Multi-model assessment of yield stability in lentil across multi-environments
4 affiliations
- 1Nuclear Institute for Agriculture and Biology (NIAB), Faisalabad, Punjab, Pakistan
- 2Barani Agriculture Research Institute (BARI), Chakwal, Punjab, Pakistan
- 3Plant Pathology Research Institute, Ayub Agricultural Research Institute (AARI), Faisalabad, Punjab, Pakistan
- 4Sugarcane Research Institute, Ayub Agricultural Research Institute (AARI), Faisalabad, Punjab, Pakistan
Abstract
Developing high-yielding and stable lentil genotypes is essential for sustaining productivity under increasingly variable climatic conditions. However, strong genotype × environment (G × E) interactions often complicate the identification of widely adapted genotypes. Although several stability models are available, each has distinct strengths and limitations, making it uncertain whether a single model or a combination of models provides the most reliable selection. Therefore, this study compared multiple stability models to identify high-yielding and stable lentil genotypes and to evaluate their effectiveness for genotype selection. A set of lentil genotypes was evaluated across four diverse environments using a randomized complete block design (RCBD). Adjusted mean data were analyzed using the Eberhart and Russell regression model, AMMI (Additive Main Effects and Multiplicative Interaction), BLUP (Best Linear Unbiased Prediction), and WAASBY (Weighted Average of Absolute Scores and Yield). Significant G × E interactions confirmed differential genotype responses across environments. The Eberhart and Russell model identified stable genotypes based on regression coefficients close to unity and minimal deviation from regression. AMMI and BLUP effectively partitioned G × E effects and distinguished genotypes with broad and specific adaptation. WAASBY integrated yield and stability into a single index and was particularly effective in identifying genotypes with superior performance and wide adaptability. Correlation analysis among stability indices and the Genotype Selection Index (GSI) further strengthened the comparison and integration of model outputs. Genotypes G4, G8, G9, and G3 consistently exhibited high yield and stable performance across environments, with G4 emerging as the most promising and widely adapted genotype. This study demonstrates that integrating regression, mixed model, and multivariate based stability analyses provides a more robust framework for selecting stable, high-yielding lentil genotypes than relying on a single method, thereby supporting breeding for diverse agro-ecological conditions in Pakistan.
Related
- Beyond prediction accuracy by integrating GWAASB into wheat stability assessment across environments in Saudi Arabia
- Integration of multi-trait stability indices for selection of nutrient-dense and yield-stable pearl millet genotypes across environments
- GGE biplot and WAASB assessment of seed-purpose watermelon for fruit yield and 100-seed weight across hot semi-arid environments of India
- Multi-omics reveals accumulation mechanisms underlying differences in sesquiterpenoid content across different growth environments of Atractylodes macrocephala rhizome
- External Evaluation of Multi-Omics Prognostic Models Across TCGA-BRCA and METABRIC: Transportability, Stability, and Incremental Performance
BibTeX
@article{asghar2026multimodel,
title = {{Multi-model assessment of yield stability in lentil across multi-environments}},
author = {Muhammad Jawad Asghar and Maria Ghaffar and Muhammad Shahid and Khalid Pervaiz Akhtar and Amjad Hameed and Aqsa Tabasum and Ghulam Rabbani and Uzma Javed and Muhammad Umar Shahbaz and Muhammad Kamran and Muhammad Ehtisham-ul-Haq},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0356741},
url = {https://doi.org/10.1371/journal.pone.0356741},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Multi-model assessment of yield stability in lentil across multi-environments
AU - Muhammad Jawad Asghar
AU - Maria Ghaffar
AU - Muhammad Shahid
AU - Khalid Pervaiz Akhtar
AU - Amjad Hameed
AU - Aqsa Tabasum
AU - Ghulam Rabbani
AU - Uzma Javed
AU - Muhammad Umar Shahbaz
AU - Muhammad Kamran
AU - Muhammad Ehtisham-ul-Haq
PY - 2026
DA - 2026/09/01
JO - PLOS One
DO - 10.1371/journal.pone.0356741
UR - https://doi.org/10.1371/journal.pone.0356741
AB - Developing high-yielding and stable lentil genotypes is essential for sustaining productivity under increasingly variable climatic conditions. However, strong genotype × environment (G × E) interactions often complicate the identification of widely adapted genotypes. Although several stability models are available, each has distinct strengths and limitations, making it uncertain whether a single model or a combination of models provides the most reliable selection. Therefore, this study compared multiple stability models to identify high-yielding and stable lentil genotypes and to evaluate their effectiveness for genotype selection. A set of lentil genotypes was evaluated across four diverse environments using a randomized complete block design (RCBD). Adjusted mean data were analyzed using the Eberhart and Russell regression model, AMMI (Additive Main Effects and Multiplicative Interaction), BLUP (Best Linear Unbiased Prediction), and WAASBY (Weighted Average of Absolute Scores and Yield). Significant G × E interactions confirmed differential genotype responses across environments. The Eberhart and Russell model identified stable genotypes based on regression coefficients close to unity and minimal deviation from regression. AMMI and BLUP effectively partitioned G × E effects and distinguished genotypes with broad and specific adaptation. WAASBY integrated yield and stability into a single index and was particularly effective in identifying genotypes with superior performance and wide adaptability. Correlation analysis among stability indices and the Genotype Selection Index (GSI) further strengthened the comparison and integration of model outputs. Genotypes G4, G8, G9, and G3 consistently exhibited high yield and stable performance across environments, with G4 emerging as the most promising and widely adapted genotype. This study demonstrates that integrating regression, mixed model, and multivariate based stability analyses provides a more robust framework for selecting stable, high-yielding lentil genotypes than relying on a single method, thereby supporting breeding for diverse agro-ecological conditions in Pakistan.
SN - 1932-6203
ER -