搜索结果：Exploit

In a dual-beam super-resolution laser direct-writing lithography system, the flatness error of the motion stage during XY-plane scanning (small fluctuations along the Z direction) is a critical factor that degrades writing accuracy. This work presents an integrated method combining dual-cylindrical-lens astigmatic measurement and machine-learning-based feedforward compensation. By employing the dual-cylindrical-lens astigmatic approach together with a four-quadrant detector (FQD) to capture variations of the spot profile, nanometer-scale Z-axis errors of the stage are converted into changes in the focus error signal (FES), enabling precise detection of flatness errors. To exploit the predictability of the error, this study further introduces random forest regression from ensemble learning. A prediction model is constructed through analysis of historical data to accurately estimate the spatial distribution of the stage flatness error, and the predicted error is then used to drive a piezoelectric actuator for feedforward compensation. Experimental results demonstrate that the proposed random-forest-based feedforward strategy markedly reduces flatness errors generated during planar motion, thereby improving both the writing accuracy and operational stability of the dual-beam super-resolution laser direct-writing lithography system.

Molecular dynamics is a key technique for exploring biomolecular systems at the atomic level. The rapid growth in accessible system sizes and time scales has intensified the need for efficient postprocessing methods that extract meaningful insights from the resulting data. Interaction fingerprint (IFP) analyses are a valuable tool for elucidating key atomic interactions within molecular ensembles; yet current specialized software often struggle with extensive trajectories or complex systems. Here, we introduce InterMap, a Python package designed to accelerate IFP detection on large-scale molecular ensembles. By actively exploiting k-d trees, InterMap efficiently handles the massive amount of distance calculations necessary to detect IFPs, particularly when dealing with intramolecular interactions. The seamless integration with MDAnalysis ensures broad format compatibility and allows using SMARTS patterns for flexible interaction definitions. InterMap adopts a deeply compressed binary encoding to manage IFPs, which makes it very memory-friendly. Furthermore, convenient interactive visualizations are provided to enhance data interpretation through a locally hosted web-browser application. Benchmark results indicate that InterMap significantly outperforms existing tools for processing complex biomolecular systems, achieving up to a 99% reduction in both runtime and peak memory usage. InterMap's code and issue tracker are available at https://github.com/rglez/intermap, while documentation and tutorials can be found at https://rglez.github.io/intermap/.

Thioester chemistry is exploited in Nature by many CoA-dependent enzymes. However, the covalent nature of CoA attachment largely prevents the use of these enzymes in many applications. Replacing the CoA moiety with simpler, truncated fragments, such as its pantetheine (PAN) moiety, is also hampered by the lack of understanding of the function of the CoA moiety in enzymatic conversions. Herein, we describe the utilization of the enzyme (2E)-enoyl-CoA hydratase (ECH) using PAN thioesters and an activator, 3',5'-ADP (PAP). ECH catalyzes the hydration of the carbon-carbon double bond of (2E)-enoyl-CoA substrates in the &#x3b2;-oxidation lipid-degrading pathway. The hydration reaction is very challenging to carry out by traditional chemical synthesis, as no selective catalysts are available. Structural enzymology of ECH and its complexes with (3S)-hydroxyacyl-CoA products show that hydrogen bonds between the adenine 6-amino group of the ADP moiety of CoA and loop-2 induce a small structural change in this active site loop, tightening the NN distance between the hydrogen bond donors of the oxyanion hole from 5.2 &#xc5; (unliganded) to 4.0 &#xc5; and forming a competent oxyanion hole at the catalytic site. A structurally similar and catalytically competent oxyanion hole is observed in the complex with (3S)-hydroxyhexanoyl PAN and the activator 3',5'-ADP, both bound at the active site. The use of 3',5'-ADP as the activator enables the synthetic use of ECH for the hydration of a wide range of (2E)-enoyl-PAN substrates with different steric demands and functionalities. The products, 3-hydroxyacyl-PAN thioesters, were obtained in good isolated yields and excellent stereoselectivities (typically >99:<1 3S:3R). Even for acyl chains that contain reactive groups such as bromide or methyl ester functionalities at C7, no side products resulting from potentially competing cyclization could be detected in the enzymatic hydration protocol.

Advanced chronic liver disease (ACLD) with cirrhosis is increasingly recognised as a condition shaped by the 'oral-gut-liver axis', in which dysbiosis within the oral microbiome contributes to systemic inflammation, infection, decompensation, and acute-on-chronic liver failure. Periodontal disease is highly prevalent in ACLD and is associated with endotoxaemia, immune dysfunction, and hepatic complications. The protected dental biofilm and keystone pathogens are key to the development of local and systemic inflammatory processes. The concept of "oralisation" of the gut microbiome further links oral dysbiosis to microbial translocation and hepatic injury. Recent advances in multi-omics, resistome profiling, and spatially resolved imaging have deepened insights into community function and host-microbial crosstalk, while salivary biomarker panels and microbial signatures across different aetiologies suggest potential tools for non-invasive diagnosis and risk stratification. Clinical priorities now lie along two paths which complement each other. The first is immediate implementation: embedding routine periodontal assessment and professional plaque removal within hepatology care; consistent advice on oral hygiene, fluoride use, diet, and smoking and alcohol cessation; careful review of proton-pump inhibitor use; and much closer coordination between hepatologists and dentists to facilitate indicated procedures. The second is innovation: development of precision microbiome-based interventional trials powered for hepatic outcomes, including targeted probiotics and postbiotics, biofilm-disrupting and quorum-quenching strategies, and phage or narrow-spectrum antimicrobial therapies supported by rapid diagnostics and robust antimicrobial stewardship. Integrating oral health into hepatology practice may represent a practical opportunity to reduce infection risk, delay decompensation, and improve survival and quality of life in people living with ACLD. This review aims to synthesise concepts around current understanding of the patho-biological mechanisms, analytical innovations, and therapeutic opportunities that define this evolving connection, as well as identify gaps in the knowledge base and propose avenues to harness and exploit the oral-gut-liver axis.

Lentil (Lens culinaris Medik.), a nutrient-rich legume cultivated worldwide, plays a vital role in combating malnutrition and hidden hunger. Understanding the genetic architecture underlying key phenological and agronomic traits in lentil is crucial for accelerating molecular breeding. In this study, genome-wide association mapping was conducted using 142 genetically diverse lentil accessions, evaluated across two field environments over two years. High-throughput sequencing generated 34,995 high-quality single-nucleotide polymorphisms, which were used for genetic characterization and for the identification of marker-trait associations for phenological and yield-associated traits. Population structure analysis identified three subpopulations (K&#x2009;=&#x2009;3), with UPGMA clustering showing a similar pattern. Association mapping was performed using multi-locus models and further confirmed through a single-locus generalized linear model. A total of 64 significant associations were identified, of which Chr5_342836807 and Chr6_200603138 were consistently detected across all environments for days to 50% flowering. Putative candidate genes located near these phenology-associated loci such as abscisate &#x3b2;-glucosyltransferase, pentatricopeptide repeat proteins, and transcription factors from the MYB, MADS-box, and GRAS families are likely involved in flowering-time regulation in lentil. These findings reveal novel associations between genetic variants and complex traits and identify putative genes that may be exploited in marker-assisted selection and genomic prediction strategies. The online version contains supplementary material available at 10.1007/s12298-026-01739-x.

Photothermal therapy (PTT) holds transformative potential for precision cancer treatment, yet clinical translation remains constrained by the scarcity of molecularly defined, biocompatible, and efficiently NIR-absorbing photothermal agents (PTAs). Here we report a rational donor-acceptor-donor (D-A-D) framework that delivers ultrasmall organic PTAs with record photothermal conversion efficiencies (49.8%) and intrinsic immunogenic cell death (ICD) activity. The design exploits 6,7-diphenyl-[1,2,5]thiadiazolo[3,4-g]quinoxaline as a &#x3c0;-extended, multi-nitrogenated acceptor core flanked by trifluoromethyl groups to deepen the LUMO, while methoxylated triphenylamine donors intensify intramolecular charge-transfer and suppress radiative decay. Nanoprecipitation furnishes monodisperse nanoparticles that exhibit intense NIR-II absorption, exceptional photostability across five hyperthermic cycles, and lysosome-directed uptake. In vitro, single-dose FTPA NPs plus 808-nm laser irradiation trigger mitochondrial depolarization, G0/G1 arrest, and apoptosis in > 70% of 4T1 cells while releasing abundant ATP and surface calreticulin-canonical ICD signals. A prophylactic vaccination model corroborates these molecular cues: mice primed with FTPA-NP-treated tumor cells reject contralateral challenge, achieving > 90% long-term survival, expansion of cytotoxic CD8+ T cells (&#x2248; 70% activation), and suppression of Tregs (&#x2248; 3%). No systemic toxicity or off-target pathology is observed. This study establishes a chemically tunable, metal-free PTA platform that synergizes thermal ablation with systemic anti-tumor immunity, providing a versatile scaffold for next-generation precision immuno-photothermal medicine.

A traditional view of selective attention distinguishes between goal-directed and stimulus-driven mechanisms of attentional control. More recently, a large (and growing) body of research has identified a third class of control system-termed selection history-wherein attentional prioritisation is shaped by our prior experience with stimuli, independently of our goals and the physical salience of those stimuli. This article reviews work within this selection history literature demonstrating that prioritisation is rapidly and automatically modulated by learning about the rewards associated with stimuli, and argues for a framework that distinguishes between history-driven processes implementing attentional exploitation (the drive to leverage reliable information) and attentional exploration (the drive to resolve uncertainty, with the aim of validating potential new sources of information). Findings such as these highlight a fundamental and intricate interaction between learning and attention, wherein our prior experience shapes the way in which we extract information from our environment - with potential consequences for understanding the subsequent decisions that we make and choices that we take.

Quadrupolar nuclei with half-integer spin, which represent 66&#xa0;% of the NMR-active isotopes, are present in a wide range of materials with applications in various fields, including heterogeneous catalysis, optoelectronics and energy. The solid-state NMR spectra of these isotopes are affected by quadrupolar interactions, which provide unique information on the local environment of these nuclei, in addition to their chemical shifts. These anisotropic interactions, which are generally larger than other internal spin interactions, split and broaden the NMR transitions, which reduce the sensitivity for the detection of these isotopes. In addition, the large dimensions of their density matrices and the numerous NMR transitions complicate the spin dynamics and can reduce the efficiency of coherence transfers, such as cross-polarization under magic-angle spinning (CPMAS), which is widely employed to boost the sensitivity for the detection of spin-1/2 isotopes. In the last decade, sensitivity gains provided by dynamic nuclear polarization (DNP) have been exploited to detect half-integer quadrupolar nuclei in solids. This review discusses the advantages and limitations of the different DNP-NMR techniques that have been proposed for the detection of these isotopes, including direct excitation and CPMAS, and two more recently introduced methods called PRESTO (Phase-shifted Recoupling Effects by Smooth Transfer of Order) and D-RINEPT (Dipolar-mediated Refocusing Insensitive Nuclei Enhanced by Polarization Transfer). We also show how these techniques can be applied to obtain new insights on the structure of materials, notably of their surfaces, and hence, contribute to extend the range of applications of the surface-enhanced NMR spectroscopy (DNP-SENS).

Two-dimensional (2D) inorganic materials provide a powerful platform for electronic-structure engineering through precise control of the composition and crystal structure. While cation substitution has been widely exploited in oxide nanosheets, anion engineering remains far less developed, particularly in molecularly thin oxynitride systems with controlled nitrogen doping. Here, we report a generalizable route to nitrogen-doped perovskite oxide nanosheets that overcomes long-standing challenges associated with nitridation and structural instability. Using Dion-Jacobson (DJ)-type perovskite oxynitrides, RbSr2(Nb1-xTax)3O10-yNy, as a model platform, we demonstrate that the combination of cation substitution and nitrogen doping enables systematic modulation of both composition and electronic band structure in 2D perovskites. DJ-type perovskite oxynitrides with substantial nitrogen incorporation can be obtained via an unexpected transformation from pseudo-Ruddlesden-Popper-type phases, induced by alkali metal salt-assisted nitridation followed by simple aqueous treatment, without altering the anion composition. These oxynitrides are subsequently exfoliated into single-layer nanosheets that preserve the perovskite framework and the designed cation stoichiometry. Direct determination of both valence and conduction band edges by combined ultraviolet and inverse photoelectron spectroscopy reveals composition-dependent, nonmonotonic band alignment behavior that cannot be resolved by indirect optical or electrochemical approaches. This work establishes an integrated materials and characterization framework for the rational electronic-structure design in 2D oxynitride nanosheets.

With the continuous growth of information retrieval and knowledge acquisition demands, intelligent question-answering systems have been widely adopted across various vertical domains. However, in the context of audit result announcement analysis, domain-specific models tailored to the characteristics of audit texts remain scarce, rendering a large amount of structured and semantic information difficult to exploit effectively. To address this limitation, this paper proposes a dual-channel retrieval-augmented generation framework for audit result announcements, termed Dual-Channel Retrieval-Augmented Generation. The framework establishes a dual-path retrieval mechanism over a document database and a relational database, respectively retrieving candidate evidence from unstructured audit texts and structured knowledge graphs, and integrates multi-source evidence through a unified evaluation and re-ranking strategy, thereby enhancing the system's capability in audit semantic understanding and reasoning. Building upon this design, an end-to-end intelligent question-answering model is constructed by incorporating modules for query understanding, evidence re-ranking, and response generation, enabling knowledge retrieval and answer generation driven collaboratively by multiple data sources. Experimental results demonstrate that the proposed model consistently outperforms baseline methods in both retrieval effectiveness and answer quality, accurately interpreting user queries and producing high-quality responses, thus providing a feasible solution for the intelligent application of audit result announcements.

Programmable photonic networks carry out universal unitary functions by independently operating on the amplitude and phase of guided light. Exploiting the reconfigurability and spatiospectral degrees of freedom of these systems, the majority of state-of-the-art photonics applications, ranging from microwave photonics to photonic computing and optical communication links, can be demonstrated in one unified system. Existing techniques require a large footprint due to weak modulation efficiency, and continuous power dissipation to maintain the configured state. Here, we demonstrate a programmable recirculating mesh unit cell based on the nonvolatile low-loss phase-change material Sb2Se3. The demonstrated devices achieve an ultrashort active length (<10 &#x3bc;m, more than 15 times smaller than the current state of the art of competing technologies) and zero static power, in combination with high-extinction switching (>20 dB), broadband operation (>15 nm), and low insertion loss (<2 dB). This work forms the basis for nonvolatile field-programmable coupler arrays (nv-FPCAs) and zero-static power reconfigurable optical interconnects.

Autophagy plays a dual role in cancer progression, and strategies to drive excessive autophagic flux remain a promising yet challenging therapeutic avenue. Herein, we develop an endoplasmic reticulum (ER)-targeted self-assembling peptide system (P-1-ERT@Rap) that enables localized photodynamic damage and robust ER stress, which synergizes with rapamycin (Rap) for inducing dual autophagy activation in cancer cells. The peptide P-1-ERT co-assembles with Rap into well-dispersed micelles, which exhibit pH-responsive morphological transformation from nanoparticles to nanofibers under acidic conditions, thereby facilitating lysosomal escape and cellular release of therapeutics. Importantly, P-1-ERT selectively accumulates in the ER and generates reactive oxygen species under laser exposure, triggering significant ER stress with upregulation of CHOP proteins. Concurrently, cellular delivery of Rap, an autophagy inducer, further amplifies autophagic flux with increasing LC3B-II/I ratios, ultimately promoting programmed cell death in A375 cells. Notably, the P-1-ERT@Rap system achieves higher tumor accumulation compared to free photosensitizer in vivo. Moreover, intravenous administration of P-1-ERT@Rap alongside laser irradiation significantly inhibits tumor growth in an A375-xenografted mouse model, with minimal systemic toxicity observed. This dual modulation strategy for autophagy regulation effectively enhances photodynamic therapy efficacy, and it offers a promising approach for exploiting organelle specific stress pathways in cancer treatment. STATEMENT OF SIGNIFICANCE: This study presents a endoplasmic reticulum (ER)-targeted self-assembling peptide nanoplatform (P&#x2011;1&#x2011;ERT@Rap) that integrates organelle-specific photodynamic therapy (PDT) with autophagy modulation for synergistic cancer treatment. The system uniquely exploits pH-responsive morphological transformation from micelles to nanofibers in acidic tumor environments, facilitating lysosomal escape and efficient intracellular delivery. By selectively accumulating in the ER and generating localized reactive oxygen species upon irradiation, it induces severe ER stress and upregulates CHOP, while co-delivered rapamycin further amplifies autophagic flux. This dual activation of autophagy leads to enhanced programmed cell death in melanoma cells, as demonstrated both in vitro and in vivo. Our work provides a pioneering strategy for organelle-precise therapy that leverages dual stress pathways to overcome the limitations of conventional monotherapies. We believe that this new approach has the potential to revolutionize the field of precision oncology, setting a new paradigm for significantly enhancing treatment outcomes across a broad spectrum of tumor types.

Chronic inflammation perturbs hematopoietic homeostasis, promoting aberrant myelopoiesis and clonal expansion of mutated stem cells. Here, we develop a mathematical model that integrates both local (bone marrow-intrinsic) and global (systemic/peripheral) inflammation-driven feedback mechanisms to investigate their roles in hematopoietic regulation and disease progression. Our model captures the nonlinear interplay between self-renewal, progenitor proliferation, and inflammatory cues, enabling classification of healthy, myelodysplastic, and leukemic states based on stem cell population dynamics. We show that global inflammatory feedback enhances the resilience of hematopoiesis, while excessive feedback on progenitor cells under chronic inflammation drives instability and clonal dominance. Using sensitivity analysis and parameter space mapping, we identify critical feedback thresholds governing transitions between hematopoietic states and reveal how mutated clones exploit inflammation to outcompete wild-type cells. This systems-level framework offers mechanistic insights into the emergence of myeloid malignancies and provides a computational platform for exploring potential anti-inflammatory therapeutic strategies.

Ubiquitination, a central post-translational mechanism, shapes the amplitude and duration of cellular signalling. Josephin domain-containing 2 (JOSD2), a Machado-Joseph disease (MJD) family deubiquitinase, eliminates ubiquitin moieties from ubiquitin-conjugated substrates and tunes proteostasis and signalling outputs. Emerging evidence links aberrant JOSD2 activity to diverse pathological states. This review, aims to summarize the current data regarding of JOSD2 as a regulatory node in ubiquitin-dependent signalling and discuss the role of its dysregulation in malignancies through interconnected mechanisms, including metabolic rewiring, rewiring of oncogenic signalling circuits, and altered therapeutic responses that promote resistance. Furthermore, the context-dependent roles of JOSD2 beyond cancer emphasized, with reported pathogenic or protective functions in cardiovascular disorders and inflammatory bowel disease. The literature highlights JOSD2 as a signalling-relevant deubiquitinase with pleiotropic, context-dependent functions. This review discusses key knowledge gaps-such as incomplete substrate mapping and determinants of tissue specificity-and outlines translational opportunities and challenges for exploiting JOSD2 as a biomarker and therapeutic target.

Water losses caused by hidden leaks in water distribution networks (WDNs) remain a persistent challenge for utilities, motivating reliable leak monitoring under realistic field constraints. However, acoustic leak detection (ALD) in practice is often limited by scarce labeled data, non-stationary interference, and the need for risk-aware fine-grained recognition beyond a simple leak/no-leak decision. This study develops a few-shot ALD framework that improves fine-grained performance while reducing high-risk cross-parent errors. To this end, a unified benchmark is established to evaluate diverse signal representations and model backbones on a real in-service dataset with a two-level label hierarchy (parent: leak and no-leak; child: four imbalanced subclasses). Hierarchical consistency learning (HCL) is proposed to enforce coherence between parent- and child-level predictions, and class-wise gating fusion (CWGF) is developed to adaptively exploit expert complementarity within a shared label space. Experiments show that HCL improves Macro-F1 by 5.11 percentage points and Parent-F1 by 2.12 percentage points on average over baseline. CWGF further yields a reliably positive gain of 1.93 percentage points over the stronger expert, with benefits becoming more pronounced as the training fraction decreases. Robustness and data-efficiency evaluations across group-aware splits and varying training fractions demonstrate favorable performance under few-shot regimes. This study supports a practical utility-oriented approach for data-scarce leak monitoring that improves fine-grained accuracy while enhancing operational risk control.

Iron oxide nanoparticles (IONPs) have proven to be of therapeutic potential against cancer. The feature of the surface coating can affect important properties of IONPs; it is therefore critical for further understanding how these materials react to physiological conditions, which is still needed to fully exploit the potential of IONPs for their theranostic applications. In this study, we explored the therapeutic potential of rutin and nisin conjugated IONPs as anticancer agents. One important hallmark of many cancers is the overexpression of the endoplasmic reticulum-resident chaperone, GRP78, and its translocation to many cellular compartments, including the cell membrane. We explored the potential binding affinity of rutin and nisin against the substrate-binding domain &#x3b2; (SBD&#x3b2;) of GRP78. The results show promising results for both nisin and rutin, with more enhanced binding capability of the former due to its extended structure (peptide in nature), forming more non-bonded interactions with the GRP78 surface. Our findings pave the way for the use of these coating agents against the cell-exposed chaperone, GRP78, to alleviate its chemoresistance characteristics in cancer.

Androgen receptor (AR) signaling is central to prostate cancer progression, yet resistance to AR-targeted therapies remains a major clinical challenge. Understanding the molecular consequences of AR pathway inhibition is therefore essential for improving therapeutic outcomes. Here, we identify a previously unrecognized link between AR antagonism and cuproptosis, a copper-dependent form of regulated cell death. Using integrated genomic profiling, we find that AR-targeted agents transcriptionally activate the key cuproptosis regulator Ferredoxin-1 (FDX1), thereby rendering prostate cancer cells markedly more susceptible to copper-induced lethality. Mechanistically, ligand-bound AR directly engages FDX1 cis-regulatory elements, which are rendered accessible by the pioneer factor GATA2, and drives FDX1 upregulation upon AR antagonist exposure. Consistent with this mechanism, FDX1 expression is elevated in clinical prostate cancer samples following androgen deprivation therapy or AR antagonist treatment. Increased FDX1 enhances intracellular Cu+ accumulation, destabilizes Fe-S cluster proteins, and disrupts mitochondrial metabolism, establishing a procuproptotic state. Functionally, combining AR antagonists with copper ionophores synergistically induces cuproptosis and potently suppresses tumor growth in AR-positive prostate cancer cells, three-dimensional (3D) spheroids, patient-derived organoids, and xenograft models, with minimal systemic toxicity. This synergy is abolished by FDX1 loss or copper chelation, confirming dependence on AR-FDX1 axis activation. Together, these findings uncover FDX1 as a mechanistic effector of AR pathway inhibition and propose a well-tolerated combination strategy that exploits cuproptosis to improve therapeutic responses in prostate cancer.

The management of acid gases (i.e., H2S and CO2) is a fundamental requirement in process plants, such as refineries. Current strategies typically do not exploit the hydrogen content in hydrogen sulfide, which is usually burned. The acid gas-to-syngas (AG2S) technology represents an innovative approach for converting a mixture of hydrogen sulfide (H2S) and carbon dioxide (CO2) into syngas (H2 and CO), with potential applications in fuels and chemical synthesis. This work investigates the optimal process design of AG2S in terms of the H2S/CO2 and H2S/O2 feed molar ratios to maximize syngas production. The study combines detailed kinetic and thermodynamic modeling to obtain a comprehensive process simulation, which serves as the basis for generating accurate surrogate models trained on flowsheet simulation data via a design of experiments (DoE) approach. These models allow for a reliable prediction of H2S conversion, syngas flow rate, H2/CO ratio, and selectivity. The results highlight limitations imposed by relatively low H2/CO ratios for downstream applications and illustrate the trade-off between syngas quality and quantity.

Canavalia gladiata (C. gladiata) is an important medicinal and edible plant. However, systematic research on the distribution of metabolites in different tissues of C. gladiata and their potential transcriptional regulatory mechanisms remain poorly understood. We performed integrated metabolomic and transcriptomic analyses across five tissues (roots, stems, leaves, seeds, and fruit pericarps) of C. gladiata, combined with antioxidant capacity and bioactive component content assays, to dissect the regulatory networks of flavonoid and terpenoid biosynthesis. Seeds exhibited the highest antioxidant activity and total phenolic content, whereas leaves accumulated the highest levels of total flavonoids and terpenic acids. A total of 4,405 DAMs and 25,597 DEGs were identified, revealing pronounced tissue-specific metabolic and transcriptional divergence. Flavonoid and terpenoid biosynthesis pathways were significantly enriched in comparisons between seeds and other tissues. Key structural genes, including 4CL, CHS, FLS, HMGR, and DXS, displayed strong tissue-specific expression patterns. Co-expression network analysis identified candidate regulatory modules, highlighting MYB, bHLH, and MYC2 transcription factors as central regulators of flavonoid and terpenoid metabolism in seeds and fruit pericarps. This study provides the first comprehensive landscape of tissue-specific flavonoid and terpenoid metabolism in C. gladiata, offering a theoretical foundation and valuable genetic resources for the targeted exploitation of its bioactive components and molecular breeding.