Genotype X Environment Interactions of Fodder Barley Genotypes as Estimated by AMMI, BLUP and Non Parametric Measures

Introduction

Genotype x Environment interactions of cross over type would introduce inconsistency in the behaviour  of genotypes evaluated in the various environmental conditions¹. Adaptability and stability of various crops under multi-environment field trials studied by  number of analytic measures as observed in the literature². Moreover non parametric measures to assess GxE interaction and stability analysis had been also reflected ^6. The components of analysis of variance, the regression models, non-parametric methods, AMMI methods, BLUP based mixed modelswould be most suitableanalytic methods³.AMMI stability value (ASV), ASV1, Modified AMMI stability value (MASV) & MASV1) have been registered  visibility⁴.  Random effects of the genotypes to improve their predictive accuracy had been advocated for Best linear unbiased prediction (BLUP) based measures. The stability and adaptability of genotypes were also highlighted by the harmonic mean of genotypic values (HMGV), relative performance of genotypic values (RPGV), and harmonic mean of relative performance of genotypic values (HMRPGV) ⁵. Besides that nonparametric measures S_i¹,S_i²,S_i³,S_i⁴,S_i⁵,S_i⁶,S_i⁷, NP_i⁽¹⁾, NP_i⁽²⁾, NP_i ⁽³⁾, NP_i ⁽⁴⁾ have been also utilized for genotypes x environmental conditions⁶.Allrecentanalytic measures have been compared to decipher the GxE interactions effects for fodder barley genotypes evaluated in northern hills zone of the country. 

Materials and Methods

Twenty three fodder barley genotypes were evaluated at six major centers of All India Co-ordinated Research Project at the northern hill zone of the country. To increase the barley production of this zone has been emphasised more to augment the total fodder production of the country. Randomized complete block designs with four replications has been laid out in field trials during 2020-21 cropping season. The environmental conditions of the locations and parentage details of the evaluated fodder barley genotypes reflected in table 1for ready reference.  The phenotypic value of ith genotype in jth environ­ment  de­noted by X_ijwhere i=1,2, …k, ,j =, 1,2 ,…,n while rank of genotypes as per yield values reflected by r_ij  as the rank of the ith genotype in the jth environment, and average of ranks for the ith geno­type by r_i ̅   . The corrected yield of ith genotype in jth environment reflected as (X*_ij =  X_ij– X_i ̅.+ X_.. ̅  ) as X*_ij, the corrected mean phenotypic value; .was the overall mean of ith  genotype in all environments as X_.. ̅ .

However, the composite non parametric measures were also suggested to utilize the ranks of genotypes as per yield and corrected yield in number of environments as NP_i⁽¹⁾, NP_i⁽²⁾, NP_i⁽³⁾ and NP_i⁽⁴⁾. In the formulas, r^*_ij was the rank of X^*_ij, and r_i ̅   and M_di were the mean and median ranks for original (unadjusted) grain yield, where r_i ̅  ^* and M^*_di were the same parameters computed from the corrected (adjusted) data.

AMMISOFT version 1.0 software utilized for AMMI analysis of data sets and SAS software version 9.3 for further analysis.

Table 1: The location and parentage details of fodder barley genotypes.

Results and Discussion

AMMI analysis

AMMI analysis observed highly significant variations (P>0.001) due to environments, GxE interactions, and genotypes with corresponding share of 67.5% ,14.1% , marginally 3.2% (Table 2) the total sum square of variation for yield⁷. Further Interaction effects portioned into four significant components accounted for nearly 98% of interactions sum of square variations. First component (AMMI1) contributed 53.7%, followed by 32.1% , 6.9%, 4.8% byAMMI2, AMMI3, AMMI4 respectively. Nearly85.9% of the total variation contributed by  the two AMMI components³. G×E signal and noise effects accounted for 25.7%  &74.2% in total G×E.  Share of GxEnoise effect was 3.2 times the genotypes effects.

Table 2: Interaction principal component analysis of Fodder barley genotypes.

Behaviour of genotypes as per BLUP based measures

Average higher yield showed by G6, G2, G7 genotypes while lowest yield of G23 (Table 3). IPCA’s in the AMMI analysis exploited to know about the stability or adaptability of genotypes. Absolute IPCA-1 scores pointed for G9, G4, G5 while as per IPCA-2, G2, G15, G21genotypes would be of choice. Values of IPCA-3 favouredG18,G8, G23genotypes. As per IPCA-4, G17, G7, G10 genotypes would be of stable performance.  ASV & ASV1 measures based on two IPCAs and utilized 85.9% of G×E interaction sum of squares would be useful for  dynamic concept of stability⁸.  Both measures recommended (G4, G5, G9) as of stable performance. Values of MASV1using 97.7% of GxE interactions sum of squares identified  G18, G15, G8genotypes whereas G18,  G8, G5 genotypes be of stable yield as per MASV⁹.

Table 3: Stability measures of  fodder barley genotypes  as perAMMI analysis. Click here to view Table

Performance of genotypes as per BLUP and Non parametric measures

Average yield of genotypes as per their best linear unbiased predictors⁴ pointed towards G2, G6, G16 as high yielders. Consistent yield of G11, G6, G15 recognised as per lower values of standard deviation while CV values identified G6, G5, G20 genotypes for northern hills zone of the country. More over the values of GM selected G11, G6, G5. Values of measure HM ,BLUP-based simultaneous selection, identified G11, G6, G20 while values of PRVG favored G11,G6,  G5and HMPRVG settled for G6, G11, G5 genotypes. Measures  HMGV, RPGV and HMRPGV had expressed the same ranking of genotypes as reported^2,6.

S_i¹ non parametric measure pointed forG9, G8, G1 while S_i² selected G9, G8, G1  and values of S_i³favoured G9^, G5, G1 as suitable genotypes (Table 4). G9, G8, G1selected by values of S_i⁴ , S_i⁵ favouredG9, G5, G1,S_i⁶G9, G5, G1 and lastly S_i⁷ for G3, G5, G1 (Table 4). The stability of genotype over environment in biological concept appreciated by its consistent rank over other environments¹⁰. Further composite measures NP_i⁽¹⁾  to NP_i⁽⁴⁾, considered the ranks of genotypes as per yield and corrected yield simultaneously. NP_i ⁽¹⁾ measure observed suitability of G9, G8, G1 whereas as per NP_i⁽²⁾,   genotypes G9, G4, G8  would be of choice while NP_i⁽³⁾  identifiedG9, G8, G1. Last composite measure NP_i⁽⁴⁾  found G9, G8, G4  as genotypes of choice for this zone.

Table 4: Stability measures of fodder barley genotypes as per BLUP and Non parametric  measures. Click here to view Table

Biplot analysis

Approximately 64.1% of the total variation among the AMMI, BLUP and non parametric measures explained by first two significant PC’s in biplot analysis (Table 5) with respective contributions of 35.9% &28.1% by first and second principal components respectively ^1,11. Measures S_i¹, S_i², S_i³, S_i⁴, S_i⁵,S_i⁶ ,S_i⁷,HMPRVG, ASV1, ASV, accounted more of share in first principal component whereas NP_i ⁽¹⁾, NP_i ⁽²⁾,NP_i ⁽³⁾, NP_i ⁽⁴⁾,PRVG,  S_i¹, GM, Mean, Average were major contributors  in PC2. The biplot analysis had been established to study the association among measures via graphical presentation.  Positive correlation among measures pointed out by acute angles between vectors of measures from the origin in the biplot while negative correlation expressed by obtuse or straight line angles. Moreover the right angles between vectors expressed Independent type of relationships.

Table 5: Contribution share of AMMI, BLUP and Non parametric measures.

Figure 1: Association analysis among AMMI, BLUP and Non parametric measures. Click here to view Figure

Direct association observed for IPC3, IPC1with BLUP based measures GM, HM, PRVG, HMPRVG, Average in one quadrant. CV closely related to Stdev, IPC2, IPC4 in opposite quadrant. ASV, ASV1 expressed very tight association with S_i⁶,S_i^7. Whereas NP_i⁽¹⁾, exhibited close affinity with S_i¹, S_i⁴,S_i²,S_i⁵. Closely related NP_i⁽²⁾ ,NP_i⁽³⁾ , NP_i⁽⁴⁾ were placed in same quadrant. Group CV with Stdev, IPC2, IPC4 managed right angles with group of BLUP based measures.Nonparametric measures NP_i⁽²⁾ ,NP_i⁽³⁾ , NP_i⁽⁴⁾ showed right angles with BLUP based measures^. AMMI based measures also exhibited right angles with BLUP based measures. Overall small and large sizes seven clusters observed among the measures for this study.  CV grouped with Stdev, IPC2, IPC4 in first cluster of first quadrant. Third quadrant seen two clusters first former one of IPC1 with IPC3 whereas latter one consisted of BLUP based measures. Last quadrant placed four clusters. MASV showed affinity with MASV1. Nearby cluster of  ASV,  ASV1 with S_i³, S_i⁶,S_i⁷.Adjacent cluster consisted of NP_i⁽²⁾, NP_i⁽³⁾, NP_i⁽⁴⁾ measures. WhileS_i¹, S_i⁴,S_i²,S_i⁵ managed with NP_i⁽¹⁾  in last cluster (Fig.2).

Figure 2: Clustering pattern of AMMI, BLUP and Non parametric measures. Click here to view Figure

Acknowledgement

The training by Dr J Crossa and financial support by Dr A.K Joshi & Dr RP Singh CIMMYT, Mexico sincerely acknowledged along with hard work of the staff to carry out the field evaluation of genotypes at coordinating centres.

Conflict of Interest

No conflict of interests reported by the authors.

Funding Sources

There is no funding source.

References

1. Ahakpaz F., Abdi H., Neyestani E., Hesami A., Mohammadi B., Nader Mahmoudi K., Abedi-Asl G., Jazayeri Noshabadi M.R., Ahakpaz F., Alipour H. Genotype-by-environment interaction analysis for grain yield of barley genotypes under dry land conditions and the role of monthly rainfall. Agric Water Manag., 2021; 245:10665
   CrossRef
2. Anuradha N., Patro T.S.S.K., Singamsetti A., Sandhya Rani Y., Triveni U., Nirmala Kumari A., Govanakoppa N., Lakshmi Pathy T. and Tonapi V.A. Comparative Study of AMMI- and BLUP-Based Simultaneous Selection for Grain Yield and Stability of Finger Millet [Eleusine coracana (L.) Gaertn.] Genotypes. Plant Sci., 2022;12:786839.
   CrossRef
3. Pour‑Aboughadareh A., Ali B., Ali K. S., Mehdi J., Akbar M., Ahmad G., Kamal S.H., Hassan Z., Poodineh Omid and Masoome, K.   Dissection of genotype‑by‑environment interaction and yield stability analysis in barley using AMMI model and stability statistics. Bulletin of the National Research Centre, 2022;46:19
   CrossRef
4. Sousa A.M.C.B., Silva V.B., Lopes A.C.A., Ferreira-Gomes R.L. and Carvalho L.C.B . Prediction of grain yield, adaptability, and stability in landrace varieties of lima bean (Phaseolus lunatus L.).Cr Br and App Bio., 2020; 20: e295120115
   CrossRef
5. Gonçalves G. de M. C. , Gomes R. L. F., Lopes Â. C. de A. and Vieira P. Fe. de M. J. Adaptability and yield stability of soybean genotypes by REML/BLUP and GGE Biplot. Cr Br and App Bio.,2020;20(2): e282920217.
   CrossRef
6. Pour-Aboughadareh A., Yousefian M., Moradkhani H., Poczai P., and Siddique K.H.  STABILITYSOFT: A  new online program to calculate parametric and non-  parametric  stability statistics for crop traits.  App in Pl Sci., 2019;7(1): e1211
   CrossRef
7. Mehraban R. A., Hossein-Pour T., Koohkan E., Ghasemi S., Moradkhani H., Siddique K.H . Integrating different stability models to investigate genotype × environment interactions and identify stable and high-yielding barley genotypes. Euphytica, 2019;215:63
   CrossRef
8. Silva E. M. da, Nunes E. W. L. P., Costa J. M. da, Ricarte A. de O., Nunes G. H. de S. and Aragão Fernando Antonio Souza de . Genotype x environment interaction, adaptability and stability of ‘Piel de Sapo’ melon hybrids through mixed models.Cr Br and App Bio.,2019; 19(4): 402-411.
   CrossRef
9. Gerrano A.S., Rensburg W.S.J.V., Mathew I., Shayanowako A.I.T., Bairu M.W., Venter S.L., Swart W., Mofokeng A., Mellem J., Labuschagne M. Genotype and genotype x environment interaction effects on the grain yield performance of cowpea genotypes in dry land farming system in South Africa. Euphytica, 2020;216:80
   CrossRef
10. Vaezi B., Pour-AboughadarehA. MehrabanA., Hossein-PourT., Mohammadi R. Armion  and Dorri M.  The use of parametric and  non-  parametric measures for selecting stable and adapted barley lines.  Arc of Ag and So Sci.,2018; 64: 597–611
    CrossRef
11. Bocianowski J., Tratwal A., Nowosad K . Genotype by environment interaction for main winter triticale varieties characteristics at two levels of technology using additive main effects and multiplicative interaction model. Euphytica, 2021;217:26
    CrossRef