Research Article

Stability Analysis of Rice Hybrids for Grain Yield in Telangana through AMMI and GGE Bi-plot Model

Y. Chandra Mohan, L. Krishna, S. Sreedhar, B. Satish Chandra, Ch. Damodhar Raju, P. Madhukar, Ramya Rathod6, T. Virender Jeet Singh1 and M. V. Ramana

  • Page No:  687 - 695
  • Published online: 31 Dec 2021
  • DOI : HTTPS://DOI.ORG/10.23910/1.2021.2575

  • Abstract

An investigation was carried out on fifteen rice genotypes to identify stable rice hybrids across six different agroclimatic zones in Telangana state using AMMI and GGE bi-plot analyses during July to November, 2020. Analysis of variance clearly showed that environments contributed highest (65.47%) in total sum of squares followed by genotypes×environments (21.19%) indicating very greater role played by environments and their interactions in realizing final grain yield. AMMI analysis revealed that rice hybrids viz., RNRH 39 (G6), 27P31 (G14) and RNRH 15 (G1) were recorded higher mean grain yield with positive IPCA1 scores. The hybrids, JGLH 275 (G11) and JGLH 365 (G15) were plotted near to zero IPCA1 axis indicating that these hybrids are relatively more stable across locations. GGE bi-plot genotype view depicts that the hybrids, JGLH 365 (G15) and US 314 (G8) were inside the first concentric circle and found to be more stable across environments. GGE bi-plot environment view showed that Rudrur (E4) location was the most ideal environment. However, Warangal (E6) and Jagtial (E1) locations were poor and most discriminating. Depending on dispersion of environments in different directions, six locations were partitioned into three mega zones as first zone comprised of four locations viz., Kunaram (E2), Kampasagar (E3), Rudrur (E4) and Rajendranagar (E5) whereas highly dispersed Jagtial (E1) and Warangal (E6) were identified as two separate mega environments. The bi-plot view identified that 27P31 (G14), JGL 24423 (G2) and RNRH 39 (G6) were the best performing genotypes in first zone comprising four locations.

Keywords :   Rice hybrids, GE interaction, GGE biplot, AMMI, stability

  • Introduction

    Rice (Oryza sativa L.) is the staple food for a large proportion of the world’s population.Globally rice is cultivated in an area of 162.06 million ha, with 755.47 million t paddy production and 4.66 t ha-1 productivity (Anonymous, 2019a). In India, rice has occupied an area of 43.79 million ha with 116.42 million t production and 2659 kg ha-1 productivity (Anonymous, 2019b). Whereas in Telangana state, rice is grown in an area of 2.01 million ha with a production of 7.43 million t and an average productivity of 3694 kg ha-1 (Indiastat, 2020). 

    Human selection and adaptation to diverse environments has resulted in numerous cultivars. Although more than 900 rice varieties have been released in India, many of them were no longer cultivated within a few years due to inconsistent performance in diverse environments and only a few varieties with stable performance continue under cultivation after 15 to 20 years of their release (Nitiprasad et al., 2015). Nevertheless, there is still a large gap between production and demand. To meet this challenge, there is a need to develop rice varieties with higher stability. In order to meet the domestic demand of the increasing population, the expected requirement of rice in India would be 130 and 168 million t with reduced cultivated area of only 42 and 40 million ha by 2030 and 2050, respectively (Gupta et al., 2020). Since the yield of high yielding varieties (HYVs) of rice is plateauing, it is rather difficult to achieve this target with the present day inbred varieties. Among the limited options available, hybrid rice production is the only practically proven, sustainable and eco-friendly technology currently available for stepping up rice production significantly. Hybrid rice has clearly shown a yield advantage of 1-1.5 t ha-1  (20 to 30%) over conventionally bred modern varieties. Therefore, the introduction of hybrids and popularization of their production technology are feasible and readily adoptable to achieve targeted production.

    So far in our country, an excellent progress has been made in hybrid rice research and development. As a result of concerted efforts over the last three decades, 127 hybrids have been released for commercial cultivation in different rice growing states across the country. Among these, 38 hybrids have been developed by the public sector, while remaining 89 are developed by the private sector (Anonymous, 2021). During the year, 2020, hybrid rice was planted in an area of 3.5 million ha and more than 80% of the total hybrid rice area is in the states of Uttar Pradesh, Jharkhand, Chhattisgarh, Madhya Pradesh, Odisha and Haryana. It’s a matter of concern to note that area under hybrid rice remains @ 3 million ha since 2016 and major reason would be inadequate yield heterosis and unstable performance across environments.

    Telangana state is being identified as seed hub of the country and producing huge quantities of quality seed in different crops. Especially in hybrid rice, about 80 percent of the total hybrid rice seed production of India is taken up in the three districts of Telangana viz., Warangal, Medak, and Karimnagar (Nirmala, 2015). Though the Telangana state is producing large quantities of hybrid seed, very limited area is being grown with hybrids in rice due to their poor stability and low heterosis over available inbred varieties.

    Yield is a complex character that depends on a number of other characters and is highly influenced by genetic and environmental factors. Identification of superior genotypes through GEI became complicated for a range of environments to determine their true genetic potential. Various statistical methods/ models (parametric and non-parametric), concepts, and definitions of stability have been described over the years by many researchers (Becker and Leon, 1988; Crossa et al., 1990).

    The additive main effects and multiplicative interaction (AMMI) model proposed by Gauch (1992) is a multivariate method widely used for GEI investigation. This method uses analysis of variance (ANOVA) and principal component analysis (PCA) to compute additive main effects and interaction effects, respectively. It refers to double-centered PCA. The effectiveness of AMMI procedure has been clearly demonstrated by various authors using multi-location data (Yan and Hunt, 2001). The GGE biplot model uses one-way ANOVA for estimating genotypic effects and G×E interactions. In AMMI, G×E effects is plotted in a biplot, whereas, in GGE, the genotype and G×E effects are presented using environment-centered PCA.

    The AMMI and GGE biplot analyses have been extensively used to identify promising stable cultivars with high yield in a wide variety of crops in diverse production environments. In the present investigation, an attempt was made to identify stable hybrids suitable for Telangana state through AMMI and GGE biplot models.

  • Materials and Methods

    The present investigation was carried out on fifteen genotypes of rice during crop season of July to November 2020. The list of hybrids and checks used in the present investigation is provided in Table 1 along with their parentage and source of the material. The study was conducted at six different locations of Telangana covering different agroclimatic zones as detailed in Table 2. Crop was raised by sowing the nursery during first week of July and 25–30 days age seedlings were planted in mail field under irrigated ecosystem at all the six locations. The spacing adopted was 15×15 cm2 between hills and rows with plot size of 12 m2 replicated thrice in Randomized Complete Block Design.

    Crop was managed by adopting recommended agronomic package and suitable plant protection measures to enable for realizing potential yields in all genotypes. Grain yield was recorded in each plot and expressed as kg ha-1.

    Statistical analyses for grain yield were subjected to combined ANOVA and AMMI analysis to understand the pattern of genotype performance across the six locations. ANOVA was used to partition genotype deviations, environment deviations, and G×E deviations from the grand mean. Subsequently, multiplication effect analysis (AMMI) was used to partition GE deviations into different interaction principal component axes (IPCA). The AMMI stability value (ASV) was calculated as described by Purchase et al. (2000). PB tool software (IRRI, Philippines) was used to analyze the AMMI and GGE biplot for fifteen genotypes.

    The AMMI model used for the stability analysis is as follows:

    Yij = μ + gi + ej +∑ λkaikγjk + εij

    where Yij = mean of a trait of ith genotype in jth environment;

    µ = the grand mean;

    gi = genotypic effect;

    ej = environmental effect;

    λk = eigenvalue of Interaction Principal Components Axes (IPCA) k;

    aik = eigenvector of genotype i for PC k;

    γjk= eigenvector for environment j for PC k;

    εij = error associated with genotype i in environment j.

    The GGE biplots were generated to ascertain the which-won-where pattern and to identify the genotypes best suited across environments as well as for specific environments.

  • Results and Discussion

    An effort was made to identify stable rice hybrids across different agroclimatic zones in Telangana state using AMMI and GGE bi-plot analyses. The mean grain yield ranged from 1997 kg ha-1 to 8680 kg ha-1 (Table 3).

    Mean grain yield across locations reveled that the JGLH 337 (6275 kg ha-1) was the top ranked hybrid followed by RNRH 39 (6053 kg ha-1), JGLH 275 (5945 kg ha-1) and 27P31 (6053 kg ha-1). However, among locations, Rudrur (E4) was found to be best location with highest mean grain yield (7509 kg ha-1) across genotypes followed by Rajendranagar (E5) (7018 kg ha-1), whereas Wanargal (E6) was poorest location with least mean grain yield (3959 kg ha-1). Jat et al., (2020) reported that Arize 6444 hybrid recoded higher grain yield compared to PHB 71. According to Islam et al. (2014) the average productivity of all the genotypes across environments was 6.84 t ha-1. On the basis of environmental index value considering negative and positive, E4 and E3 were poor, E5 was medium and E2 and E1 were rich environments.

    3.1.  Analysis of variance

    Analysis of variance clearly showed that grain yield was significantly different among genotypes, environments and genotypes×environment interactions depicting the presence of significant variability among genotypes, considerable influence of environments and interaction of genotypes with environments in expression of the trait (Table 4).

    Further, environments contributed highest (65.47%) in total sum of squares followed by genotypes×environments (21.19%) indicating very greater role played by environments and their interactions in realizing final grain yield. Zewdu et al. 2020 observed highly significant difference for grain yield in upland rice by genotype (6.97%), environment (61.64%), and their interaction (20.86%). Further Fentie et al. (2013) and Islam et al. (2014) also reported the similar results in rice production.

    3.2.  AMMI analysis

    The significant G×E interactions were further partitioned by PCA (Gollob, 1968) into five significant principal component axes explaining 39.6, 31.7, 14.8, 8.0 and 5.9% of GEI sum of squares, respectively (Table 5).

    The first two interaction PCAs accounted for maximum of 71.3%. The IPCA score of a genotype in the AMMI analysis is an indication of the adaptability over environments and association between genotypes and environments (Gauch and Zobal, 1996 and Mahalingam et al., 2006). Farshadfar and Sutka (2003), Asenjo et al. (2003), Das et al. (2010) and Umma et al. (2013) analyzed G×E interaction in rice by AMMI model and found significant G×E interaction for grain yield stating the usefulness of AMMI analysis for selection of genotypes for specific location/environment.

    The mean grain yield and IPCA1 (interaction effects) were plotted on x and y axis, respectively for the construction of AMMI1 bi-plot (Figure 1). The four quadrants (Q) of the bi-plot corresponded to higher mean (Q I, II), lower mean (QIII, IV), +ve IPCA1 score (QI, IV) and -ve IPCA1 score (QII, III) and a genotype falling in same quadrant denote positive interaction and vice-versa. A genotype with IPCA1 score near to zero is considered to be more stable across environments. Conversely, a genotype with high IPCA1 score is highly variable among environments (Rao et al., 2020).

    Rice hybrids, RNRH 39 (G6), 27P31 (G14), RNRH 15 (G1) and varietal check, JGL 24423 (G2) were recorded higher mean grain yield with positive IPCA1 scores (Figure 1). The hybrid, JGLH 337 (G3) had high mean grain yield and found more adaptable to Rudrur (E4) location. However, JGLH 275 (G11) and JGLH 365 (G15) were plotted near to zero IPCA1 axis indicating that these hybrids are relatively more stable across locations. The remaining hybrids had less than the mean grain yield and found specific adaptation to few tested environments. Likewise Mary et al. (2019) reported that genotypes with PC1 scores close to zero are usually widely adapted and they were considered more stable in their performance across test environments.

    3.3.  GGE bi-plot model analysis

    GGE bi-plots provide effective evaluation of genotypes and allow for comprehensive understanding of the target and test environments through various IPCAs. The genotype×environment interactions were partitioned into six significant interaction PCAs and 61.8% variance was explained by first two IPCAs together (Table 6).

    Similarly Zewdu et al. (2020) partitioned the genotype×environment interactions into six rays which divided the biplot into seven sections.

    GGE bi-plot genotype view depicts that the hybrid JGLH 365 (G15) and US 314 (G8) were inside the first concentric circle and found to be more stable across environments (Figure 2). However, the hybrid RNRH 98 (G7) and JGLH 373 (G12) were known as highly unstable across locations with longest vector from origin.

    GGE bi-plot environment view showed that Rudrur (E4) location was the most ideal environment (Figure 3) with least vector length from origin. Conversely Warangal (E6) and Jagtial (E1) locations had longest vectors and indicating that they were poor and most discriminating. Correspondingly, Zewdu et al., 2020 reported that E6, E1, E3, and E2 environments were ideal with short vectors, while E4 and E5 had long spokes and indicated as high discriminating ability of these environments. Similarly, Kripa et al. (2020) reported that the most powerful interpretive tool for AMMI models is Bi-plot analysis and identified that environment E6 and E5 had short vectors and they did not exert strong interactive forces while E1, E2, E3 and E4 with long vectors were more differentiating environments.

    Depending on dispersion of environments in different directions, six locations were partitioned into three mega zones. First zone comprised of four locations viz., Kunaram (E2), Kampasagar (E3), Rudrur (E4) and Rajendranagar (E5) with less than 60° angle among the respective vectors, whereas highly dispersed Jagtial (E1) and Warangal (E6) were identified as two separate mega environments. Keeping in view of similarity among the locations and cost reduction of genotype testing, one environment either E2 (Kunaram) or E4 (Rudrur) pertaining to Northern Telangana Zone and likewise E3 (Kampasagar) or E5 (Rajendranagar) from Southern Telangana Zone might be ignored as a testing location to ascertain hybrid performance. In parallel, Krishnamurthy et al. (2017), the biplot showed four sectors containing all the test environments in 2011 and accordingly four mega-environments were identified, whereas three mega-environments were identified in 2012 and 2013.

    3.4. Mean performance and stability of genotypes

    The magnitude of interaction can be visualized for each genotype and each environment using IPCA vs. mean yield and IPCA1 vs. IPCA2 biplot model (Yan et al., 1998). The hybrid RNRH 39 (G6) was identified as ideal genotype followed by JGLH 275 (G11) with higher mean yield and good stability. Though the hybrids, JGLH 337 (G3) and 27P31 (G14) had high mean grain yield, found to be relatively not stable across locations falling out of the concentric circle (Table 7 and Figure 4). Considering AMMI stability value and respective rank, US 314 (G8), RNRH 18 (G10) and RNRH 78 (G9) were found to be highly stable across environments whereas RNRH 98 (G7) and JGLH 375 (G12) as most unstable. Similarly, among locations, Rudrur (E4) was identified as best location for realizing higher grain yields.

    Further, the hybrid JGLH 365 (G15) was identified as highly stable with least dispersion from AEA axis and also recorded reasonably good mean grain yield. However, the hybrid, JGLH 373 (G12) and RNRH 98 (G7) were found to be most unstable across locations. Similarly, Mary et al., 2019 found G10 and G9 as ideal genotypes having the highest mean yield, followed by G2 and were also considered the most stable genotype due to its close proximity to AEA, whereas genotype G3 was considered unstable and the poorest yielder.

    3.5.  What-won-where bi-plot

    The what-won-where view of the GGE bi-plot (Yan et al., 2000) is the best model for multi-environment trial data for grouping the environments and also to identify best performing genotype in each. Many researchers found this biplot intriguing, as it graphically addresses important concepts such as crossover GE, mega environment differentiation, specific adaptation (Yan and Tinker, 2006).

    Genotypes located on the vertices of the polygon performs either the best or the poorest in one or more environments. The biplot view classified that 27P31 (G14), JGL 24423 (G2) and RNRH 36 (G6) were the best performing genotypes in Kampasagar (E3), Rajendranagar (E5), Kunaram (E2) and Rudrur (E4) locations (Figure 5).

    Similarly, the hybrids JGLH 337 (G3), JGLH 275 (G11) were found to have good performance in Jagtial (E1) and RNRH 98 (G7), JGLH 373 (G12) in Warangal (E6) locations. Whereas the hybrids, RNRH 99 (G4) and Bio 799 (G5) fall in separate group with poor performance in many of the locations. Comparably, Mary et al., (2019) reported that the biplot for yield during the wet season showed that G10 was the winner genotype in E4, and G7 in E8 and E9. Rukmini Devi et al., (2020) reported the rice genotypes, G2 (WGRH-6) and G3 (WGRH-10), had better performance in mega environment E3 and the genotype, G9 (WGRH-18), exhibited better performance in the second mega environment consisting of E1 and E2.

  • Conclusion

    Rice hybrids, US 314 (G8) and RNRH 39 (G6) were more stable across locations with higher grain yields. GGE bi-plot environment view and ASV ranks confirmed that Rudrur (E4) location as the most ideal environment to obtain higher grain yields. Six locations were partitioned into three mega zones and to economize the genotype testing. One environment each from Northern and Southern Telangana Zone might be ignored as a testing location to ascertain hybrid performance. Further, 27P31 (G14) and RNRH 39 (G6) were the best performing genotypes in first mega zone.

  • Acknowledgement

    The authors are obliged to PJTSAU and ICAR-IIRR, Hyderabad, India for funding under AICRIP programme to different Rice Research Centers of Telangana, India.

  • Reference
  • Asenjo, C.A., Bezus, R., Acciaresi, H.A., 2003.Genotype×environment interaction in rice (Oryza sativa L.) in temperate regions using the Joint Regression Analysis and AMMJ methods.Cereal Research of Communication 32, 97–104.

    Becker, H.C., Leon, J., 1988. Stability analysis in plant breeding. Plant Breeding 101, 1–23.

    Crossa, J., Gauch, H.G., Zobel, R.W., 1990. Additive main effects and multiplicative interactions analysis of two international maize cultivar trials. Crop Science 30, 493–500.

    Das, S, Misra, R.C, Patnaik, M.C., Das, S.R., 2010. GxE interaction, adaptability and yield stability of mid-early rice genotypes. Indian Journal of Agricultural Research 44(2), 104–111.

    Anonymous, 2019a. FAOSTAT. Food and Agriculture Organization of the United Nations, Rome, Italy.

    Anonymous, 2019b. Pocket Book of Agricultural Statistics 2019, Ministry of agriculture & Farmers Welfare Department of agriculture, cooperation & Farmers Welfare Directorate of Economics & Statistics New Delhi pp:24-27.

    Farshadfar, E., Sutka, J., 2003. Locating QTLs controlling adaptation in wheat using AMMI model. Cereal Research of Communication 31, 249–255.

    Fentie, M., Assefa, A., Belete, K., 2013. AMMI analysis of yield performance and stability of finger millet genotypes across different environments. World Journal of Agricultural Sciences 9(3), 231–237.

    Gauch, H.G., Zobel, R.W., 1996. AMMI analysis of yield trials. In: Kang, M.S., Gauch, H.G. (Eds.), Genotype-by-environment interaction. Boca Raton, FL, CRC Press. Pp, 85–122.

    Gauch, H.G., 1992. Statistical analysis of regional yield trials: AMMI analysis of factorial designs. Elsevier, Amsterdam, The Netherlands, 278.

    Gollob, H.F., 1968. A statistical model that combines features of factor analysis and analysis of variance techniques. Psycrometrika 33, 73–115.

    Gupta, H.S., Dwivedi J.L., Sridhar, R., Kundu, S., Birthal P.S., Subba Rao L.V., Singh, T.V.K., 2020. Quinquennial review team report of ICAR-Indian Institute of Rice Research (2012-2020) and All India Co-ordinated Rice Improvement Project.

    Anonymous, 2021. Progress Report, 2020, Vol.1, Varietal improvement, all india coordinated rice improvement project, Rajendranagar, Hyderabad, India.

    Indiastat, 2020.

    Islam, M.R., Anisuzzaman, M., Khatun, H., Sharma, N., Islam, M.Z., Akter, A., Biswas, P.S., 2014. AMMI analysis of yield performance and stability of rice genotypes across different haor areas. Journal of Eco-friendly Agriculture 7(2), 20–24.

    Jat, A.L., Srivastava, V.K., Chongtham, S.K., Singh, R.K., 2020. Integrated Nitrogen Management Influences the Yield Performance of Rice under Different Methods of Cultivation. International Journal of Bio-resource and Stress Management 11(5), 472–481.

    Krishnamurthy, S.L., Sharma, P.C., Sharma, D.K., Ravikiran, K.T., Singh, Y.P., Mishra, V.K., Burman, D.,  Maji, B., Mandal, S., Sarangi, S.K., Gautam, R.K., Singh, P.K., Manohara, K.K., Marandi, B.C., Padmavathi, G., Vanve, P.B., Patil, K.D., Thirumeni, S., Verma, O.P., Khan, A.H., Tiwari, S., Geetha, S., Shakila, M., Gill, R., Yadav, V.K., Roy, S.K.B., Prakash, M., Bonifacio, J., Ismail, A., Gregorio, G.B., Singh, R.K., 2019. Identification of mega environments and rice genotypes for general and specific adaptation to saline and alkaline stresses in India. Scientific Reports 7, 1–14.

    Mahalingam, L., Mahendran, S., Chandra Babu, R., Atlim, G., 2006. AMMI Analysis for Stability of Grain Yield in Rice (Oryza sativa L). International Journal of Botany 2(2), 104–106

    Mary, A.I.A., Mallikarjuna Swamy, B.P., Amparado, A.F., Gwen Iris, L., D.E., Emily, C.A., Reinke, R., 2019. Stability and GGE analysis of zinc-biofortified rice genotypes evaluated in diverse environments. Euphytica 215, 1–17.

    Nirmala, B., 2015. Hybrid rice seed production in telangana and andhrapradesh states of india: a situation analysis. International Journal of Agriculture Sciences 7(14), 883–886.

    Nitiprasad, N.J., Bose, L.K., Pande, K., Singh, O.N., 2015. Genotype by environment interaction and stability analysis in rice genotypes. Ecology, Environment and Conservation 21(3), 1427–1430.

    Purchase, J.L. Hatting, H., Deventer, C.S.V., 2000. “Genotype×environment interaction of wheat in South Africa: stability analysis of yield performance,” South African Journal of Plant and Soil, 17(3), 101–107.

    Rao, P.J.M., Kishore, N.S., Sandeep, S., Neelima, G., Rao, P.M., Das, D.M., Saritha, A., 2020. Evaluation of performance and yield stability analysis based on AMMI and GGE-biplot in promising Pigeonpea [Cajanus cajan (L.) Millspaugh] genotypes. Legume Research, DOI: 10.18805/LR-4299.

    Rukmini Devi, K., Venkanna, V., Lingaiah, N., Rajendra Prasad, K., Satish Chandra, B., Hari, Y., Jagan Mohan Rao, P., 2020. AMMI Biplot analysis for genotype×environment interaction and stability for yield in hybrid rice(Oryza sativa L.) under different production seasons. Current Journal of Applied Science and Technology 39(48), 169–175.

    Umma, K.M., Hasan, M., Jamil Akter, A., Rahman, H., Biswas, P., 2013. Genotype-environment interaction and stability analysis in hybrid rice: an application of additive main effects and multiplicative interaction. Bangladesh Journal of Botany 42(1), 73–81.

    Yan, W., Hunt, L.A., 1998. Genotype by environment interaction and crop yield. Plant Breeding Review. 16, 35-178.

    Yan, W., Hunt, L.A., 2001. Interpretation of genotype ×environment interaction for winter wheat yield in Ontario. Crop Science 41, 19–25.

    Yan, W., Hunt, L.A., Sheng, Q., Szlavnics, Z., 2000. Cultivar evaluation and mega-environment investigation based on the GGE biplot. Crop Science 40, 597–605.

    Yan, W., Tinker, N.A., 2006. Biplot analysis of multi-environment trial data: Principles and applications. Canadian Journal of Plant Science 86, 623–645.

    Zewdu, Z., Abebe, T., Mitiku, T., Worede, F., Dessie, A., Berie A., Atnaf, M., 2020. Performance evaluation and yield stability of upland rice (Oryza sativa L.) varieties in Ethiopia. Cogent Food & Agriculture 6, 1–13.

People also read

Short Research

Preference Towards Online Mode of Distance Education Courses–conjoint Analysis

M. Malarkodi, V. M. Indumathi and S. Praveena

Conjoint analysis, distance education, preference, online learning

Published Online : 07 Feb 2018

Review Article

Cereal Residues - Not a Waste Until We Waste it: A Review

Amit Anil Shahane and Yashbir Singh Shivay

Cereal residue, ecological services, energy source, residue burning

Published Online : 07 Feb 2016