Identifying male-sterile genes and developing biotechnology-based male-sterility systems are crucial for advancing hybrid maize breeding. However, this progress is hampered by the limited number of characterized key regulators and their incomplete mechanistic understanding in male sterility, as well as fertility instability of male-sterile lines and a lack of efficient maintainer lines in maize. Here, we elucidate the multifaceted roles of ZmMS1, an LBD transcription repressor. ZmMS1 coordinates timely tapetal PCD by repressing ROS-scavenging genes to regulate anther ROS homeostasis, while its DNA-binding activity is redox-sensitive, suggesting a potential redox-dependent feedback mechanism. In addition, ZmMS1 balances lipid allocation between anther cuticle and pollen exine by directly repressing sporopollenin biosynthesis and indirectly promoting cutin/wax formation. Constitutive overexpression of ZmMs1 induces dwarfism associated with GA and ABA homeostasis, and produces ~50% sterile and small pollen grains in maize and rice, offering a potential route for dwarf and male-sterile breeding. Leveraging the findings that loss and precocious expression of ZmMs1 cause recessive and dominant sterility, respectively, we develop a non-transgenic multi-control sterility system and a dominant genic male-sterility system, both showing stable and complete sterility across diverse backgrounds without yield penalty, thereby providing flexible options for hybrid maize breeding. Our findings reveal that ZmMS1, as a redox-sensitive transcription factor, regulates male fertility with previously unrevealed mechanisms, and provide practical tools for efficient hybrid maize breeding.

As the world's most widely cultivated and highest-yielding food crop, maize serves as a cornerstone of global food security, feed supply, and industrial raw material systems. Its breeding efficiency and the level of heterosis utilization are directly linked to global food security, sustainable agricultural development, and the stability of the agricultural industrial chain. In recent years, the deep integration of biotechnology (BT), information technology (IT), and artificial intelligence (AI) has become a transformative force, propelling maize breeding from the traditional "experience-driven" model into a new era of precision, intelligent, and data-intensive breeding jointly empowered by data, algorithms, and technologies. This paper systematically reviews the cutting-edge progress of the three core technology pillars in modern maize breeding. BT, represented by gene editing, doubled haploid induction, genomic selection, and multi-omics integration, has greatly accelerated the discovery, cloning, and functional validation of key genes governing yield, quality, stress resistance, and ecological adaptability. IT underpins breeding research through multi-omics big data, high-throughput phenomics, precision environmental monitoring, and genotype-phenotype-environment association analysis, establishing a solid data foundation for intelligent decision-making. AI, including machine learning, deep learning, predictive breeding models, intelligent breeding robots, and autonomous breeding platforms, realizes efficient data mining, intelligent prediction, and automatic design, acting as the "intelligent brain" of modern breeding. The review focuses on the cloning and regulatory mechanisms of representative functional genes in the past five years, as well as the research, development, and industrial application of disruptive innovative tools such as the GEAIR intelligent breeding robot and the AutoGP intelligent breeding platform. From the perspective of "trinity" synergy, this study constructs for the first time an integrated intelligent breeding framework in which BT provides core molecular tools, IT constructs a full-chain data system, and AI serves as intelligent decision-making engine. It further dissects the internal logic, synergistic mechanisms, and practical application closed-loop of the three technologies, and proposes a new precision design pathway for maize heterosis based on the "ontogenic hypothesis". Finally, the paper analyzes the key bottlenecks faced by the current technology integration, and prospects the future development trends of intelligent design breeding in stages, and provides important theoretical references and technical paths for China's maize seed industry to break through key bottlenecks in germplasm innovation, intelligent breeding, and efficient seed production, so as to support the high-quality, green, and sustainable development of the maize industry.

Natural genetic variation in diverse populations has long served as a key resource for crop improvement through genomic selection and genome editing. Despite advances in sequencing, genome assembly, and phenotyping technologies that enable millions of alleles to be linked to traits, the discovery of functional alleles has not kept pace. Although some quantitative trait loci (QTLs) have been fine-mapped to relatively small genomic intervals associated with phenotypic variation in target traits, their regulatory patterns have not been systematically summarized. In this review, we curate 762 functional alleles from maize, cotton, and rice and propose a practical methodological framework spanning six mechanistic categories, including transcriptional, post-transcriptional, and protein-structural regulation. We also provide detailed insights into their origins, identification strategies, and applications in hybrid breeding. This review provides valuable guidance for the future identification and utilization of functional variation in crop molecular breeding.

Gene chip biotechnology has emerged and been developed in the molecular breeding era. The rapid development of genomics, pan-genomics, molecular biology, and bioinformatics has driven the improvement and application of gene chip technology, supporting significant progress in crop genetics and evolutionary analysis, functional genomics research, and breeding applications. Gene chip technology is also integrated with traditional hybrid breeding and modern biological breeding technologies due to the high-throughput genotyping and high-precision marker detection capabilities, which greatly improve breeding efficiency and accuracy. More importantly, with the rapid development and practical application of big data and artificial intelligence technology, the deep integration of gene chips and intelligent breeding technologies will further improve the efficiency and precision of crop breeding and achieve rapid breakthroughs in variety breeding. It is worth noting that gene chip biotechnology has evolved from a high-throughput genotyping tool to a core engine for intelligent crop breeding. However, to achieve this transformation, gene chip technology and its applications still need to be closely integrated with the practical problems of biological breeding and the development trend of artificial intelligence technology. Here, with the focus on the transformation trend, we first review the history of the application of gene chip technology during the four stages (Breeding 1.0 to Breeding 4.0) of the crop breeding process. Then, we summarize the development progress of gene chips in major crops such as rice, wheat, corn, soybean, rapeseed, and cotton and describe the characteristics of solid-phase and liquid-phase chips, gene chips with whole genome coverage markers and functional markers, as well as developmental trends including increased sources for marker design, higher marker density, and diversified marker types. Subsequently, we summarize the four aspects of applications of gene chips, including genetic background testing (genetic map, DNA fingerprint), functional gene mining (QTL mapping, GWAS), breeding line screening (marker-assisted selection, genomic selection), and biosafety detection (detection of genetically modified ingredients). The application of gene chips runs through the entire process of biological breeding from germplasm identification to variety management. Finally, we forecast the new development progresses and possible development paths of crop breeding gene chips in the intelligent breeding era, including intelligent developments in chip design, automatic test, multi-source and heterogeneous data integration, data analysis, decision support, and construction of an integrated platform that supports above functions. In the future, an intelligent gene chip system integrating with big data, artificial intelligence, and advanced biotechnologies will form a core engine force to support the development and application of intelligent breeding technology in crops.

Maize ear rot severely restricts maize yield and quality, making the breeding of disease-resistant varieties the core strategy for disease prevention and control. Due to the highly uneven spatial distribution of lesions on maize ears, precise full-surface detection is essential for objectively quantifying disease severity. However, traditional manual disease grading is highly subjective, and conventional RGB-based detection methods struggle to precisely identify lesion regions associated with maize ear rot. These limitations hinder the precise identification and quantitative analysis of maize ear rot infection regions, thereby limiting the reliability of phenotypic data used for resistance evaluation and subsequent genome-wide association studies (GWAS). To address these challenges, this study developed an integrated full-surface hyperspectral imaging system featuring line-scan imaging and synchronous rotation control. Non-redundant full-surface ear images were then generated using the oriented FAST and rotated BRIEF (ORB) algorithm combined with random sample consensus (RANSAC), hereafter referred to as ORB-RANSAC. Furthermore, after Savitzky-Golay (SG) preprocessing and feature selection using a genetic algorithm (GA), three machine learning models and three deep learning models were established, and their classification performance was compared. The results showed that the convolutional neural network-bidirectional long short-term memory network (CNN-Bi-LSTM) model achieved the best average performance, with an average overall accuracy (OA) of 95.61 ± 0.36%. It also achieved higher overall accuracy than traditional machine learning models such as random forest (RF), indicating that CNN-Bi-LSTM can achieve high-precision pixel-level detection of lesion regions showing Fusarium-associated maize ear rot symptoms. Additionally, this model was deployed in locally developed automatic analysis software, enabling an integrated analysis workflow from raw hyperspectral data input to the quantification of disease-related phenotypic parameters. This study not only fills the technical gap in the non-destructive full-surface detection of maize ear rot but also provides an efficient and reliable automated tool for high-throughput phenomics research, which holds great significance for accelerating the discovery of maize resistance genes and ensuring food security.

Developing green and efficient agriculture is a critical strategy for addressing global environmental and demographic challenges, ensuring long-term national food security, and promoting agricultural sustainability. This review synthesizes key insights from presentations by fifteen academicians at the Second Crop Heterosis and Bio-breeding Conference of China. It focuses on how biological and information technologies empower the green transformation of modern agriculture, outlining a strategic pathway from scientific discovery to field application. Advancements in biological breeding form the genetic cornerstone of this transformation. Significant progress has been made in the molecular dissection of heterosis mechanisms, particularly through breakthroughs in elucidating "killer-protector" genetic systems, which have effectively overcome reproductive barriers in indica-japonica hybrid rice. Additionally, a novel paradigm for synergistically regulating growth and metabolism has been established. This approach, exemplified by key regulators such as the transcription factor GRF4, successfully decouples high yield from excessive fertilizer dependency. Concurrently, precision breeding systems have matured, underpinned by high-quality genomics and efficient gene-editing platforms. For example, highly optimized CRISPR-Cas systems in cotton have achieved editing efficiencies exceeding 85%, accelerating the genetic improvement of a diverse range of crops, including cotton, fruit trees, and specialty crops, and enabling rapid, targeted trait customization. In the research field of smart agriculture, the integration of big data analytics, artificial intelligence, and multi-scale phenotyping technologies is fundamentally reshaping research and production. This convergence enables intelligent design breeding, where machine learning models leverage genomic and phenotypic data to predict optimal crosses, potentially shortening breeding cycles by 30%-50%. Concurrently, a paradigm shift in crop protection is underway, moving from a reactive chemical-centric model to a proactive, ecosystem health-centered strategy. This smart plant protection framework utilizes integrated networks of sensors, drones, and AI-driven image recognition for real-time pest and disease monitoring, enabling precise, targeted interventions that minimize environmental impact. For sustainable green production, the review emphasizes the critical importance of integrated and optimized agronomic management to translate genetic potential into widespread, high-yielding, and low-footprint farming. It showcases successful regional practices, notably a systematized suite of eight key rice cultivation techniques, including optimized seedling cultivation, precise density management, and water-fertilizer coupling, and so on. This integrated approach has demonstrably closed the yield gap in demonstration zones while significantly reducing pesticide and fertilizer inputs. This is complemented by continuous innovation in green agricultural inputs, including the bio-rational design of novel, low-risk pesticides and the promotion of ecological farming systems like crop rotation, which collectively enhance soil health and system resilience. Looking forward, the scope of agriculture is expanding into functional and value-added dimensions. Directed breeding for enhanced nutritional and health-promoting traits, exemplifies this trend towards functional agriculture. Cross-disciplinary insights, particularly from biomedicine, underscore the universal importance of translating foundational research into tangible applications and the necessity of proactive science communication to foster public understanding and acceptance of innovative technologies. Despite these substantial advances, significant systemic challenges persist. These include a need for more original innovation in foundational tools and gene discovery; fragmented data ecosystems that impede the development of unified, powerful smart breeding platforms; a shortage of simplified, scalable, and economically viable technology packages tailored for smallholder farmers; and a pronounced "valley of death" between laboratory proof-of-concept and large-scale, market-driven commercialization. The review concludes that overcoming these hurdles and achieving the deep, systemic integration of biological, information, and green technologies is essential to construct a resilient, productive, and sustainable modern agricultural system. This integrative pathway is crucial for cultivating new quality productive forces within the agricultural sector, thereby securing China's food future and contributing to global food security and sustainable development goals.

As global climate change intensifies, the frequency and severity of crop abiotic stress events (e.g., droughts, heatwaves, floods, and salinity) are increasing, posing severe threats to agricultural production and national food security. In China, crop abiotic stresses exert pronounced impacts on staple crops, highlighting the urgent need for accurate, timely, and scalable monitoring approaches. Traditional approaches based on field surveys, empirical indices, and expert judgment struggle to capture early-stage physiological responses and spatiotemporal dynamics, limiting their ability to meet the practical demands of intelligent and large-scale crop stress monitoring. Consequently, they fall short of the requirements of modern smart agriculture and disaster risk management. In recent years, the deep integration of remote sensing and artificial intelligence (AI) technologies, known as "AI + remote sensing", has made significant progress in crop abiotic stress monitoring. This review provides a comprehensive overview of the integrated "space-air-ground" monitoring framework across different application scenarios. We discuss the applications of satellite platforms, unmanned aerial vehicle platforms, and ground-based sensor networks in monitoring crop abiotic stresses. The integration of these platforms provides multi-scale and multi-source datasets for crop stress detection, damage assessment, and phenotyping. Moreover, we outline the evolution from machine learning to deep learning approaches, as well as recent advances in multimodal remote sensing data fusion and explainable AI. AI-driven multimodal data fusion strategies integrate optical, thermal infrared, hyperspectral, and synthetic aperture radar data, substantially enhancing the capability for continuous, reliable, and large-scale monitoring of crop physiological responses under complex environmental conditions. In addition, emerging explainable AI approaches are highlighted for their ability to improve model interpretability and robustness while strengthening the link between model outputs and underlying agronomic or physiological mechanisms. Furthermore, we analyze the industrial demands and practical challenges of agricultural disaster prevention and mitigation in China, including data fragmentation, limited model generalization, and gaps between research outputs and operational deployment. We also explore how to address these challenges by coordinating data infrastructure, algorithmic innovation, and application-oriented system design. Overall, this review offers valuable insights into the development of intelligent agricultural disaster monitoring and early-warning systems and facilitates the translation of advanced "AI + remote sensing" technologies into practical applications for smart agriculture, sustainable crop production, and food security. By systematically reviewing technological breakthroughs in integrating remote sensing and AI technologies for monitoring crop abiotic stresses, this review provides scientific guidance for establishing smart monitoring and early-warning systems in China. It will drive the transformation of state-of-the-art research outcomes into practical technologies and operational tools across the crop production lifecycle.

Plant male sterility (MS) is a fundamental model for studying reproductive development, yet a critical but often overlooked distinction exists between sporophytic male sterility (SMS) and gametophytic male sterility (GAMS). Unlike SMS, which is controlled by the sporophyte's genotype, GAMS is uniquely determined by the gametophyte's own genome, leading to subtle phenotypes that are challenging to identify. Here, we systematically classify 59 identified GAMS genes in Arabidopsis, rice, and maize into three functional categories based on their defective phenotypes during pollen development: Pollen Maturation Defects (PMD), Pollen Germination Defects (PGD), and Male-Female Communication Defects (MFCD). Leveraging the evolutionary conservation of these pathways, we employed a comparative genomics approach to identify 18 high-confidence candidate GAMS genes in maize. Functional validation of ZmCESA6 as a PGD-type GAMS gene confirms the effectiveness of our strategy. This study provides a comprehensive functional map of GAMS and offers valuable genetic resources for hybrid breeding in major crops.

Plant cuticular waxes form a critical hydrophobic barrier covering aerial organs, serving as the first line of defense against abiotic and biotic stresses and playing a vital role in reproductive development. However, regulatory networks that orchestrate cuticular wax deposition in response to environmental cues and developmental programs, particularly in cereal crops, remain elusive. This review integrates current knowledge by identifying genes implicated in wax formation in Arabidopsis and major graminaceous crops. We detail the molecular mechanisms of wax biosynthesis and export, and place a major focus on the intricate transcriptional regulatory modules that integrate signals from drought, salinity, and pathogens, as well as developmental signals critical for anther cuticle formation and male fertility. Conserved and species-specific adaptations in these networks are highlighted, emphasizing how natural variation in these pathways underpins adaptive traits. We also discuss evolutionary perspectives and critically identify key knowledge gaps, such as the unresolved trade-offs between abiotic and biotic stress resistance and the mechanistic basis of anther cuticle development under heat stress, providing insights into leveraging cuticular traits for climate-resilient crop design.

Over the past few decades, China has been engaged in improving maize (Zea mays L.) production to ensure food security. However, it remains unclear whether the spring and summer maize planting regions had different responses to a changing climate. Therefore, this study aims to use smart agricultural techniques to analyze the sensitivities of maize growth to recent climate change in China during 2000-2020. The main objectives are to reveal differences in the relationship between climate and agriculture across the two regions, and to provide insights into optimizing agronomic practices in space and time. Satellite-based observations indicated that summer maize faced more challenging growth conditions (i.e., significant warming and drying) than spring maize during the last two decades. However, as the spring maize regions shifted toward favourable cooler and wetter conditions, its growth and yield increase rates were mostly higher than those of summer maize. To estimate the sensitivities of maize growth to climate factors and atmospheric carbon dioxide concentration (CO2), we employed random forest-based simulation experiments. The results demonstrated that spring maize growth was more sensitive to climate and CO2 than summer maize during the period. We found that the sensitivities were significantly different between the two regions. It seems that spring maize growth was more sensitive, likely due to its predominantly rain-fed nature, making it more vulnerable to climate fluctuations. Conversely, summer maize showed greater resistance, likely buffered by more well-developed irrigation facilities, allowing it to withstand adverse environments. Specifically, spring maize's high sensitivity to water supply illustrates the necessity of expanding irrigation to mitigate extreme events, while summer maize's exposure to heat stress calls for the deployment of heat-tolerant hybrids. These findings provide an evidence-based framework for region specific adaptation strategies, such as breeding improved plant architecture, optimizing planting density, and promoting the northward expansion of cropping systems, thereby ensuring crop yield stability under diverse climate challenges. Overall, this study provides decision-makers with essential support for developing adaptation strategies, highlighting the differential impacts of climate change on crops across various regions.

Global warming increasingly threatens crop productivity, necessitating new varieties that combine high yield potential with climate resilience. Optimizing plant architecture, particularly by reducing leaf angle (LA), enables dense planting and enhanced light capture, yet the mechanisms coordinating architecture improvement with heat tolerance remain largely unknown. Here, we identify the microRNA ZmMIR319A as a key regulator that simultaneously regulates LA and thermotolerance in maize. ZmMIR319A promotes LA by post-transcriptionally repressing Teosinte branched1/Cycloidea/PCF (TCP) transcription factors ZmTCP5 and ZmTCP44, thereby relieving their suppression of auxin transporter genes ZmPIN1 and ZmPIN8 to stimulate parenchyma cell proliferation in the ligular region. Under heat stress, rapid downregulation of ZmMIR319A leads to accumulation of ZmTCP5/44 proteins, which suppress the heat shock chaperone genes ZmHSP70-4 and ZmHSP90-5 in leaves to modulate thermotolerance, while continuing to restrain ZmPIN1/8-mediated growth. Thus, the ZmMIR319A-ZmTCP5/44 module bifurcates into two pathways in maize that coordinately govern LA via auxin transport and thermotolerance via heat shock protein signaling. Notably, ZmMIR319A knockout mutants achieve increased grain yield under high-density planting conditions despite reduced thermotolerance, highlighting a trade-off between architectural optimization and stress adaptation. Collectively, our findings reveal that a single regulatory module can simultaneously optimize plant architecture and stress adaptation, providing a dual-target strategy for molecular breeding of maize varieties with ideal plant architecture and climate resilience.

Hybrid maize performance depends strongly on the genetic purity of hybrid seeds, but female self-pollinated seeds and target hybrid seeds are difficult to distinguish by conventional visual inspection because of their highly similar phenotypes. This study developed a nondestructive and interpretable maize hybrid purity detection framework by integrating hyperspectral imaging and RGB-derived texture features. Hyperspectral images were acquired from the embryo and endosperm sides of five female parents, one common male parent, and their corresponding hybrids. Texture-based, full-band spectral, characteristic-band spectral, and texture-spectral fusion models were systematically constructed and compared. Competitive Adaptive Reweighted Sampling (CARS), Successive Projections Algorithm (SPA), and Synchronous Two-Dimensional Correlation Spectroscopy (Sync2D) were used for characteristic wavelength selection. The results showed that the embryo side provided more stable and discriminative spectral information than the endosperm side. Texture-only models showed limited ability to distinguish hybrids from female self-pollinated seeds, whereas embryo-side texture-spectral fusion models combined with CARS or SPA and Support Vector Machine (SVM) or Partial Least Squares Discriminant Analysis (PLS-DA) achieved average test accuracies of 0.99-1.00, meeting the national maize hybrid seed purity requirement of 97%. In the optimal low-dimensional models, the retained high-dimensional spectral variables were compressed to 28-69 key features, corresponding to a dimensionality reduction ratio of approximately 88%-95%. SHAP analysis identified mean saturation and seed size as important texture features; among spectral intervals, the 450-462 nm region appeared among the top-ranked embryo-side SHAP features in all five maize lines and showed the highest embryo-side mean absolute SHAP magnitude (0.0107 ± 0.0028). Overall, the proposed framework provides a high-throughput, low-dimensional, and interpretable solution for maize hybrid seed purity detection.

The rising prevalence of metabolic disorders underscores the need for a comprehensive understanding of dietary influences on health outcomes. Cereal grains are a major dietary source of fiber and are widely recommended for cardiometabolic health. Yet their main fiber fractions differ in structure and function, and whether these differences translate into distinct metabolic effects remains unclear. This review synthesizes evidence from randomized controlled trials to examine the metabolic effects of cereal-derived dietary fiber components, principally resistant starch, arabinoxylan, and composite dietary fiber, derived from the major cereal staples (wheat, maize, and rice). These fiber components consistently improved postprandial glucose and insulin responses, whereas fasting glucose, blood pressure, and most blood lipids changed little over the short term; the principal exception was a small but significant reduction in low-density lipoprotein cholesterol, confined to the composite fiber subgroup. The three components acted through partly distinct routes. Together these patterns point to an immediate, matrix-dependent mode of action alongside a slower, fermentation-linked one. Because fiber-rich, structurally intact matrices are a key feature distinguishing whole-grain from refined foods, these componentlevel patterns may help explain the metabolic effects observed in whole-grain trials. Recognizing both the functional fiber components and the food matrix in which they are consumed therefore provides a mechanistic framework for interpreting complex intervention results, and informs cereal processing strategies, dietary recommendations, and public health policies that aim to preserve the metabolic benefits of whole-grain foods.

Leaf area index (LAI) is an important structural parameter of crops and it is usually estimated non-destructively using reflectance spectra from various reflectometers. Prevailing models, often trained on single-crop and singleyear data, lack generalizability. As rotation crops with similar morphology, rice and wheat present an opportunity to develop generalized models; however, their spectral response patterns are not well compared, and adaptable multi-year, multi-crop LAI models remain scarce. To bridge this gap, we developed a generalized LAI estimation model for both crops by integrating physically-based simulation with data-driven deep learning. Key steps included canopy spectral simulation, data augmentation, and model construction with a 1D-CNN and transfer learning. The PROSAIL model was employed to simulate canopy reflectance spectra, with crop growth stages stratified into two phenological phases: sowing-heading stage (LAI: 0.01-5, increment: 0.2) and headinggrouting stage (LAI: 3-8, increment: 0.2). To enhance ecological fidelity, the LSMM was integrated to simulate mixed spectral scenarios involving soil background, water interactions, and spike contributions, while 5% Gaussian noise was systematically introduced to approximate real-world environmental variability. The results showed that the R2 values of the SMOTE-1D-CNN model for the different datasets (four rice and two wheat) ranged from 0.62 to 0.87, and the RMSE values ranged from 0.55 to 1.22. The model achieved a relatively high R2 (0.79 f 0.09) for rice LAI estimation but exhibited a larger RMSE (0.8 f 0.29). For wheat, the R2 was slightly lower (0.74 f 0.17), while the RMSE was smaller and more stable (0.56 f 0.01). These discrepancies reflect how crop characteristics or data distribution may influence estimation accuracy. SMOTE is used as a data enhancement to reduce the "high underestimation" phenomenon of the model, and the model performance is stabilized when the multiplicity of the sample size (n) is greater than or equal to 5. And the model input feature importance is only related to the original sample (the original unenhanced dataset) and does not change with "n". This study demonstrates that a hybrid methodology, fusing physically-based simulation with deep learning, offers significant potential for robust, multi-crop LAI inversion, providing novel insights and technical support for crop monitoring and management.

Soil salinization poses a global challenge to agricultural sustainability, crop productivity, and food security. In maize, salinity stress severely restricts root and shoot development, ultimately compromising yield and quality. Unlike the traditional descriptive structure, this review presents a method-validation-oriented workflow that summarizes the molecular and genetic basis of salinity tolerance and aims to integrate salt-responsive genic resources, mine candidate genes, and clarify their functional roles in maize. First, we synthesize independent studies on maize salinity tolerance and compile a curated set of reported salt-responsive genes. On this basis, we construct a regulatory network underlying plant responses to salinity stress, thereby outlining the evolving landscape of their genetic and molecular regulation. Second, we catalogue genic resources, including quantitative trait loci (QTLs), quantitative trait nucleotides (QTNs), and functionally validated genes, identify QTL/QTN hotspots, and validate a multi-omics integration strategy by mapping transcriptomic, proteomic, and metabolomic salt-responsive signals onto hotspot regions to prioritize candidate genes. Third, comparative collinearity analyses across maize, rice, wheat, and sorghum further reveal orthologous genes associated with salinity tolerance in maize. Through this workflow, we identify 19 previously uncharacterized genes involved in salinity stress responses, 14 of which are predicted to participate in three salt-responsive pathways: proline biosynthesis, ABA signaling, and the PEP bypass. Importantly, we further validate the practical utility of this review-derived prioritization by functionally testing two candidates using virus-induced gene silencing (VIGS). Collectively, this workflow provides a reusable, quality-controlled set of actionable targets for developing high-yielding, salt-tolerant maize and other crops through integrated genomics, systems biology, and advanced breeding technologies.

High-throughput maize phenotyping is essential for addressing the challenge of large-scale, precise identification and creation of elite maize germplasm, thereby supporting the revitalization of China's seed industry and promoting high-quality agricultural development. Traditional manual phenotyping suffers from low efficiency, strong subjectivity, and destructiveness, failing to meet the demand for large-scale, multi-temporal, and high-precision data in modern breeding. The rapid advancement of high-throughput phenotyping technologies offers an effective solution. Through the synergistic integration of multi-sensor technologies, phenotyping algorithms, and phenotyping equipment, high-throughput, nondestructive acquisition of phenotypic parameters has been achieved across four dimensions: structural phenotyping, physiological phenotyping, quality phenotyping, and yield prediction. Structural phenotyping focuses on threedimensional plant architecture and organ-scale geometric features. Physiological phenotyping addresses photosynthetic efficiency, water stress, and nutrient status. Quality phenotyping enables rapid assessment of kernel composition and seed vigor. Yield prediction builds estimation models driven by multi-temporal, multi-modal data. However, current phenotyping algorithms face limitations in data fusion interpretability and model generalizability. First, the fusion mechanisms of multi-source heterogeneous data remain unclear, and the "black-box" nature of deep learning models lacks biological interpretability. Second, existing models show insufficient generalizability across different ecological zones and varieties, with significant performance degradation during cross-scene transfer. Meanwhile, phenotyping equipment exhibits deficiencies in multi-sensor coordination and system integration, including limited spatiotemporal registration accuracy, mismatched acquisition and processing capabilities, and prominent cost-throughput tradeoffs. This paper systematically reviews the principles and applicable scenarios of RGB, multispectral, hyperspectral, LiDAR, thermal imaging, and fluorescence sensors, clarifying their advantages and limitations at the data acquisition level. It then summarizes research progress in high-throughput phenotyping across the four dimensions, covering two-dimensional image segmentation, three-dimensional point cloud reconstruction, spectral feature analysis, and deep learning modeling. Furthermore, the application status and limitations of commercial ear phenotyping platforms, three-dimensional reconstruction platforms, root phenotyping platforms, small-scale mobile field platforms, and large-scale fixed field platforms are analyzed, with emphasis on tradeoffs between throughput, accuracy, cost, and environmental adaptability. Finally, future directions are discussed. To address the specific bottlenecks posed by tall maize plants and dense canopies, this paper proposes developing maize-adapted sensors and heterogeneous collaborative equipment, as well as constructing large-scale phenotyping analysis models for complex organs. The specific value of high-throughput phenotyping in early breeding screening and genomic selection is also clarified, including assisting early evaluation of breeding materials, reducing breeding cycle time, and improving genomic prediction accuracy. This review aims to provide a systematic technical reference and promote the large-scale application and continuous innovation of high-throughput phenotyping technologies in maize breeding.