Search Results (8,333)

Adenosine triphosphate (ATP) serves as a central energy currency and signaling molecule in cellular processes, with ATP-binding sites in proteins playing critical roles in enzymatic catalysis, signal transduction, and gene regulation. The accurate identification of ATP-binding sites is essential for understanding protein function mechanisms and facilitating drug discovery, enzyme engineering, and disease pathway analysis. In this study, we present a novel hybrid deep learning framework that synergizes heterogeneous learning paradigms based on protein sequence information for accurate ATP-binding site prediction. Our approach integrates two complementary base classifiers. One is a Transformer-based model, which leverages high-level contextual embeddings generated by Evolutionary Scale Modeling 2 (ESM-2), a state-of-the-art protein language model, combined with a local–global dual-attention mechanism that enables the model to simultaneously characterize short-segment and long-range contextual dependencies across the entire protein sequence. The other is a deep Q-network (DQN)-inspired classifier that achieves residue-level prediction as a sequential decision-making process. The final predictions are generated using a weighted ensemble strategy, where optimal weights are determined via cross-validations to leverage the strengths of both models. The prediction results on benchmark independent testing sets indicate that our method achieves satisfactory performance on key metrics. Beyond predictive efficacy, this work uncovers the intrinsic biological mechanisms underlying protein–ATP interactions, including the synergistic roles of local structural motifs and global conformational constraints, as well as family-specific binding patterns, endowing the research with substantial biological significance. The research in this work offers a deeper understanding of the protein–ligand recognition mechanisms and supportive efforts on large-scale functional annotations that are critical for system biology and drug target discovery. Full article

Esophageal cancer is a highly aggressive malignancy with a poor prognosis. More precise prognostic biomarkers are therefore needed. Disulfidptosis is a recently identified form of regulated cell death driven by disulfide stress. It has been implicated in tumor progression. However, its prognostic role in esophageal cancer remains largely unexplored. This study aimed to develop a disulfidptosis-related gene signature for risk stratification and outcome prediction in esophageal cancer patients. Based on 23 disulfidptosis-related genes, consensus clustering was performed to identify molecular subtypes. Differentially expressed genes (DEGs) between subtypes were subjected to functional enrichment, immune microenvironment, and drug sensitivity analyses. Univariate and multivariate Cox regression were used to construct a prognostic risk model, which was evaluated using time-dependent receiver operating characteristic (ROC) curve and Kaplan–Meier analysis. A clinical nomogram integrating the risk score and clinicopathological factors was developed and validated. Two distinct disulfidptosis-related subtypes were identified, showing significant differences in gene expression, immune infiltration, and stromal scores. A total of 1080 DEGs were enriched in pathways related to epidermal differentiation, NRF2 signaling, and glucocorticoid receptor activity. A five-gene prognostic signature was established and effectively stratified patients into high- and low-risk groups. The risk model exhibited strong discrimination for 1-, 3-, and 5-year overall survival outcomes. The predictive accuracy was further maximized through an integrated clinical nomogram, which achieved an outstanding area under the curve (AUC) of 0.94 for 5-year survival predictions. Drug sensitivity analysis revealed subtype-specific therapeutic vulnerabilities, supporting potential precision treatment strategies. This study proposes a novel disulfidptosis-related five-gene signature and nomogram that robustly predict prognosis in esophageal cancer. The findings highlight the clinical relevance of disulfidptosis in tumor biology and offer a potential tool for risk stratification and personalized therapeutic decision-making. Full article

Background and Objectives: Coronavirus disease 2019 (COVID-19) is characterized by excessive inflammatory responses, including the so-called cytokine storm, which contributes substantially to morbidity and mortality in hospitalized patients. The vagus nerve, through the cholinergic anti-inflammatory pathway, represents a theoretically attractive therapeutic target for modulating systemic inflammation. Vagus nerve stimulation (VNS) has emerged as a potential adjunctive treatment for COVID-19, with several randomized controlled trials (RCTs) investigating its efficacy on inflammatory biomarkers and clinical outcomes. The quality of this evidence base has not been rigorously evaluated. This systematic review critically appraises all available RCT evidence for VNS in hospitalized COVID-19 patients. Materials and Methods: We systematically searched PubMed, Scopus, Cochrane (CENTRAL), and Web of Science from database inception to January 2026, for RCTs evaluating any form of VNS (invasive, non-invasive, cervical, or auricular) in hospitalized patients with confirmed acute COVID-19. Two reviewers independently screened titles, abstracts, and full texts according to pre-specified eligibility criteria. Risk of bias was assessed using the Cochrane Risk of Bias 2 (RoB 2) tool, with assessments initially performed using multiple artificial intelligence tools and subsequently validated by the authors in accordance with PRISMA 2020 guidelines. Given substantial heterogeneity and high risk of bias, narrative synthesis was performed rather than meta-analysis. Also, GRADE assessment was performed. Results: From 437 records identified, six RCTs comprising 221 patients met the inclusion criteria. Five trials (83%) were rated as high risk of bias, primarily due to inadequate blinding, substantial baseline imbalances, significant missing data and extensive multiple testing without statistical correction. The single double-blind trial with a credible sham control (Rangon et al.) found null results across all outcomes, including clinical progression, ICU transfer, and mortality, while the five “high” risk-of-bias trials generally reported positive findings on various inflammatory markers and clinical outcomes. One trial (Corrêa et al.) measured heart rate variability as a direct indicator of vagal activation and found no change despite claiming anti-inflammatory effects, contradicting the proposed mechanism of action. Significant cognitive findings from an interim analysis (Uehara et al., n = 21) disappeared in the larger completed trial (Corrêa et al., n = 52), providing empirical demonstration of false positive findings in small, underpowered studies. Conclusions: Currently available evidence supporting the use of VNS for acute COVID-19 remains scarce; however, the physiological rationale remains sound, although the absence of reliable target engagement markers in the included studies limits confidence in this treatment method. Large-scale, double-blind, sham-controlled trials are required before VNS can be firmly recommended for COVID-19 management. Full article

With the rapid development in data mining and learning analytics, integrating predictive analytics into educational data has become increasingly critical for supporting students’ learning trajectories. In many schooling systems, the academic tracks (such as Liberal Arts or Science) and the performance of junior high school students can substantially shape their subsequent university pathways and career planning. Despite the long-term impact of these decisions, academic track selections and the evaluation of students’ potential are often made without systematic and evidence-based guidance. Predictive computer applications can assist, but the training of accurate models and the selection of adequate features remain key challenges. This paper details our process of engineering such an application comprising two tasks based on 1357 real-world junior high school academic performance records. The first task applies a classification approach to predict students’ academic track orientation, while the second task employs a multi-output regression model to forecast students’ future academic performance in senior high school. Our approach shows that the stacking ensemble model achieved a classification accuracy of 85.76%, whereas the Bi-LSTM model with multi-head attention attained an overall R² exceeding 82% in performance forecasting; both models demonstrated strong and reliable predictive capability. Moreover, the proposed approach provides inherent interpretability by decomposing predictions at the subject level. Feature importance analysis reveals how different academic subjects contribute variably to both academic track decisions and future academic performance, offering actionable insights for academic counselling and future study planning. By bridging predictive modelling with students’ educational and career planning needs, this study advances the practical application of educational data mining and provides support for evidence-based academic guidance and future career choices in real-world contexts. Full article

Background/Objectives: Retinoblastoma represents the most common intraocular malignancy in childhood; however, the clinical applicability of mitomycin C (MMC) is restricted by dose-dependent ocular toxicity. Consequently, the development of pharmacological strategies that sensitize tumor cells to MMC while allowing dose reduction remains an unmet therapeutic objective. In this context, quercetin, a bioactive flavonoid with pleiotropic anticancer properties, has emerged as a potential chemosensitizing agent. Methods: Human retinoblastoma cell lines Y79 and WERI-Rb1 were exposed to MMC and quercetin, administered either individually or in fixed-ratio combinations. Cytotoxic responses were quantified through dose–response modeling and IC₅₀ determination following 24 and 48 h of treatment. Drug–drug interactions were quantitatively characterized using the Chou–Talalay combination index (CI) approach and isobologram analysis. Cell cycle distribution was assessed by propidium iodide (PI)-based flow cytometric analysis to evaluate treatment-associated alterations in cell cycle progression. Apoptotic cell death was assessed by Annexin V-FITC/PI flow cytometry, while transcriptional modulation of genes associated with apoptosis, cell cycle regulation, and oxidative stress (BAX, BCL-2, TP53, CASP3, CDKN1A, and HMOX1) was evaluated by qRT-PCR. Modulation of tumor-supportive signaling was examined by measuring VEGF and IL-6 secretion. Translational relevance was further investigated using a three-dimensional (3D) tumor spheroid model, and the functional contribution of reactive oxygen species (ROS) was interrogated through N-acetyl-L-cysteine (NAC) rescue experiments. Results: Quercetin significantly enhanced the cytotoxic activity of MMC in both retinoblastoma cell lines, with CI values below 1 across IC₅₀–IC₉₀ effect levels, indicating a synergistic pharmacological interaction. PI–FACS analysis revealed that combined MMC and quercetin treatment induced a pronounced accumulation of cells in the G2/M phase, consistent with cell cycle arrest, with a more marked effect observed in Y79 cells compared with WERI-Rb1 cells. Combination treatment resulted in a pronounced increase in apoptotic cell populations compared with single-agent exposure and triggered a coordinated pro-apoptotic transcriptional response, characterized by increased expression of BAX, TP53, CASP3, CDKN1A, and HMOX1, alongside suppression of BCL-2 and a marked shift in the BAX/BCL-2 ratio. Concurrently, VEGF and IL-6 secretion were significantly reduced, reflecting attenuation of pro-angiogenic and pro-inflammatory signaling. Notably, synergistic cytotoxicity was maintained in 3D tumor spheroids, where combined treatment induced spheroid shrinkage, architectural disruption, and reduced viability. NAC pretreatment diminished ROS accumulation and partially restored cell viability, indicating that oxidative stress contributes to, but does not solely account for, the observed synergistic cytotoxic effect. Conclusions: Collectively, these findings indicate that quercetin appears to function as an effective chemosensitizing adjuvant to MMC in retinoblastoma models, through transcriptional changes consistent with p53-associated apoptotic signaling at the transcriptional level, G2/M cell cycle arrest, and partial involvement of ROS-related cellular stress responses, along with suppression of tumor-supportive signaling pathways. The preservation of synergistic activity in 3D tumor spheroids supports the potential preclinical relevance of this combination. However, these findings are based on transcriptional and phenotypic analyses and should be interpreted as hypothesis-generating, requiring further validation through protein-level and in vivo studies before translational application. Full article

Spontaneous fermentation processes can promote uncontrolled microbial growth and increase the risk of foodborne contamination, making the characterization of artisanal beverages essential for consumer safety. This study investigated the microbial composition of quinoa-based rejuvelac, a homemade fermented drink often perceived as a functional food, with the objective of identifying potential microbiological hazards associated with its preparation. High-throughput sequencing of the 16S rRNA V3–V4 region was combined with shotgun metagenomics to profile bacterial communities and recover metagenome-assembled genomes. The analysis revealed a strong dominance of Pseudomonadales, mainly Pseudomonas, Acinetobacter, Enterobacter and Burkholderiales, while lactic acid bacteria typically responsible for stable and safe fermentations were not detected. Shotgun metagenomics recovered medium- to high-quality genomes from Burkholderiaceae and Clostridiales, supporting the overrepresentation of non-beneficial taxa and indicating deviations from expected fermentation microbiota. These results show that the spontaneous preparation of rejuvelac may favor bacterial groups associated with environmental contamination rather than fermentative pathways, underscoring the importance of hygiene practices, controlled starter cultures and monitoring strategies to mitigate microbiological risk. The study highlights the need for improved safety standards in artisanal fermented foods to prevent unintended microbial contamination and protect consumers. Full article

Background: Ulcerative colitis (UC) is an intestinal disease characterized by long-term inflammation. Circadian rhythm disorder (CRD) affects various biological activities and has been linked to several diseases, including UC. This study aimed to investigate the role and significance of CRD in UC. Methods: Bulk RNA-seq data from five independent UC cohorts were obtained from the Gene Expression Omnibus (GEO) database and integrated into a single dataset. The dataset underwent differential analysis to identify differentially expressed genes (DEGs) in association with CRD. Expression levels and pathway enrichment of CRD genes were analyzed, and signature genes were identified using machine learning algorithms. Based on these signature genes, a UC risk prediction model and CRD-related molecular subtypes were established. Furthermore, single-cell RNA-seq data of UC were analyzed to discuss the key role of CRD and signature genes in the UC microenvironment. RT-PCR analysis was employed to validate the expression levels of the identified signature genes. Results: 247 DEGs associated with CRD in UC were identified (referred to as CRD-DEGs). Gene set enrichment analysis (GSEA) revealed a strong association between CRD and inflammation, as well as immune cell infiltration in UC. This association potentially impacts intestinal fibrosis. A comparison of three machine learning algorithms (Lasso, SVM-RFE, and Random Forest) resulted in the identification of 12 signature genes. A UC risk prediction model and two UC CRD subtypes were developed using these genes. Among them, STXBP1 was identified by all three machine learning algorithms and was further analyzed. STXBP1 was predominantly enriched in pathways related to inflammatory response. Elevated levels of STXBP1 are mainly caused by reduced levels of methylation of its gene promoter. RT-PCR confirmed elevated expression of certain genes in mouse UC models. Conclusions: This study is the first to establish a strong association between CRD and the onset of UC. The newly developed UC nomogram based on CRD demonstrated high predictive accuracy, although further clinical validation is required. Understanding the intrinsic relationship between CRD and UC enhances our understanding of the potential pathogenesis of UC. This study introduces novel ideas and methods for early diagnosis, treatment, and prognosis of UC. Full article

This study examines the impact of family multidimensional poverty on cognitive function among older adults in China using the 2018 Chinese Longitudinal Healthy Longevity Survey (CLHLS). Filling a critical gap in the existing literature, we construct a multidimensional poverty index (MPI) based on the Alkire-Foster methodology to evaluate cognitive decline within the context of China’s post-poverty-eradication landscape. Utilizing quantile regression analysis, our findings demonstrate that multidimensional poverty exerts a significant, negative effect on cognitive function, which is more pronounced among individuals at lower cognitive quantiles, consistent with the cumulative disadvantage theory. Furthermore, we identify substantial urban–rural and regional disparities, revealing unique socio-economic inequalities. By linking multidimensional poverty to elderly cognitive health through psychosocial pathways, this study provides empirical evidence that reducing multidimensional deprivation among older adults is integral to achieving both SDG1 and SDG3 in China’s post-eradication context, demonstrating that income-based metrics alone are insufficient to capture the full burden of poverty on elderly cognitive health. Full article

Tamusu, the only identified super-large sandstone-hosted uranium deposit in the Bayingobi Basin, provides an important natural laboratory for evaluating ore-controlling factors and genetic models of sandstone-type uranium mineralization. Based on core descriptions from more than 200 boreholes, log facies analysis and geochemical environmental proxies, this study constrains the sedimentary–mineralization architecture and key controlling factors of the deposit. Uranium orebodies are mainly hosted in the upper member of the Lower Cretaceous Bayingobi Formation (Sq2) within a gravity flow-dominated fan-delta–lacustrine system. Braided distributary channel sands on the fan-delta plain and subaqueous distributary channel sands on the delta front constitute the principal uranium reservoirs, controlling both the migration pathways and storage space for U-bearing fluids. Mineralization is jointly governed by fan-delta architecture, interlayer oxidation zonation and reducing agents. The interlayer oxidation zone displays a north-thick–south-thin geometry, and uranium orebodies are concentrated at redox transition positions, with grades of 0.01–0.33 wt%. The metallogenic evolution can be summarized in three stages: syndepositional uranium pre-enrichment, interlayer oxidation mineralization, and a late hydrothermal/diagenetic overprint that mainly modified reservoir properties, favored ore preservation, and did not contribute to the primary uranium budget. Accordingly, a genetic model of “fan-delta architecture + interlayer oxidation control + late overprint and preservation” is proposed to guide exploration in the Bayingobi Basin and analogous sandstone-type uranium systems. Full article

Resource-based cities face unique land use challenges due to resource dependence and path lock-in, yet the driving mechanisms and future trajectories of their land use transitions remain underexplored. This study examines the Huaihai Economic Zone (HEZ), a representative coal-rich region in eastern China, to analyze land use changes from 2000 to 2023 and simulate 2036 scenarios under different development pathways. Using land use transfer matrices, dynamic degree metrics, and the Patch-generating Land Use Simulation (PLUS) model, we systematically identified spatiotemporal evolution patterns, quantified the contributions of driving factors, and projected multi-scenario future land use patterns. Results reveal that land use change in the study area was dominated by the conversion of cultivated land to construction land, alongside spatial restructuring from a monocentric to a polycentric network pattern. Notably, construction land expansion was least evident in the central Mining-Affected Zone, where land use changes remained relatively sluggish compared to other sub-regions. Driving factor analysis indicates that socio-economic factors primarily influenced changes in construction and cultivated land, while natural factors strongly affected ecological land and unused land. Multi-scenario simulations for 2036 demonstrate diverging trajectories: an urban development scenario would accelerate cultivated land loss and unused land expansion; a natural development scenario would maintain current pressures; and an ecological protection scenario would effectively curb urban sprawl while actively promoting ecological land recovery. This study concludes that transcending simple land use control to actively orchestrate “mining-urban-rural-ecological” spatial synergy is critical for achieving a sustainable transition in resource-based regions facing similar transformation pressures. Full article

Deoxynivalenol (DON), a Fusarium-derived mycotoxin widely found in grain-based feed, has become a major global environmental contaminant. Reproductive toxicity is one of its most important toxic effects, yet systematic investigations covering both male and female reproductive injury remain limited. This study aimed to establish a combined strategy of network toxicology, molecular docking, molecular dynamics simulation, and single-cell RNA sequencing to evaluate the reproductive toxicity of DON. AKT1, EGFR, PIK3CA, PIK3R1, and SRC were identified as key targets involved in DON-induced reproductive injury. For testicular injury, the prolactin, Ras, HIF-1, and AGE-RAGE signaling pathways were closely associated with DON toxicity. For ovarian injury, the PI3K-Akt, HIF-1, prolactin, insulin, and AGE-RAGE signaling pathways were strongly implicated. Molecular docking demonstrated favorable binding affinities between DON and the hub targets, while molecular dynamics simulation further confirmed the stability of the DON–PIK3CA complex. Single-cell RNA sequencing analysis revealed that these five hub genes were highly expressed in both testicular (SRA667709:SRS3065430) and ovarian (SRA638923:SRS2797100) tissues. These findings deepen current understanding of DON-induced reproductive toxicity, provide new insights into the effects of environmental toxins on reproductive health, and offer a theoretical basis for future studies integrating DON exposure with in vivo validation of core targets and signaling pathways. Full article

Chemoresistance is a major challenge in lung adenocarcinoma (LUAD) treatment and is associated with mitochondrial metabolism. Using publicly available LUAD transcriptome data, we established a five-gene prognostic signature (YWHAZ, HSPD1, NOTCH3, PGK1, and PPARG) for LUAD through differential gene expression profiling, univariate Cox analysis, and machine learning–based feature selection. Patients with LUAD were classified into a high-risk group (HRG) and a low-risk group (LRG) based on their risk scores. Enrichment analysis revealed significant differences between the HRG and LRG in multiple pathways related to metabolism and immunity. The immune microenvironment also differed significantly between the two groups, and the prognostic genes were correlated with infiltrating immune cells. A total of 110 compounds exhibited differential sensitivity across the groups. Molecular docking demonstrated a favorable binding affinity between the prognostic genes and the predicted drugs. Furthermore, YWHAZ knockdown significantly suppressed cancer cell proliferation in cell and animal models. In addition, YWHAZ knockdown markedly reduced cisplatin resistance by downregulating key regulators of the DNA replication and repair pathway, including POLA1 and MCM4. These results provide insight into the molecular mechanisms underlying chemoresistance and identify putative therapeutic targets for LUAD treatment. Full article

Acute myeloid leukemia (AML) is an aggressive hematologic malignancy characterized by profound molecular heterogeneity and high relapse rates, posing significant clinical challenges. Programmed cell death (PCD), encompassing diverse regulated modalities such as apoptosis, necroptosis, and ferroptosis, plays a key role in leukemogenesis and therapeutic response; however, a comprehensive prognostic framework integrating multi-modal PCD pathways in AML remains elusive. In this study, we performed a systematic transcriptomic analysis of 1624 genes associated with 13 distinct PCD forms. A novel computational pipeline combining a variational autoencoder (VAE) for dimensionality reduction and a multilayer perceptron (MLP) for classification was employed to identify robust PCD-related biomarkers, interpreted via SHapley Additive exPlanations (SHAP) analysis. This approach identified 48 candidate genes with discriminative potential between AML and normal bone marrow. Unsupervised consensus clustering based on these genes delineated two molecular subtypes exhibiting divergent clinical outcomes and immune microenvironment profiles. The subtype demonstrated an immunosuppressive phenotype, characterized by enriched regulatory T cells, M2 macrophages, and elevated expression of inhibitory immune checkpoints, correlating with inferior survival. We developed an 8-gene prognostic signature (SORL1, PIK3R5, RIPK3, ELANE, GPX1, VNN1, CD74, and IL3RA) that effectively categorized patients into high- and low-risk groups with notable survival differences, validated across independent cohorts. A prognostic nomogram combining the risk score, age, and cytogenetic risk enhanced the prediction accuracy for overall survival. Our study presents an integrative model that connects multi-modal PCD pathways to AML prognosis, offering a new molecular subtyping system and a clinically applicable risk assessment tool for improved prognostication and personalized treatment strategies. Full article

Background: Extracellular vesicles (EVs) released from skeletal muscle mediate metabolic communication via microRNAs (miRNAs). While both circadian rhythms and exercise influence metabolism, the joint modulation of the muscle-derived EV miRNA landscape by circadian rhythms and chronic exercise remains undefined, particularly under the metabolic stress of obesity. Methods: Employing a 2×2 factorial design (Phase: ZT3 vs. ZT15; Condition: sedentary vs. exercise; ZT, Zeitgeber Time), EV-enriched fractions were isolated from ex vivo quadriceps muscle (QUA) cultures of high-fat diet-fed mice following an 8-week treadmill training regimen using polymer-based precipitation, and comprehensive miRNA profiling was performed by small RNA sequencing. Results: Principal component analysis (PCA) revealed that circadian phase accounted for a greater proportion of global variance in EV miRNA profiles than exercise. Differential expression analysis identified miR-1a-3p and miR-1b-5p as upregulated across both composite phase and exercise contrasts; however, condition-specific analyses indicated that this signal was primarily driven by the sedentary-phase comparison (ZT15-sed vs. ZT3-sed), in which the miR-29 family was also prominently co-upregulated, rather than constituting independent phase and exercise effects; this phase-associated signature was absent in the corresponding exercise-condition comparison. Exploratory functional enrichment of experimentally validated targets revealed phase-preferential association with metabolic and iron–heme pathways, whereas exercise-associated miRNAs mapped to signaling, inflammatory, and transcription-related networks. Conclusions: Circadian phase was the dominant contributor to global variance in muscle-derived EV-enriched miRNA profiles in obesity, as reflected by the phase-associated separation along principal component 1 (PC1, 33.47% of total variance), with exercise introducing context-dependent adaptive modulation. This study provides a foundational basis for investigating the temporal regulation of muscle secretome dynamics under high-fat diet conditions, highlighting temporal specificity as a key dimension in EV-mediated exercise physiology research. Full article

Bothriochloa ischaemum is a key forage species with strong grazing tolerance and high nutritional value, making precise quantification of spike and seed traits essential for germplasm evaluation and yield prediction. However, the compact architecture and minute seed size in natural field conditions render manual counting inefficient and labor-intensive. To address this limitation, this study presents a non-destructive and automated quantification framework integrating advanced object detection and regression analysis for accurate in situ estimation of spikes and seed numbers. To further address the challenges of dense spike detection caused by occlusion and small object sizes, this study developed a modified model named YOLOv12-DAN by integrating DySample dynamic upsampling, ASFF feature fusion, and NWD loss, which achieved a mean average precision (mAP) of 91.6%. Meanwhile, for the detection of dense kernels on compact spikes, an improved YOLOv12 architecture incorporating an Explicit Visual Center (EVC) module was proposed to enhance multi-scale feature representation. The optimized model attained a bounding box precision of 96.5%, a recall rate of 86.4%, an mAP50 of 94.3%, and an mAP50-95 of 73.9%. Furthermore, a univariate linear regression model based on 132 spike samples verified the reliable consistency between the predicted and actual seed counts, with a mean absolute error (MAE) of 6.30, a mean absolute percentage error (MAPE) of 9.35, and an R-squared (R²) value of 0.808. Finally, the model was deployed through a lightweight end-to-end web application, enabling real-time field operation and promoting its applicability in breeding programs and agronomic decision-making. This study provides a robust technical pathway for automated phenotyping and precision forage improvement. Full article