Stability analysis of sugarcane genotypes for cane yield using AMMI in sub-tropical multi-environment trials

Introduction

India's tropical and sub-tropical regions are important growing regions for sugarcane (Saccharum officinarum L.). With a total production of 405.4 million tons, the crop covers more than 5.64 million hectares in the country (Directorate of Economics & Statistics, 2024) [5] . In fact, after Brazil, India is the world's second-largest producer of sugar. A wide range of goods i.e. sugar and ethanol, are made from the crop. Furthermore, in an energy-poor nation like India, the bagasse residue left over after the juice is extracted in the sugar mill has become crucial for producing electricity (Quereshi et al. 2022) [20] . The high demand of sugarcane products has prompted policy makers to promote the expansion of this crop (Solomon and Swapna 2022) [25] . The variation in the environmental condition in sub-tropical India is wide during the sugarcane growth and maturity period with temperature ranging from 0-48 °C, photoperiod from 4-8 hr and relative humidity from 8-100% (Shukla et al. 2017) [24] . Hence, the sugarcane yield is generally low in sub-tropical part of India (Guo et al. 2021) [8] . However, sugarcane is primarily a tropical crop capable of producing high cane yield under ideal conditions. But in sub-tropics it is cultivated in erractic environments leading to varied varietal performance over different locations. The search for sugarcane genotypes with high cane and sugar yield which are also adapted in most varied environments is one of the most important objectives of sugarcane breeders. The development of stable varieties requires a broad genetic base within the populations under investigation. Such genetic diversity enables the identification of genotypes that demonstrate stable performance across a range of environmental conditions. This can be achieved by understanding the interaction between genotype and environment (Appunu et al. 2024) [2] . Genotype × Environment Interaction (GEI) refers to the variation in genotypic performance across different environmental conditions and plays a critical role in how genotypes perform under varying environments (Pour-Aboughadareh et al. 2022) [19] . The efficiency of selection and accuracy of varietal recommendation is reduced by G × E interaction (Mehareb et al. 2022) [15] .

Due to the interaction between genotype and environment (G × E), it is essential to analyze this interaction prior to the introduction of new high-yielding and stable genotypes across diverse environments. Various statistical methods, categorized into two main groups: parametric and non-parametric approaches have been developed to uncover the G × E interaction patterns. Parametric methods are firther divided into two groups: univariate and multivariate. Univariate methods include stability factor (Shojaei et al. 2022) [23] and regression-based approaches (Hashim et al. 2021) [10] . Multivariate methods include the AMMI (Additive Main Effects and Multipicative Interaction) model (Gupta et al. 2022) [9] and Genotypic Main Effect plus Genotype by environment (GGE) biplot analysis (Mullualem et al. 2024) [17] . Additive main effects and multiplicative interaction (AMMI) analysis has been proved as a useful analytic approach for linear and non-linear response of genotypes over the environmental conditions (Achenef 2022) [1] , which combines ANOVA (with additive parameters) and principal component analysis (with multiplicative parameters) into a single analysis (Gupta et al. 2022) [9] and interprets multi-environment data structure in breeding programs (Crossa et al. 2022) [4] . It is also an effective tool to diagnose genotype environment interaction patterns graphically (Mehareb et al. 2022) [15] . So, for the breeders to develop a variety suitable for different environments, the analysis of the stability of genotypes is the most important tool. In this study, nine sugarcane clones along with 3 checks were evaluated for cane yield across different environments to identify stable genotypes for general and specific adaptation in different sowing conditions and to estimate genotype × environment interaction and stability parameters. 1). The experiment at three different locations over two crop seasons with three crops constituted nine environments (Table 2). The experiments were laid out in randomized block design (RBD) with three replications at each location. Plots were of eight rows with 6.0m length having 0.90 m spacing between rows. Three budded sets were used for planting with a seed rate of 12 buds/m 2 at all the three locations. A combined analysis of variance across test environment and an AMMI analysis (Table 3) were carried out using the programme R software.

Materials and Methods

The AMMI model equation is

Where, Y ij = mean yield of i th genotype in the j th environment μ = general mean gi = i th genotypic effect ej = j th location effect λ n = eigen value of the Principal Component Axis n α in γ jn = i th genotype, j th environment Principle component analysis (PCA) scores for the PCA axis n θ ij = residual n = number of PCA axes retained in the model.

Environment-wise analysis and pooled analysis of variance were conducted as per normal procedure. If genotypes, environments and G x E interaction are significant, the analysis may proceed further for AMMI analysis.

Results and Discussion

AMMI Biplot Analysis

To increase the reliability of multi-environmental trial analysis, AMMI model was used to study the G × E interaction as AMMI model explains the genotype-environment interaction (Momotaz et al. 2021) [16] , which is used for reliable yield estimates (Yadawad et al. 2023) [28] and forms a robust base for enhancing the application of other analytical models (Oghan et al. 2022) [18] . The analysis of variance from the AMMI model revealed that cane yield was significantly (p < 0.001) affected by environment, genotype and genotype-environment interaction, which accounted for 53.41%, 26.80% and 19.78% of the total variation, respectively. Furthermore, the analysis showed that two principal components with significant differences collectively captured 71.00% of total genotype-environment (GEI). The first principal component explained 42.95% of the GEI while the second accounted for 28.04% of the genotype-environment interaction (Table 4). The present results clearly demonstrated significant contributions of environment, genotype and their interaction towards cane yield variation, in agreement with earlier findings (Mehareb et [15,22,6] . AMMI revealed that a major part of the variation in yield is explained by environment, which indicates that the environments were diverse. The results are line with the findings of Vinu et al. 2025 [27] and Kumar et al. 2023 [12] . The AMMI 1 Model

The AMMI biplot has the main effect as cane yield per plot in the abscissa and the IPCA1 as the ordinate where the genotypes or environments that lie on the same vertical line have the same yield, and those that lie on the same horizontal line have the same interaction pattern. Genotypes and environments exhibiting IPCA1 scores near zero were associated with minimal interaction effects and are therefore regarded as stable across environments (Elayaraja et al. 2022) [7] . In the AMMI 1 biplot, the sugarcane genotypes G7 and G12 are relatively stable and high yielding genotypes indicating their suitability as elite and broadly adapted cultivars. In contrast, G1 and G3 also exhibit stability across environments but are associated with lower yield performance, suggesting they are broadly adapted but less productive lines (Figure 1). Similar to this research, Durai et al. 2025 [6] also identified stable and having high cane yield genotypes Co 17001, CoC 13339 and Co 86032, inferring little interaction with the environment. The sugarcane genotypes G2, G4, G5, G6, G8 and G11 are relatively unstable in yield because these lines are far from the origin and can be specifically adapted to particular environment. Especially, genotypes CoPant 17223 and CoS 17235 were likely to perform better in the second plant crop at Uchani, whereas the genotype CoH 17262 was identified as specially adapted in ratoon crop at Gurdaspur. Ratoon crop at Uchani location was the most responsive environment for genotype CoH 17261 (Figure 1). Additionally, Meena et al. 2017 [14] identified clones such as Co 98014, Co 07025 and Co 07023 had larger interaction effects because they fall almost on same perpendicular line and away from IPCA-1 axis origin. Especially, the plant crop of Co 07023 and Co 07025 have appeared to have positive interactions with autumn (E1) and spring (E2) seasons and resulted in high cane yield in comparison to their ratoon crop yield in E3 and E4.

The AMMI2 Model

The AMMI2 biplot (Fig. 2 [16] , Vinu et al. (2025) [27] and Solomon & Swapna (2022) [25] and highlight the importance of using multiple stability measures for reliable selection.

According to the AMMI2 biplot, the best-adapted genotype with respect to 1 st plant crop in Uchani was CoH 17261, whereas the genotype CoS 17235 was highly adapted to 2 nd plant crop in Uchani, genotype CoS 17236 grouped in site E2, Co 05011 grouped into site E2 and E5. Similarly Kumar et al. 2018 [11] concluded that the genotypes Co 05018 and CoSe 092423 had shown the positive response for cane yield with E8 and E9, CoSe 05452 with E1 and E3. Similar results were further confirmed by Magalhães et al. 2018 [13] while analyzing the stability of 14 RB sugarcane genotypes by AMMI in the in the state of Pernambuco. Therefore, the above-mentioned genotypes were found most stable for cane yield and can be incorporated as breeding stocks in any future breeding programs aiming to produce high yielding cultivars of sugarcane.

Acknowledgement

Acknowledgments:The authors would like to express sincere thanks to CCSHAU, Hisar, Haryana and PAU, Ludhiana for providing the facilities to accomplish this research work.

References

1. Achenef G. Advancement of analytical models quantifying G × E interactions and stability analysis in multi-environment trial. Int J Res Agric Sci. 2022;9(4):103-120.
2. Appunu C, Hemaprabha G, Sreenivasa V, Durai AA, Mohanraj K, Elayaraja K, et al. Evaluation of sugarcane genotypes (Saccharum sp. hybrid) for multi-trait stability analysis across diverse environments. Ind Crops Prod. 2024;219:118993.
3. Barati A, Taherian M, Saeidnia F, Nikkhah HR, Arazmjoo E, Mahlooji M, et al. Analysis of genotype by environment interaction in barley across various locations in Iran using the AMMI model. Iran J Genet Plant Breed. 2024;13(1):11-19.
4. Crossa J, Montesinos-Lopez OA, Pérez-Rodríguez P, CostaNeto G, Fritsche-Neto R, Ortiz R, et al. Genome and environment-based prediction models and methods of complex traits incorporating genotype × environment interaction. In: Ahmadi N, Bartholomé J, editors. Genomic prediction of complex traits: Methods and protocols. New York (NY): Springer; 2022. p. 245-283.
5. Directorate of Economics and Statistics, Department of Agriculture and Cooperation. 3rd advance estimate of production of foodgrains 2023-24. New Delhi: Government of India; 2024.
6. Durai AA, Kumar RA, Hemaprabha G. Elucidating the G × E interaction using AMMI, AMMI stability parameters and GGE for cane yield and quality in sugarcane. Trop Plant Biol. 2025;18(1):3.
7. Elayaraja K, Govindaraj P, Mahadevwamy HK, Appunu C, Hemaprabha G, Rajesh V, et al. Stability analysis for identification of stable genotypes of sugarcane (Saccharum spp.) through AMMI model. Indian J Genet Plant Breed. 2022;82(4):616-622.
8. Guo H, Huang Z, Tan M, Ruan H, Awe GO, Are KS, et al. Crop resilience to climate change: A study of spatio-temporal variability of sugarcane yield in a subtropical region, China. Smart Agric Technol. 2021;1:100014.
9. Gupta V, Kumar M, Singh V, Chaudhary L, Yashveer S, Sheoran R, et al. Genotype by environment interaction analysis for grain yield of wheat (Triticum aestivum L.) genotypes. Agriculture. 2022;12(7):1002.
10. Hashim N, Rafii MY, Oladosu Y, Ismail MR, Ramli A, Arolu F, et al. Integrating multivariate and univariate statistical models to investigate genotype-environment interaction of advanced fragrant rice genotypes under rainfed condition. Sustainability. 2021;13(8):4555.
11. Kumar D, Kumar S, Meena LK, Meena LR, Singh SP. Stability analysis of sugarcane (Saccharum spp.) genotypes for metric and quality traits by AMMI model. Indian J Agric Sci. 2018;88(4):657-664.
12. Kumar R, Dhansu P, Kulshreshtha N, Meena MR, Kumaraswamy MH, Appunu C, et al. Identification of salinity-tolerant stable sugarcane cultivars using AMMI, GGE and some other stability parameters under multienvironments of salinity stress. Sustainability. 2023;15(2):1119.
13. Magalhães TA, Silva GC, Rocha PR, Silva JD, Mendonça Júnior AD, Oliveira FD. Adaptability and phenotypic stability of the sugarcane RB genotype by the AMMI method. J Agric Sci. 2018;10:484-492.
14. Meena MR, Karuppiayan R, Ram B, Kumar R, Kulshreshtha N. Genotype × environment interactions and stability analysis of sugarcane clones (Saccharum spp.) by AMMI model in sub-tropical regions of India. Indian J Genet Plant Breed. 2017;77(4):540-546.
15. Mehareb EM, Osman MAM, Attia AE, Bekheet MA, Abo Elenen FF. Stability assessment for selection of elite sugarcane clones across multi-environments based on AMMI and GGE-biplot models. Euphytica. 2022;218(7):95.
16. Momotaz A, Davidson RW, Zhao D, McCord PH, Sandhu HS, Baltazar M, et al. Genotype-by-environment interaction analysis across three crop cycles in sugarcane. J Crop Improv. 2021;35(2):276-290.
17. Mullualem D, Tsega A, Mengie T, Fentie D, Kassa Z, Fassil A, et al. Genotype-by-environment interaction and stability analysis of grain yield of bread wheat (Triticum aestivum L.) genotypes using AMMI and GGE biplot analyses. Heliyon. 2024;10(12):e16701.
18. Oghan HA, Bakhshi B, Rameeh V, Faraji A, Askari A, Fanaei HR. Comparison of AMMI, parametric and nonparametric models in identifying high-yielding and stable oilseed rape genotypes. Turk J Field Crops. 2022;27(2):224234.
19. Pour-Aboughadareh A, Khalili M, Poczai P, Olivoto T. Stability indices for deciphering the genotype-byenvironment interaction (GEI) effect: An applicable review for use in plant breeding programs. Plants. 2022;11(3):414.
20. Quereshi S, Naiya TK, Mandal A, Dutta S. Residual sugarcane bagasse conversion in India: Current status, technologies, and policies. Biomass Convers Biorefin. 2022;12(9):3687-3709.
21. Sanghera GS. Sugarcane disorders associated with temperature extremes and mitigation strategies. East Afr Sch J Agric Life Sci. 2020;3:101-114.
22. Sheelamary S, Karthigeyan S. Evaluation of promising commercial sugarcane genotypes for stability by AMMI analysis. Electron J Plant Breed. 2021;12(2):371-378.
23. Shojaei SH, Mostafavi K, Lak A, Omrani A, Omrani S, Mousavi SMN, et al. Evaluation of stability in maize hybrids using univariate parametric methods. J Crop Sci Biotechnol. 2021;24:1-8.
24. Shukla SK, Sharma L, Awasthi SK, Pathak AD. Sugarcane in India. In: Package of practices for different agro-climatic zones. Lucknow: All India Coordinated Research Project on Sugarcane, Indian Institute of Sugarcane Research; 2017. p. 1-64.
25. Solomon S, Swapna M. Indian sugar industry: Towards selfreliance for sustainability. Sugar Tech. 2022;24(3):630-650.
26. Srivastava AK, Rai MK. Sugarcane production: Impact of climate change and its mitigation. Biodiversitas J Biol Divers. 2012;13(4):214-221.
27. Vinu V, Alarmelu S, Elayaraja K, Appunu C, Hemaprabha G, Parthiban S, et al. Multi-environment analysis of yield and quality traits in sugarcane (Saccharum sp.) through AMMI and GGE biplot analysis. Sugar Tech. 2025;27(2):540-558.
28. Yadawad A, Patil SB, Kongawad BY, Kadlag AD, Hemaprabha G. Multi-environmental evaluation for selection of stable and high-yielding sugarcane (Saccharum officinarum L.) clones based on AMMI and GGE biplot models. Indian J Genet Plant Breed. 2023;83(3):389-397.