TL;DR
The results indicate that incorporating advanced computational methods into forage breeding programs can speed genetic progress while safeguarding sustainability in stress-prone agroecosystems and offer a robust, transparent, and reproducible means of evaluating multiple traits in genotype selection.
Background: Chickpea (Cicer arietinum L.) is one of the most important pulse crops in the world and it is greatly contributing to global protein security and sustainable agricultural systems. However, fluctuations in agro-climatic conditions and genotype-environment interactions considerably influence yield performance and seed quality traits. Accurate prediction and assessment mechanisms are therefore essential to support precision agriculture and breeding programs. Methods: This paper proposes an integrated biological-computational framework utilizing a novel hybrid attention-based random forest optimizer using artificial intelligence (HARFO-AI) for yield prediction and seed quality assessment in chickpea. Field experiments were conducted to determine the important morphological, physiological, and seed-related variables like plant height, chlorophyll index, number of pods per plant, weight of 100 seeds, germination percentage, vigor index and moisture content. These biological variables were used for training and testing the HARFO-AI model. Statistical parameters like coefficient of determination, root mean square error, mean absolute error, and classification accuracy were used to evaluate the performance of the HARFO-AI model. Result: The proposed framework yielded a prediction with R = 0.94, RMSE = 0.19 t ha, and MAE = 0.15 t ha, accuracy of 93.1%. The classification accuracy of seed quality was 95.2%. The comparative analysis showed that the performance was about 20% higher than that of the traditional regression models. The findings indicate that integrating biologically significant parameters with advanced AI optimization techniques enhances predictive reliability and provides an effective decision-support tool for precision breeding and sustainable chickpea production.
Background Rainfed cereal production in southwestern Saudi Arabia depends largely on farmer-saved landraces, creating an important opportunity to strengthen grain and fodder productivity in mixed crop-livestock systems exposed to heat and variable rainfall. This study evaluated introduced dual-purpose sorghum and pearl millet germplasm under contrasting arid, supplemental-irrigated rainfed environments, and applied complementary quantitative approaches to identify high-performing, stable candidates that balance grain yield and green-fodder production while accounting for genotype × environment interaction (G×E). Methods Twenty-seven sorghum and 15 pearl millet entries were evaluated in alpha-lattice trials at the contrasting Jazan lowland and Abha highland sites across four site × season/sowing-window environments. Linear mixed models quantified environment, genotype, and G×E effects and estimated entry-mean reliability. Stability and adaptation patterns were evaluated using the weighted average of absolute scores from BLUPs (WAASB) and GGE biplots. Dual-purpose performance was assessed using a standardized 50:50 grain-yield-fresh-biomass index as a transparent baseline in the absence of validated local economic or farmer-preference weights. Pareto-frontier analysis was used to identify non-dominated genotypes representing efficient grain-green-fodder trade-offs. Results The trials revealed substantial genetic variation and contrasting genotype responses across environments, demonstrating considerable scope for improving grain and green-fodder productivity. Selection reliability was very high for sorghum fresh biomass and pearl millet grain yield, and high for pearl millet fresh biomass, providing a strong basis for prioritizing selection candidates within the target production environments. Sorghum grain yield showed lower reliability because only two grain-evaluable environments were available and G×E variance was large, highlighting a clear priority for expanded grain testing. SSV 74, WM 89/90#1615, and 89WM 5003 achieved the highest sorghum DualIndex values, while ICSV 15021 combined the highest mean fresh biomass with the lowest WAASB, identifying it as a particularly promising forage-oriented candidate. In pearl millet, LCICMV-1 (SOSAT-C-88) and ICMV 91450 displayed favorable grain-green-fodder profiles, ICMV 221 recorded the highest mean grain yield, and LCICMV-4 (Jirani) showed promising grain-yield performance under Jazan-associated conditions. Conclusions The study identified a valuable portfolio of sorghum and pearl millet candidates addressing complementary production objectives, including high green-fodder productivity, balanced grain-fodder performance, high grain yield, and targeted environmental adaptation. By integrating mixed-model reliability, stability analysis, a transparent grain-green-fodder index, and Pareto-frontier assessment, the study provides a rigorous and practical framework for converting early multi-environment data into evidence-based advancement decisions. The identified genotypes provide a strong foundation for expanded multi-location and participatory on-farm validation, standardized fodder-quality assessment, and staged seed-system development. Collectively, these findings advance the development of productive and resilient dual-purpose cereal options capable of strengthening grain and fodder security in southwestern Saudi Arabia and comparable arid mixed crop-livestock systems.
Plant breeding forms a vital part of horticulture, contributing significantly to the creation of new varieties with desirable traits. Diverse breeding methodologies have been adopted to achieve improvements in yield, biotic stress resistance, nutritional attributes, and adaptability to varying environments. In this pursuit, stability models are developed to release a promising line as a variety at the end of the evaluation trials. However, regular approach results in various stable line for various traits, making difficult for a holistic recommendation. Thus, through trait pyramiding concept, non-parametric stability models are developed to identify stable line(s) across all traits. This concept is demonstrated using various yield attributing traits of 15 Chrysanthemum lines evaluated at ICAR-Indian Institute of Horticultural Research, Bengaluru during 2023 to 2025. R-codes were developed for the analysis. Based on the Non-parametric stability models of Venugopalan indices, it was found that IIHR 4-8 and IIHR 2-16 are stable across all traits evaluated such as flower diameter (cm), number of flowers/plant, flower stalk length (cm), plant height (cm), plant spread (cm). These stable lines will be utilized effectively to release a variety holistically combining all desirable traits in crop hybridization trials, emphasizes the need to construct a non-parametric index for assessing the stability of a set of lines collectively based on various traits evaluated over multiple seasons or years in an experimental setup
Phenomic selection (PS) offers a cost‐effective , breeder‐friendly approach for public breeding programmes with limited access to genotyping or restricted financial resources for laboratory infrastructure. Since PS relies on high‐throughput phenotyping data, which is often derived from near‐infrared spectroscopy (NIRS) of harvested seeds, prediction is challenged by the high dimensionality and strong intercorrelation of NIRS data, which means that only a subset of wavelengths is informative. This study evaluates spectral variable selection models for predicting key morpho‐agronomic and quality traits in malt barley ( Hordeum vulgare L.) and assesses their performance under realistic breeding scenarios. Four NIRS‐based regularized regression models (Lasso, Enet, Ridge and a heritability‐filtered Ridge model) were tested to predict 10 morphological, agronomic and quality traits measured in two malt barley trials conducted during the 2022 and 2024 cropping seasons at three locations in Ethiopia using 100 genotypes in each trial. Model performance was evaluated across four practical breeding scenarios: within‐location unseen genotype prediction (WL‐uG), leave‐one‐location‐out prediction (LOLO), target environment unseen genotype prediction (TargetEnv) and across‐location wide adaptability (RuG). Accordingly, Cross‐validation identified stable, informative spectral predictors for each trait, scenario and trial. Among the models tested, Lasso and Enet consistently outperformed Ridge regression, with Enet showing the best predictive performance across scenarios. Prediction ability (r) ranged from 0.15 to 0.85 for quality traits, 0.16 to 0.79 for agronomic traits and 0.07 to 0.89 for morphological traits across scenarios. Thus, these findings underscore the importance of spectral predictor selection in improving predictive ability and demonstrate the transferability of PS in barley breeding.
Introduction The identification and advancement of superior maize hybrids under the All India Coordinated Research Project (AICRP) on Maize rely on multi-environment evaluation integrating grain yield, maturity, and agronomic performance. Interpretation of large multi-environment datasets is often complex, time-consuming, and susceptible to subjectivity, highlighting the need for objective and reproducible decision-support tools. This study evaluated the effectiveness of REMATTOOL-R (Relative Maturity Adjustment Tool in R) in validating the existing hybrid advancement framework adopted under the AICRP on Maize. Methods Multi-environment trial data from the National Initial Varietal Trial (NIVT)-Late conducted during Kharif 2020–21 across five locations representing the Central West Zone (CWZ) of India were analysed. The dataset comprised 45 entries, including 40 experimental hybrids, four commercial checks, and one filler entry. REMATTOOL-R integrated grain yield with days to 50% anthesis, grain moisture at harvest, and harvested plant stand to facilitate simultaneous evaluation of grain yield, maturity, and adaptation-related traits. Least-square means generated from mixed-model analysis were used to identify superior hybrids based on a predefined grain yield superiority threshold (≥5%) over the standard check while maintaining comparable maturity and agronomic performance. Results REMATTOOL-R enabled rapid visualization and integrated assessment of multiple agronomic traits, allowing objective identification of superior hybrids. Five experimental hybrids—PM 21109L (Entry 30), R8050 (Entry 35), PM 21111L (Entry 32), BIO 978 (Entry 4), and DKC 9226 (Entry 9)—recorded ≥5% higher grain yield than the standard check Bio 9682 while maintaining statistically comparable days to 50% anthesis, grain moisture at harvest, and harvested plant stand. All five hybrids identified by REMATTOOL-R corresponded with the official AICRP decisions for advancement from NIVT to Advanced Varietal Trial-I (AVT-I), while three hybrids (R8050, PM 21111L, and DKC 9226) progressed further to AVT-II during subsequent testing cycles, confirming the reliability of the analytical framework. Discussion The findings demonstrate that REMATTOOL-R provides an efficient, transparent, and reproducible framework for the simultaneous evaluation of grain yield, maturity, and adaptation-related traits in maize multi-environment trials. By complementing the existing AICRP hybrid evaluation procedure, the tool facilitates objective advancement decisions and reduces subjectivity associated with manual interpretation of complex datasets. REMATTOOL-R therefore represents a valuable decision-support approach for coordinated maize breeding programmes and has considerable potential for application in large-scale hybrid evaluation systems.