Genotype-by-environment interaction for grain yield among novel cowpea (Vigna unguiculata L.)selections derived by gamma irradiation

Lydi Horn,Hussein Shimelis,Ftm Srsu,Lernmore Mwdzingeni,*,Mrk D.Ling

a

aAfrican Centre for Crop Improvement,University of KwaZulu-Natal,P/Bag X01,Scottsville 3209,Pietermaritzburg,South Africa

bDirectorate of Research and Training,Plant Production Research,Ministry of Agriculture,Water and Forestry,Private Bag 13184,Windhoek,Namibia

cPlant Breeding and Genetics Section,Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture,Vienna,Austria

1.Introduction

Cowpea(Vigna unguiculata L.;2n=2×=22)is an important legume crop widely grown in South America,sub-Saharan Africa,and Asia.It withstands harsh growing conditions,particularly drought stress[1,2].The grain,young pods,and succulent leaves are used for human food,while the foliage is an important livestock feed[3].Further,cowpea restores soil fertility through nitrogen fixation,making it an ideal component crop in rotation systems.Production and productivity of cowpea has been low in Namibia,owing to unavailability of seeds of improved cultivars,harsh climatic conditions,diseases,insect pests,and parasitic weeds[2].The present yield of cowpea ranges from 0.10 to 0.60 t ha?1in Namibia,a range far below the potentially attainable yields of 1.5 to 3.0 t ha?1reported elsewhere[4].There is thus a need to breed improved cowpea cultivars with enhanced grain yieldand quality to meet local and regional production and productivity and market demands.

Design,development,and deployment of improved cultivars require adequate genetic variation,achieved through crosses involving selected parents or targeted mutagenesis[5,6].The potential of chemical and physical mutagenic agents to induce genetic variation of cowpea for grain yield and yield-related traits has been well reported[7,8].Induced mutagenesis has been successfully used to modify several agronomic traits of cowpea,such as plant height,maturity,seed shattering resistance,disease resistance,seed color,seed size,and yield[7–9].

Genotype-by-environment(G×E)interaction analysis is an important prerequisite for recommendation of novel selections for large-scale production.It enables assessment of the relative performance and stability of genotypes for yield and yield-related traits[1,10,11].The performance of tested genotypes is influenced by the genotype,the environment,and G×E interaction[1].The growing environment often masks the potential genetic expression,leading to poor genetic gain from artificial selection,especially for quantitative traits such as grain yield.G×E analysis involves evaluation of novel selections across representative growing environments,which will assist breeders to recommend promising genotypes based on their narrow or broad adaptation.G×E analyses are valuable during the final stages of selection of elite breeding materials.Several statistical techniques have been widely adapted to analyze and interpret G×E data,including the additive main effect and multiplicative interaction(AMMI)and the genotype main effect plus genotype-by-environment interaction(GGE)biplot analysis[12,13].

A joint cowpea mutation breeding project was initiated between the government of Namibia and the International Atomic Energy Agency(IAEA)under a Technical Cooperation project to develop improved cultivars with better adaptation[2].This project resulted in the selection of promising mutants with high yield potential,drought tolerance,and insect pest resistance through continuous selfing and selection from the M2 to M7 generations[14].The selected M6 and M7 elite mutants needed to be evaluated across representative growing environments to determine their performance and yield stability for effective cultivar recommendation and to identify suitable production environments.Accordingly,the objectives of this study were to evaluate the effects of G×E interaction and yield stability among elite cowpea selections derived by gamma irradiation and to identify promising genotypes with narrow or broader adaptation for production or future breeding programs in Namibia or similar environments.

2.Material and methods

2.1.Study sites and plant material

The study was conducted at three sites:Bagani(?18°09′61.93″S,21°56′24.14″E),Mannheim(19°12′21.4″S,17°42′29.1″E),and Omahenene(?17°44′29.04″S,14°78′48.21″E)during the 2014/2015 and 2015/2016 cropping seasons.This plan provided six testing environments including Bagani 2014/2015,Bagani 2015/2016,Mannheim 2014/2015,Mannheim 2015/2016,Omahenene2014/2015,and Omahenene2015/2016.The physicochemical properties of soils at Bagani,Mannheim and Omahenene research sites are described by Horn et al.[14].Mean monthly and total rainfall(mm)at the three sites during 2014/2015 and 2015/2016 are presented in Table 1.The study used37cowpea genotypes comprising34newly developed mutant lines,selected for their superior agronomic performance,and three parental checks(Bira,Nakare and Shindimba).The mutants were at the M6 generation in 2014/2015 and M7 in 2015/2016.Details of the genotypes are presented in Table 2.

2.2.Experimental design and data collection

The experiments were performed using a randomized complete block design with three replications.Experimental units consisted of 8 rows of 4 m length with spacings of 20 cm within and 75 cm between rows.The crops were established under rainfed conditions with supplementary irrigation when required.Two middle rows(net plots)were harvested to estimate grain yield per plot,later converted to yield per hectare(t ha?1).The outer rows were not used for yield estimation in order to control border effects and to minimize experimental error.

2.3.Data analysis

Grain yield data was subjected to a combined analysis of variance(ANOVA)using GenStat 18 statistical software[15].The following AMMI model according to Gauch(16)was used for G×E and yield stability analyses based on the principal component analysis(PCA):

where Ygeis the yield of genotype g in environment e,μ is the grand mean, αgis the genotype mean deviation, βeis the environment mean deviation,λnis the eigenvalue of the nthprincipal component(PCA)axis,Υgnand ηenare the genotype and environmental PCA scores for the nthPCA axis,and θge,is the residual.The AMMI stability value(ASV)was calculated according to Purchase,Hatting and Van Deventer[17]as follows:

where SS is the sum of squares of the IPCAs and IPCA1 and IPCA2 are the first and second interaction principal component axes,respectively.Means of the genotypes were used for GGE biplot analysis.

3.Results

3.1.AMMI analysis

Mean yield for the studied traits varied widely,from 0.74 to 2.83 t ha?1.Table 3 shows the mean grain yields(t ha?1)of the34 cowpea mutant genotypes and their three parental lines in six environments in northern Namibia.AMMI analysis of variance revealed highly significant main effects(P＜0.001)of genotypes,environments and their interactions(Table 4).Genotype,G×E interaction,and the AMMI model explained respectively 37.95%,33.83%,and 77.49%of the total observed variation.In contrast,interaction principal component axes IPCA1 and IPCA2 explained respectively 44.63%and 23.41%of the total variation.Genotype G9 was ranked first across all the test environments.Mutant lines G19 and G22,developed from the parent Nakare irradiated at 150 Gy,were among the high and stable yielders.Based on the AMMI biplot(Fig.1),acute angles were observed between vectors of genotypes G4,G5,and G15 and those of environments E1,E3,and E5.The acute angle between the lines that connect the biplot origin and environments E1 and E3,as well as E2,E4,and E6 showed their close relationships.Genotype G20 was the most stable,with an ASV of 0.08(Table 5).

Table 1–Mean monthly and total rainfall(mm)during the study period in 2014/2015 and 2015/2016 at three field sites.

See codes of genotypes(G1 to G37)in Table 3.Min,minimum;Max,maximum,CV,coefficient of variance.

3.2.GGE biplot analysis

A “which won where”polygon view of the relationship between genotypes and environments is presented in Fig.2.The biplot explained 75.57%of the total variation observed,of which 63.57%was explained by the first principal component(PC1),while the second principal component(PC2)explained 12%.Genotypes G3,G6,G9,G24,and G29 were situated at the corners of the “which won where”polygon indicating that they were outstanding genotypes in particular environments[13].Among these,G9 was the highest-yielding genotype in all the test environments.Other genotypes including G1,G2,G13,G17,and G20 were located close to the origin or center of the GGE biplot,indicating that they showed stable performance across the test sites[13].In contrast,all six test environments were grouped into one mega-environment,in which the genotypes G9,G10,G12,and G13 were associated.The best-performing mutant line was G9,followed by G10 and G12 with above-average yield in environments E6 and E3(Fig.3).Fig.4 presents the average-environment coordination(AEC)view comparing environments relative to an ideal environment.It indicates that environments E1 and E3 were located in the direction of the ideal environment.Large IPC1 scores of 0.8 and 1.0 were obtained from E1 and E5,respectively,while E2 and E4 displayed a low IPC1 score of 0.25.G9 fell closer to the centre of the concentric circle of the AEC view,next to E3.Other desirable genotypes were G4,G10,G12,and G14,located on the third and fourth concentric circles.

Table 2–List of 34 cowpea mutant genotypes and three parental lines evaluated at three sites(Bagani,Mannheim,and Omahenene)during the 2014/2015 and 2015/2016 cropping seasons at the M6 and M7 generations,respectively.

Table 3 –Mean grain yield(t ha?1)of 34 cowpea mutant genotypes and their three parental lines tested in six environments in northern Namibia.

4.Discussion

Significant G×E effects observed in the present study indicate that the genotypes evaluated do not show consistent performance across test environments.This allows for an investigation of the nature and magnitude of G×E,which cannot be achieved by a standard analysis of variance[16,17].Genotype G9,which was ranked as the highest yielder across all environments,could be the best candidate for production across sites.The AMMI biplot reveals the relationship between genotypes and environments,while AMMI stability values provide more information on the variation among genotypes.Stable genotypes have ASV values close to zero[18].Thus,G20,with an ASV of 0.08,could harbor genes for adaptability to various agroclimatic conditions.This mutant line can be used during breeding for yield stability.Similarly,IPCA scores are an indication of genotype stability.The greater the IPCA scores,either negative or positive,the more specifically adapted is a genotype to particular environments.The closer the IPCA scores approach to zero,the more stable or adapted is the genotype across all the test environments,as observed for line G20.

GGE biplot analysis provides a graphical representation of the relationships between genotypes and environments and can effectively reveal genotype performance and stability[13].The vertex mutant lines G3,G6,G9,G24,and G29 were among the environmentally most responsive genotypes and can be recommended for specific adaptation.In contrast,G1,G2,G13,G17,and G20,located close to the origin,were among the environmentally least responsive lines and can be used in breeding for wide adaptation.The presence of only onemega-environment in the present study suggests that the six sites did not differ significantly in terms of discriminating capacity,so that deploying genotypes in any one of those environments would give similar results[13].This finding implies that future evaluation of the same set of materials could be performed in the most representative of the environments in order to save costs.In this case,the ideal test environment is the one with the largest PC1 scores and should have more power to discriminate genotype main effects[19,20].Thus,E3 and E1,located closest to the ideal environment with a large PC1 score could be the best sites for germplasm evaluation.Despite this observation,genotypes G9,G10,and G12 could be targeted specifically for production in environments E6 and E3,where they performed above average.

Table 4–AMMI analysis of variance for seed yield of 34 cowpea mutant genotypes and their three parental lines tested in six environments in northern Namibia.

An ideal genotype is the one that shows the highest mean performance and is highly stable across all test environments[13,19].Based on the average-environment coordination(AEC)view comparison biplot,an ideal genotype is associated with greatest vector length of the high-yielding genotypes,and a desirable genotype is the one that is located closer to an ideal genotype,which is usually at the center of the concentric circles.Mutant line G9 appears to be adapted specifically to E3.This genotype fell at the corners or vertices of the polygon view close to E3(Fig.2),performing above average and close to E3(Fig.3)and positioned close to the ideal environment(Fig.4).This genotype showed the highest yield in all the test environments.Thus,it may be recommended for production over all the present study sites.Genotypes that can be selected for cultivation across the studied environments or for future breeding include G4,G10,G12,and G14 located on the third and fourth concentric circles close to the average environment.Genotype G14(Shindimba)is one of the check varieties,known for high yield and large white grains,but is disfavored by farmers because of its coiled pod shape.The newly developed mutant derivatives of Shindimba,namely G3,G4,G9,G10,and G12 had straight pods,indicating that in addition to grain yield,mutagenesis also created variation for other key traits.

5.Conclusions

Table 5–AMMI adjusted combined mean grain yield(t ha?1),IPCA scores of 33 cowpea mutant genotypes and their three parental lines tested in six environments in northern Namibia.

This study selected promising cowpea mutant genotypes using G×E analyses involving different agroecological conditions.Four mutant selections:G9(ShL3P74),G10(ShR3P4),G12(ShR9P5),and G4(ShL2P4),showed the high grain yields,2.83,2.06,1.99,and 1.95 t ha?1,respectively.Elite mutant selections derived from the parental line Shindimba:G4,G9,G10,and G12,were among the highest grain yielders with the straight pod shape desired by cowpea farmers in northern Namibia.Accordingly,the above novel selections can be recommended for direct production or future cowpea breeding programs in Namibia or similar environments.

Acknowledgments

Fig.2––The“which won where”view of the GGE biplot showing which genotypes performed best in which environment.E1,Bagani 2014/2015;E2,Bagani 2015/2016;E3,Mannheim 2014/2015;E4,Mannheim 2015/2016;E5,Omahenene 2014/2015;E6,Omahenene 2015/2016.Dotted vertical and horizontal lines indicate points where the PC1 and PC2 axes had respective values of zero.Vertices of the polygon indicate superior genotypes in each sector.See codes of genotypes(G1 to G37)in Table 3.

Fig.3–Average-environment coordination(AEC)view ranking test environments in terms of the relative performance of genotypes.E1,Bagani 2014/2015;E2,Bagani 2015/2016;E3,Mannheim 2014/2015;E4,Mannheim 2015/2016;E5,Omahenene 2014/2015;E6,Omahenene 2015/2016.Dotted vertical and horizontal lines indicate points where the PC1 and PC2 axes had respective values of zero.Vertices of the polygon indicate superior genotypes in each sector and green dotted lines help to visualize the distance of genotypes and environments from the biplot origin.See codes of genotypes(G1 to G37)in Table 3.

Fig.4–The average-environment coordination(AEC)view comparison biplot comparing environments relative to an ideal environment(the center of the concentric circles).E1,Bagani 2014/2015;E2,Bagani 2015/2016;E3,Mannheim 2014/2015;E4,Mannheim 2015/2016;E5,Omahenene 2014/2015;E6,Omahenene 2015/2016.Dotted vertical and horizontal lines indicate points where the PC1 and PC2 axes had respective values of zero.The small circle on the arrowed line shows the average environment,the arrow indicates the ideal environment,and concentric circles indicate the distances of genotypes and environments from the ideal environment.See codes of genotypes(G1 to G37)in Table 3.

This work was supported by funds from the International Atomic Energy Agency(IAEA)through theTC Project(NAM5012):Developing High Yielding and Drought Tolerant Crops through Mutation Breeding)and the Ministry of Agriculture,Water and Forestry of Namibia.The University of KwaZulu-Natal and the Ministry of Agriculture,Water and Forestry(MAWF)of the government of Namibia are thanked for overall research support to the first author.Loide Aron,Rose-Marry Hukununa,Kangumba Annethe and Nghishekwa Alfeus are thanked for technical support and data collection.

[1]B.Adewale,C.Okonji,A.Oyekanmi,D.Akintobi,C.Aremu,Genotypic variability and stability of some grain yield components of cowpea,Afr.J.Agric.Res.5(2010)874–880.

[2]L.Horn,H.Shimelis,M.Laing,Participatory appraisal of production constraints,preferred traits and farming system of cowpea in the northern Namibia:implications for breeding,Legum.Res.38(2015)691–700.

[3]O.Agbogidi,Screening six cultivars of cowpea(Vignia unguiculata L.)walp for adaptation to soil contaminated with spent engine oil,J.Environ.Chem.Ecotoxicol.2(2010)103–109.

[4]O.A.Gbaye,G.J.Holloway,Varietal effects of cowpea,Vigna unguiculata,on tolerance to malathion in Callosobruchus maculatus(Coleoptera:Bruchidae),J.Stored Prod.Res.47(2011)365–371.

[5]J.A.de Ronde,M.Spreeth,N.Mayaba,W.J.van Rensburg,N.Matole,in:Q.Y.Shu(Ed.),Evaluation and characterization of mutant cowpea plants for enhanced abiotic stress tolerance,Induced Plant Mutations in the Genomics Era.Proceedings of an International Joint FAO/IAEA Symposium,Vienna,Austria,2008Joint FAO/IAEA Division of Nuclear Techniques in Food and Agriculture,International Atomic Energy Agency,Vienna,Austria,2009.

[6]C.Mba,Induced mutations unleash the potentials of plant genetic resources for food and agriculture,Agron.J.3(2013)200–231.

[7]S.Goyal,S.Khan,Induced mutagenesis in urdbean(Vigna mungo L.Hepper):a review,Int.J.Bot.6(2010)194–206.

[8]D.Singh,S.Sharma,M.Lal,B.Ranwah,V.Sharma,Induction of genetic variability for polygenic traits through physical and chemical mutagens in cowpea[Vigna unguiculata(L.)walp],Legum.Res.36(2013)10–14.

[9]R.M.Gaafar,M.Hamouda,A.Badr,Seed coat color,weight and eye pattern inheritance in gamma-rays induced cowpea M2-mutant line,J.Genet.Eng.Biotechnol.14(2016)61–68.

[10]Y.Ali,Z.Aslam,F.Hussain,A.Shakur,Genotype and environmental interaction in cowpea(Vigna unguiculata L.)for yield and disease resistance,Int.J.Environ.Sci.Technol.1(2004)119–123.

[11]O.Ariyo,Assessment of selection techniques in genotype×interaction in cowpea Vigna unguiculata(L.)walp,Afr.J.Agric.Res.2(2007)352–355.

[12]W.Yan,M.S.Kang,B.Ma,S.Woods,P.L.Cornelius,GGE biplot vs.AMMI analysis of genotype-by-environment data,Crop Sci.47(2007)643–653.

[13]W.Yan,N.A.Tinker,Biplot analysis of multi-environment trial data:principles and applications,Can.J.Plant Sci.86(2006)623–645.

[14]L.N.Horn,H.M.Ghebrehiwot,H.A.Shimelis,Selection of novel cowpea genotypes derived through gamma irradiation,Front.Plant Sci.7(2016)1–13.

[15]R.Payne,A Guide to ANOVA and Design in GenStat,VSN International,Hemel Hempstead,Hertfordshire,UK,2014.

[16]H.G.Gauch,A simple protocol for AMMI analysis of yield trials,Crop Sci.53(2013)1860–1869.

[17]J.L.Purchase,H.Hatting,C.S.Van Deventer,Genotype×environment interaction of winter wheat(Triticum aestivum L.)in South Africa:II.Stability analysis of yield performance,S.Afr.J.Plant Soil 17(2000)95–100.

[18]N.Mahmodi,A.Yaghotipoor,E.Farshadfar,AMMI stability value and simultaneous estimation of yield and yield stability in bread wheat(Triticum aestivum L.),Aust.J.Crop.Sci.5(2011)1837–1844.

[19]Y.Kaya,M.Ak?ura,S.Taner,GGE-biplot analysis of multienvironment yield trials in bread wheat,Turk.J.Agric.For.30(2006)325–337.

[20]M.Ding,B.Tier,W.Yan,H.X.Wu,M.B.Powell,T.A.McRae,Application of GGE biplot analysis to evaluate genotype(G),environment(E),and G×E interaction on Pinus radiata:a case study,N.Z.J.For.Sci.38(2008)132–142.