi Zone region [3]. which is lower as Experimental Design: The trial was conducted in six compared to e world average productivity, which locations during 23 crop season and e detailed is 2.7 t/ha [6]. which is far below from experimental descriptions of e locations (Table ). 22 bread wheat yields of over 5 t ha [7]. This is because of e many genotypes (Table 2) were laid out in randomized complete production constraints of which e major ones block design (RCBD) wi four replications. A total of 6 include: poor soil fertility, high incidence of weeds, rows wi row spacing of.2m and wi a total plot size of pests and diseases, drought and shortage of well-.2 meter by 2.5 meter and spacing between plots was.5 adapted improved varieties. All ese constraints meter while spacing between block was maintained at result in yield fluctuation from season to season and meter. Seed rate of 5 kg ha and planting was made by from location to location [8]. Thus e huge gap between drilling to e six rows. A fertilizer rate of 4 kg N ha and e research centers by which e genotype being 46 kg P2O 5 ha of split application were used during released and e target environment farmers field could sowing and tillering period and data for yield were be due to e less representation production environment collected from e four middle rows. at leads to e significant genotype by environment interaction. Statistical Analysis: The additive main effect and The National wheat breeding program in Eiopia had multiplicative interaction effect (AMMI) were done based started about half a century ago [8]. Nationally, still 4 on e model suggested by [3].Using e crop stat 7.2 improved bread wheat varieties had been released. software released by e international rice research Testing of nationally released and exotic introduced institute wheat genotypes have been continously under going by e Kulumsa Agricultural Research Institute (KARI) for n eir high yielding performance and adaptation. yij = + Gi + Ej + ( KU n nisnj ) + Qij + eij The GxE interaction reduces association between phenotypic and genotypic values and us, a genotype where: (i =, 2.22: j =.6); Y ij = The at performs well in a given environment may not performance of e i genotype in e j environment; µ= necessarily respond well in oer environment. So if The grand mean; G i = Additive effect of e i genotype environments are sufficiently different, GxE interaction (genotype mean minus e grand mean); K n = Eigen value can result in different yield ranking of evaluating of e PCA axis n,; E j = Additive effect of e j genotypes. In addition, e relationship between selection environment (environment mean deviation); U ni and S nj = environments and target production environment had Scorer of genotype i and environment j for e PCA axis been a fundamental problem because many of e selected n; Q ij = Residual for e first n multiplicative components activities performed by e conventional approach are in and ; e ij = error. on-stations which are good production environments [9]. To see e yield stability analysis, e formula Multi-environment yield trails are essential in estimation suggested by [4] was used and e output of e crop of genotype by environment interaction (GEI) and stat software was calculated using e Microsoft excels identification of superior genotypes in e final selection 2. cycles [,]. Different meods are commonly used to analyze SS IPCA 2 2 ASV ( IPCA score) IPCA2sccore SS multi location yield trial data to reveal patterns of IPCA2 genotype by environment interaction. The additive main where effects and multiplicative interaction (AMMI) model gives ASV= AMMI stability value good insight in separating complicated patterns of IPCA = interaction principal component analysis. genotypic by environment interaction by using combining IPCA2 = interaction principal component analysis 2. e classical analysis of variance and e principal SSIPCA = sum of square of e interaction principal component analysis [2].This study is initiated wi e component one. objective of to evaluate e yield performance and SSIPCA2 = sum of square of e interaction principal stability of bread wheat genotypes. component two. 22

3 World J. Agric. Sci., (3): 2-27, 25 Table : Description of e study site Locations Annual Rainfall (mm) Altitude (m.a.s.l.) Latitude Longitude Temperature( C) Soil type Kulumsa ' ''N 39 9' ''E Luvisol Bokoji ' 6''N 39 5' ''E Niosols Asassa N' E' Clay loam Holeta N' E' Red Sinana ' 6''N 4 2'''E -22 clay Areka '.''N ''E 5-3 Sandy loam Source: BoARD (23) Table 2: Pedigree and Selection history of materials used in is study Entry Genotype Pedigree Danda'a Breeder Seed (Check) 2 ETBW 6728 ROELFS F27 3 ETBW 6729 TRCH/SRTU/5/KAUZ//ALTAR84/AOS/3/MILAN/KAUZ/4/HUITES 4 ETBW 673 WAXWING*2/6/PVN//CAR422/ANA/5/BOW/CROW//BUC/PVN/3/YR/4/TRAP# 5 ETBW 673 FRET2*2/4/SNI/TRAP#/3/KAUZ*2/TRAP//KAUZ/5/PARUS/6/FRET2*2/KUKUNA 6 ETBW 6732 PBW343*2/KUKUNA*2//YANAC 7 ETBW 6733 FRET2/KUKUNA//FRET2/3/PARUS/5/FRET2*2/4/SNI/TRAP#/3/KAUZ*2/TRAP//KAUZ 8 ETBW 6734 ROLF7*2/KIRITATI 9 ETBW 6735 ROLF7/YANAC//TACUPETO F2/BRAMBLING ETBW 6736 WBLL/KUKUNA//TACUPETO F2/5/WAXWING/4/SNI/TRAP#/3/KAUZ*2/TRAP//KAUZ ETBW 6737 WAXWING/6/PVN//CAR422/ANA/5/BOW/CROW//BUC/PVN/3/YR/4/TRAP# 2 ETBW 6738 FRET2*2/4/SNI/TRAP#/3/KAUZ*2/TRAP//KAUZ*2/6/PVN//CAR422/ANA/5/BOW/CROW//BUC/PVN/3/YR/4/TRAP# 3 ETBW 6739 SERI.B//KAUZ/HEVO/3/AMAD*2/4/KIRITATI 4 ETBW 674 PBW343*2/KHVAKI//PARUS/3/PBW343/PASTOR 5 ETBW 674 FRET2*2/4/SNI/TRAP#/3/KAUZ*2/TRAP//KAUZ/5/PFAU/WEAVER//BRAMBLING 6 ETBW 6742 WBLL//UP2338*2/VIVITSI 7 ETBW 6743 WAXWING/WHEAR//WAXWING/KIRITATI 8 ETBW 6744 FRET2/KUKUNA//FRET2/3/YANAC/4/FRET2/KIRITATI 9 ETBW 6745 TRCH//PRINIA/PASTOR 2 ETBW 6746 BAV92//IRENA/KAUZ/3/HUITES/4/DOLL 2 ETBW 6747 PBW343*2/KUKUNA//PARUS/3/PBW343*2/KUKUNA 22 Digelu Breeder Seed (Check) RESULTS AND DISCUSSION The AMMI analysis (Multiplicative effect) was furer demonstrated ree significant interaction Additive Main Effect and Multiplicative Interaction principal components and according to e significant Analysis: The AMMI for e additive main effect showed F-test provided by [6] e ree multiplicative a significant variance (P<.) for e locations, interaction principal components was significant genotypes and genotype by location interaction (Table 3). (P<.) where e remaining interaction principal The environment captured e maximum sum of squares component was not significant e result was not in 5.97 % followed by e genotype by location interaction agreement wi [7] recommended an AMMI model sum of squares which (2.68 %) and e genotype sum of wi e first ree IPCAs predicates e genotype by square were e least (8.97%).The large sum of squares for enviroment interaction. adequately predict model environment showed at e environment was diverse fitness of e additive main effect and multiplicative wi large differences among environmental means and interaction (AMMI). The adequate number of caused variation in performance of e genotypes and is interaction principal component in e AMMI model could be attributed due to e unequal distribution of rain affected by type of traits measured, crop type but fall in e growing season and heterogeneity of location according to [8, 9] in multi location yield trial e pattern in soil type and altitude range in discriminating e of interaction is mainly explained by e two interaction performance of genotypes. Large environmental sum of principal component analysis and using e two squares was reported by [5, ] at who found very interaction principal component e genotypes can be large and significant environmental sum of squares. recommend. 23

4 World J. Agric. Sci., (3): 2-27, 25 Table 3: AMMI analysis of variance for grain yield of 22 bread wheat genotypes in six locations, in e production year 23/24 in Sou Eastern Eiopia Source df SS MS %SS Total Treatments ** 75.4 Genotypes ** 8.97 Environments ** 5.97 Block ** Interactions ** 2.68 IPCA ** 6.77 IPCA ** 5.7 IPCA ** 3.23 IPCA ns IPCA ns IPCA6 5 ns Error ** significant at p., ns = not significant The AMMI analysis of variance (Multiplicative genotypes G (Danda a), G6 (ETBW 6732) and effect) was furer broken by decomposing it into G2(ETBW6747) were genotypes wi lower mean grain principal components (Table3). The first principal yield. component (IPCA explained 6.77% of e genotype by The genotype G22 (Digelu), G2 (ETBW6738), G4 environment interaction and e AMMI had a model (ETBW673), G7 (ETBW6743) and G2 (ETBW6747) was fitness of 89.2 % of e treatment sum of square in e located near to e origin wi lower contribution to e genotype by environment interaction of e bread wheat magnitude genotype by environment interaction implying genotypes and 8.97% was explained due to noise. Hence at e genotypes were stable. The genotypes G e genotype by environment data of e 22 bread wheat (Danda a), G (ETBW 6736) G6 (ETBW6732) and G2 genotype were best explained by AMMI The result of (ETBW6746) ey were e most unstable genotypes e study was in agreement wi [2] at e AMM2 had (Fig ). better fitness in food barley. The testing locations E5 (Asassa), E4 (Kulumsa), E6 (Bokoji) and E(Holeta) were favorable testing location AMMI Biplot Analysis: The AMMI analysis provides located to e right side of e grand mean where as e a graphical representation (biplot) to summarize only testing location E3(Areka) was un favorable testing information on main effects and interactions effect of bo location placed to e left Side of e perpendicular line genotypes and environments simultaneously. The (grand mean). The testing location E3 (Areka) was located intimacy between pairs of locations or pairs of genotypes distant from e origin implying e testing locations had in e biplot is proportional to eir similarity for genotype higher contribution to e magnitude of genotype by by location interaction effects [2]. The interaction environment interaction and caused unstable genotype principal component (IPCA) represented in e y- axis performance. The testing locations E5 (Asassa) and E4 where as e genotype and environment mean (Kulumsa) were nearly placed to e origin wi lower represented on e x-axis (Figure ). Genotypes or contribution to genotype by environment interaction and Location located in e right side of e midpoint of e implying e testing locations had less contribution to e perpendicular line have higher yields an genotypes or genotype by location interaction and contributes to e location placed to e left side of e perpendicular line stable performance of e genotypes. The AMMI biplot (grand mean). analysis e first interaction principal component(ipca) Genotypes or Environments located in e right side had explained 6.77% of e genotype by enviroment of e midpoint of e perpendicular line have higher interaction and AMMI had a model fitness of 89.2% of yields an ose on e left side hence genotypes G22 e treatment sum square in e genotype by enviroment (Digelu), G2 (ETBW6738),G4(ETBW673), G9 interaction of bread wheat genotypes and 8.97% was (ETBW6745),G2 (ETBW6746) and G3 (ETBW6739) were explained e noise. The genotype by enviroment higher yielder genotypes wi mean grain yield of interaction of e bread wheat genotypes e AMMI (44.66,44,44.5,43.62,45.9 Qt/ha) respectively. The model gives e best model fit. The result of e study 24

5 AMMI BIPLOT OF MAIN EFFECTS AND INTERACTIONS IPCA -.6 E3 World J. Agric. Sci., (3): 2-27, 25 E E E4 E E MEANS VARIATE: YIELD DATA FILE: MM MODEL FIT: 89.2% OF TABLE S Fig. : AMMI biplot for grain yield of 22 bread wheat genotypes tested in six locations of Sou East Eiopia Genotypes plotted as G, G2 and G3 and Enviroment plotted as Hol, Si, Are, Kul, Asas and Bok N.B. abbreviations in e Biplot are mentioned as follows: Genotypes: G = Danda a, G2 =ETBW6728, G3 = ETBW6729, G4 = ETBW673, G5= ETBW673,G6= ETBW6732, G7 = ETBW6733, G8 = ETBW6734, G9= ETBW6735,G= ETBW6736,G= ETBW6737, G2= ETBW6738, G3= ETBW6739, G4= ETBW674,G5= ETBW674, G6= ETBW6742,G7= ETBW6743, G8= ETBW6744,G9= ETBW6745,G2= ETBW6746, G2= ETBW6747 and G22= Digelu Environments: E = Holeta, E2 = Sinana, E3 = Areka, E4 = Kulumsa, E5= Asassa, E6= Bokoji Table 4: AMMI stability value for yield of 22 bread wheat genotypes in six testing locations, in e production year 23/4 during main cropping season Genotype Gm IPCA IPCA2 ASV Rank Danda'a ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW ETBW Digelu

6 World J. Agric. Sci., (3): 2-27, 25 was in agreement wi [22] in food barley and similar reports was been made by [23] in Evaluation of yield and seed requirements stability of bread wheat at e AMMI2 had better fitness in bread wheat genotypes and deviation of current study wi e previous auors could explained e genotype by enviroment interaction. AMMI Stability Value (ASV): The Bread wheat genotypes showed significant genotype by testing location interaction effect and e additive and multiplicative interaction effect stability analysis (ASV) implied to decompose e interaction effect. Considering mean grain yield as a first criteria for evaluating e bread wheat genotypes ETBW6728 was wi a higher mean grain yield (46.28Qt/ha) followed by e genotypes ETBW6734 and ETBW6746 wi e mean grain yield of (46.26 and 45.9 Qt/ha) while e genotypes ETBW6729, ETBW6747 and ETBW6742 was wi low mean yields across e testing locations (Table 2). The interaction principal component one (IPCA) scores and e interaction principal component two in e AMMI model are indicators of stability [4]. Considering e first interaction principal component (IPCA) e genotypes ETBW6728, were e most stable genotype wi IPCA value (-.79) followed by ETBW6733, ETBW6737 and ETBW6734 wi IPCA value of (-.66,-.62and -.5). When e second interaction principal component (IPCA2) was considered ETBW6746 was e most stable genotype wi interaction principal component value (-.67) followed by e genotype ETBW6737 wi e IPCA2 value (-.4). The two principal components have eir own extremes, but calculating e AMMI stability value (ASV) is a balanced measure of stability [4]. The Genotypes wi lower ASV values is considered more stable and genotypes wi higher ASV are unstable. According to e ASV ranking in e (Table 2) e genotype Danda a was e most stable wi an ASV value of (-.8) followed by e genotype ETBW6744 wi ASV value (-5.5).The genotypes ETBW6735 was e most unstable wi ASV value (5.49). The Stable genotypes was followed wi mean grain yield above e grand mean and is result was in agreement wi [7] who has used ASV as one meod of evaluating grain yield stability of bread wheat varieties in Tigray and similar reports been made by [22, 24] in barley in Tigray and bread wheat Using AMMI stability value. CONCLUSION The AMMI analysis for e additive main effect and multiplicative interaction effect revealed significant variance for Genotype, location and genotype by location interaction. For e study of genotype by environment interaction of e 22 bread wheat genotypes across six locations e AMMI gives e best model fitness. In multi-location, adaption trial considering bo e stability and mean grain yield is vital. According to e ASV and AMM biplot e bread wheat genotypes ETBW6742, ETBW673 and ETBW6732 were stable genotypes coupled wi higher mean grain yield greater an e grand mean. The genotypes Danda a, ETBW6732 and ETBW6747 were unstable wi lower mean grain yield less an e grand mean. Using e AMMI biplot analysis E2 (Sinana), E5 (Asassa), E6 (Bokoji), E4 (Kulumsa) and E (Holeta) were favorable testing locations while e testing location Areka was unfavorable on some response variables. REFERENCES. Tsenov, N., D. Atanasova, J. Todorov and I. Ivanova, 2. "Quality of winter common wheat advanced lines depending on allelic variation of Glu-A3." Cereal Research Communications, 38(2): Curtis, B.C., 22. Wheat in e world. pp: -7. In: Curtis Mac pherson (eds). Bread Wheat Improvement and Production. FAO, Rome. 3. Central Statistical Auority (CSA), 29. Agricultural Sample Survey 28/29. Central Statistical Auority. Addis Ababa, Eiopia. 4. Abera, H.B., 28. Adoption of improved teff and wheat production technologies in crop-livestock mixed systems in norern and western Shewa zones of Eiopia (Doctoral dissertation, University of Pretoria). 5. Bachewe, F., 29. The state of subsistence agriculture in Eiopia: Sources of output grow and agricultural inefficiency (Doctoral dissertation, University of Minnesota). 6. FAO, 23. FAOSTAT. Food and Agricalture Organization of United nations, Rome, Italy. 7. Central Statistical Agency (CSA), 23. Agricultural sample survey 22/3. Report on crop and livestock product utilization. Addis Ababa, Eiopia. 26

7 World J. Agric. Sci., (3): 2-27, Hailu, G., 99. Wheat production and research in 6. Gollob, H.F., 968. Confounding of sources of Eiopia. In: Hailu Gebre-Mariam, D.G. Tanner and variation in factor-analytic techniques. Psychological M. Hulluka (eds). Wheat Research in Eiopia: Bulletin, 7(5): 33. A Historical Perspective. Addis Ababa: 7. Hintsa, G. and F. Abay, 23. Evaluation of Bread IAR/CIMMYT. wheat Genotypes for eir Adaptability in wheat 9. Ceccarelli, S. and S. Grando, 27. growing Areas of Tigray Region, norern Eiopia. Decentralized-participatory plant breeding: Journal of Biodiversity and Endangered Species. an example of demand driven research. Euphytica, 8. Gauch Jr, H.G., 988. Model selection and 55(3): validation for yield trials wi interaction. Biometrics,. Kaya, Y., M. Akçura and S. Taner, 26. GGE-biplot pp: analysis of multi-environment yield trials in bread 9. Yan, W., M.S. Kang, B. Ma, S. Woods and wheat. Turkish journal of agriculture and forestry, P.L. Cornelius, 27. GGE biplot vs. AMMI analysis 3(5): of genotype-by-environment data. Crop science,. Mitrovic, B., D. Stanisavljevi, S. Treski, 47(2): M. Stojakovic, M. Ivanovic, G. Bekavac and 2. Gebremedhin, W., M. Firew and B. Tesfye, 24. M. Rajkovic, 22. Evaluation of experimental Maize Stability analysis of food barley genotypes in hybrids tested in Multi-location trials using AMMI norern Eiopia. African Crop Science Journal, and GGE biplot analysis. Turk. J. Field Crops, 22(2): (): Crossa, J., 99. Statistical analysis for multi-location 2. Gauch, H.G., 26. Winning e accuracy game. trials. Adv. Agrono., 44: American Sci., 94: Abay, F. and A. Bjørnstad, 29. Specific adaptation 3. Crossa, J., P.N. Fox, W.H. Pfeiffer, S. Rajaram and of barley varieties in different locations in Eiopia. H.G. Gauch, 99. AMMI adjustment for statistical Euphytica, 67(2): analysis of an international wheat yield trial. Theor. 23. Mladenov, V., B. Banjac and M. MILOŠEVI, 22. App. Gen., 8: Evaluation of yield and seed requirements stability of 4. Purchase, J.L., 997. Parametric analysis to describe bread wheat (Triticum aestivum L.) via AMMI model. genotype x environment interaction and yield Turk. J. Field Crops, 7(2): stability in winter wheat. Ph.D. Thesis, Department of 24. Sivapalan, S., L. O?brien, G. Ortiz-Ferrara, Agronomy, Faculty of Agriculture of e University G.J. Hollamby, I. Barclay and P.J. Martin, 2. of e Free State, Bloemfontein, Sou Africa. An adaptation analysis of Australian and 5. Farshadfar, E., H. Safari and A. Yaghotipoor, 22. CIMMYT/ICARDA wheat germ plasm in Australian Chromosomal Localization of QTLs Controlling production enviroments. Crop Science pasture, Genotype X Environment Interaction in Wheat 5(7): Substitution Lines Using Nonparametric Meods. Journal of Agricultural Science, 4(2): 8. B. Feil, 992. Breeding progress in small grain cereals: a comparison of old and modern cultivars. Plant Breed., 8: -. 27