Search Results (38,660)

Graph Neural Networks (GNNs) have become a central methodology for modelling biological systems where entities and their interactions form inherently non-Euclidean structures. From protein interaction networks and gene regulatory circuits to molecular graphs and multi-omics integration, the relational nature of biological data makes GNNs particularly well-suited for capturing complex dependencies that traditional deep learning methods fail to represent. Despite their rapid adoption, the effectiveness of GNNs in bioinformatics depends not only on model design but also on how biological graphs are constructed, parameterised and trained. In this review, we provide a structured framework for understanding and applying GNNs in bioinformatics, organised around three key dimensions: (1) graph construction and representation, including strategies for deriving biological networks from heterogeneous sources and selecting biologically meaningful node and edge features; (2) GNN architectures, covering spectral and spatial formulations, representative models such as Graph Convolutional Networks (GCNs), Graph Attention Networks (GATs), Graph Sample and AggregatE (GraphSAGE) and Graph Isomorphism Network (GIN), and recent advances including transformer-based and self-supervised paradigms; and (3) applications in biomedical domains, spanning disease–gene association prediction, drug discovery, protein structure and function analysis, multi-omics integration and biomedical knowledge graphs. We further examine training considerations, including optimisation techniques, regularisation strategies and challenges posed by data sparsity and noise in biological settings. By synthesising methodological foundations with domain-specific applications, this review clarifies how graph quality, architectural choice and training dynamics jointly influence model performance. We also highlight emerging challenges such as modelling temporal biological processes, improving interpretability, and enabling robust multimodal fusion that will shape the next generation of GNNs in computational biology. Full article

This study explores entities’ strategy to combine multiple governmental funding sources to complement their research and innovation activities. It focuses on the case of Spanish aeronautics entities and their participation in both European and national research and innovation programs over two successive periods. Using a mixed-methods approach, this study combines qualitative interviews with three representative entities and a quantitative analysis of project-level data. The interviews are first used to identify key variables and analytical categories, such as entity type, Joint Undertaking membership, technological focus, and temporal evolution, which then guide the quantitative analysis. Quantitative data on budgets, funding, participation, and technologies are analyzed across both periods and programs, including pairwise correlation analysis. The findings show that Spanish entities actively seek to align national and European funding at both financial and technological levels, although with uneven success in some cases. Joint Undertaking membership and position within the aeronautical value chain strongly influence the ability to participate in both programs and to accumulate funding. While many entities develop informal alignment strategies, these efforts often exceed their organizational capacity, particularly in the second period. The results highlight the need for formal, government-level coordination mechanisms to support effective alignment between European and national aeronautics funding programs. Full article

Mangrove ecosystems are important blue carbon systems and play a critical role in understanding carbon cycling and responses to climate change. However, accurate regional estimation of Net Ecosystem Exchange (NEE) remains challenging due to the environmental complexity and spatial heterogeneity. This study combined eddy covariance observations from four mangrove sites along China’s southeastern coast (natural and restored mangrove forests) with multi-source remote sensing and environmental reanalysis data to construct three variable schemes (site observations only, with added vegetation indices, and comprehensive multi-source variables). We compared three machine learning models for daily NEE prediction, including eXtreme Gradient Boosting (XGBoost), Random Forest (RF), and Support Vector Machine (SVM). The results showed that: (1) Restored and natural mangroves exhibited similar temporal NEE dynamics and consistently functioned as carbon sinks, restored mangrove sites showed greater cross-site variability. Among the study sites, CN-LZR exhibited the strongest cumulative carbon uptake. (2) Scheme 3 combined with the XGBoost algorithm achieved the highest predictive accuracy, reaching an R² of 0.73 across sites. Differences among machine learning models were primarily associated with their ability to capture nonlinear interactions between atmospheric and hydrological variables, with tree-based models outperforming SVM. (3) SHAP analysis indicated that radiation-related variables were the dominant drivers of NEE, while hydrological influences were site-dependent; and (4) Regional upscaling indicated that all sites consistently functioned as long-term carbon sinks, with CN-LZR exhibiting slightly higher daily mean carbon uptake than the other sites. This study presented the first machine learning framework for estimating daily-scale NEE in mangroves, providing methodological and data support for regional carbon flux assessment and blue carbon management. Full article

The deployment of Discrete Variable Quantum Key Distribution (DV-QKD) in high-traffic, short-reach environments, such as intra-data center networks, is currently constrained by the saturation of single-photon detectors. While CMOS Single-Photon Avalanche Diodes (SPADs) offer a cost-effective solution, their Secure Key Rate (SKR) is limited by detector dead time. To the best of the authors’ knowledge, this work is the first to derive a generalized detection rate model for SiPMs that addresses the dead-time bottlenecks of gigahertz-rate quantum cryptography. While methods for managing deadtime via active optical switching have been proposed, our model quantifies the benefits of passive spatial multiplexing inherent in standard SiPM arrays. Furthermore, contrasting with models designed to optimize energy resolution or characterize nonlinear charge response to light pulses, our work focuses on maximizing the detection count rate. We derive exact detection rate models for both analog (paralyzable) and digital (non-paralyzable) SiPM architectures, incorporating correlated noise sources such as optical crosstalk and afterpulsing. Simulation results indicate that SiPMs can increase detection rates by over an order of magnitude compared to single SPADs. Full article

High-resolution marine gravity fields are critical for interpreting seafloor structure, investigating marine geodynamics, and enabling gravity-aided navigation. However, sparse shipborne observations, heterogeneous multi-source geodetic datasets, and the inability of conventional methods to handle nonlinear inversion limit accurate gravity recovery. To overcome these limitations, we propose a spectral physics-informed constraint deep-learning framework based on a multi-channel Swin Transformer to reconstruct high-resolution marine gravity anomaly fields. The model ingests multi-source geodetic inputs organized as 64 × 64 grid patches centered near each computation point and fuses them to predict the target gravity anomaly. We adopt a remove–compute–restore (RCR) strategy that isolates residual gravity signals, which improves numerical stability and accelerates training. Inputs include satellite-altimetry-derived vertical gravity gradients, vertical deflections, mean sea surface height, and topography; the model is trained on over 430,000 shipborne gravity samples from the South China Sea (0–30° N, 105–125° E). To enforce physical consistency, we embed a spectral-domain physics constraint derived from potential-field theory into the loss function; this constraint helps recover short-wavelength gravity signals. We also introduce an adaptive multi-domain multi-scale feature fusion module (AMAMFF) to improve the integration of heterogeneous inputs, and we demonstrate its benefits in experiments across complex terrain. Validation against independent shipborne gravity checkpoints yields an RMS error of 3.09 mGal, indicating a substantial performance advantage over existing deep-learning approaches and conventional gravity-field models. Full article

Coverage expansions and affordability reforms often presume that improved access to care yields better population health. We examine this premise in a capacity-constrained healthcare system, where congestion and throughput determine whether potential access translates into realized care. Using U.S. state-year panel data from 2006 to 2023, we study (i) how healthcare workforce density relates to multiple access margins and (ii) whether the mortality effects of access improvements depend on local delivery capacity. Reduced-form estimates show that higher workforce density is associated with higher insurance coverage and fewer cost-related barriers to care, while associations with having a usual source of care are weaker. With full controls these relationships attenuate, and Medicaid expansion and poverty explain much of the remaining variation. Instrumental variable models suggest that policy-driven improvements in effective access are associated with lower mortality, although the first-stage strength varies across specifications. Interaction-IV estimates indicate capacity dependence: for all-cause and external-cause mortality, implied benefits are larger in lower-capacity settings and diminish as workforce density increases; for endocrine mortality, benefits are concentrated in higher-capacity settings, while respiratory effects are not detectable. Overall, the results support a systems perspective in which the health returns to access expansions depend on local delivery capacity, underscoring the importance of aligning access reforms with constraints in healthcare production and flow. Full article

Regional epidemic risk prediction requires holistic modeling of heterogeneous data sources such as demographic structure, health capacity, geographical features and human mobility. In this study, a unique and multi-modal epidemiological data set integrating demographic, health, geographic and mobility indicators of Türkiye and its neighboring countries was collected. Türkiye’s neighboring countries are Greece, Bulgaria, Georgia, Armenia, Iran, and Iraq. This dataset, created by combining raw data from these neighboring countries, provides a comprehensive regional representation that allows for both quantitative classification and spatial mapping of epidemiological risk. To address the class imbalance problem, Conditional GAN (CGAN), a class-conditional synthetic example generation approach that enhances high-risk category representation was used. In this study, we proposed a multi-view deep learning model named MV-RiskNet, which effectively models the multi-dimensional data structure by processing each view into independent subnetworks and integrating the representations with an attention-based fusion mechanism for regional epidemic risk prediction. Experimental studies were compared using Multi-Layer Perceptron (MLP), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), Autoencoder classifier, and Graph Convolutional Network (GCN) models. The proposed MV-RiskNet with CGAN model achieved better results compared to other models, with 97.22% accuracy and 97.40% F1-score. The generated risk maps reveal regional clustering patterns in a spatially consistent manner, while attention analyses show that demographic and geographic features are the dominant determinants, while mobility plays a complementary role, especially in high-risk regions. Full article

The increasing integration of electric vehicles (EVs) and renewable energy sources has accelerated the adoption of DC microgrids, where maintaining voltage stability and effective power sharing remains a critical challenge. Hybrid energy storage systems (HESS), combining batteries and supercapacitors, are commonly employed to address dynamic power variations. However, conventional proportional–integral (PI)-based control strategies for HESS can exhibit performance limitations under nonlinear and varying operating conditions. To overcome this drawback, this paper presents an adaptive neuro-fuzzy inference system (ANFIS)-based control strategy for HESS located in a DC microgrid, with comparative evaluation against both conventional PI and traditional Fuzzy Logic controller (FLC) schemes. The proposed approach is evaluated using a detailed MATLAB/Simulink R2024a model of a DC microgrid including EVs. Simulation results show that, under normal operating conditions, the ANFIS-based control demonstrates improved transient response, reduced voltage fluctuations, and effective coordination between the battery and supercapacitor during renewable power variations, compared to PI and FLC-controlled systems. In addition to nominal performance assessment, this work investigates the vulnerability of the ANFIS controller to cyber-attacks. Two representative attack scenarios, false data injection (FDI) and denial-of-service (DoS), are applied to critical measurement and control signals of HESS. Simulation results reveal that, although the DC-bus voltage regulation is largely maintained during attack intervals, cyber manipulation significantly disrupts the intended HESS power-sharing behavior. Full article

Background: Saliva has been identified as a valuable diagnostic biofluid due to its non-invasive collection and its capacity to reflect oral and systemic biological processes. Advances in analytical chemistry, biosensing technologies, and artificial intelligence (AI)-assisted data integration have broadened the applications of salivary diagnostics. Among salivary exposome components, heavy metals such as lead, cadmium, mercury, nickel, chromium, arsenic, and aluminum serve as biologically and clinically relevant indicators of environmental exposure, toxic burden, and disease-associated molecular disorders. Methods: This structured review integrates clinical, experimental, and translational studies published between January 2020 and January 2026 that examined salivary heavy metal profiling in relation to oral health. Evidence was identified using systematic searches of PubMed/MEDLINE and supplementary sources. Studies were qualitatively assessed regarding analytical methodologies, reported concentration ranges, biological mechanisms, disease associations, and the development of digital and AI-assisted diagnostic applications. Results: Thirteen human clinical studies and six animal or in vivo investigations met the inclusion criteria. Across these studies, altered salivary metal profiles were linked to oxidative stress, inflammatory signaling, immune dysregulation, microbiome disturbances, and genotoxic markers relevant to periodontal disease, oral mucosal pathology, and the risk of oral squamous cell carcinoma. Inductively coupled plasma mass spectrometry was the predominant analytical platform, while emerging biosensor technologies showed potential for rapid detection and monitoring. Digital and AI-based approaches were identified as promising tools for integrating metallomic data with clinical and molecular biomarkers to support exposure-informed risk stratification. Conclusions: Salivary heavy metal profiling represents a biologically informative, non-invasive method for exposure-aware risk assessment in oral health. Although current clinical translation is limited by methodological variability, small cohort sizes, and the lack of standardized reference ranges, integration with digital biosensing platforms and explainable AI frameworks might facilitate scalable, precision-oriented salivary diagnostics. Full article

The dominant sound sources in different urban spaces influence residents’ multidimensional emotional responses, and the interaction of various sound sources across temporal and spatial dimensions forms a complex sound environment. This study aims to develop a comprehensive index to quantify the emotional impacts of dominant sound sources. Through field measurements, this study classified the collected audios into four major categories (natural, social, construction, traffic) and 15 subcategories, with each sound source characterized by SPL and primary frequency. A total of 1266 questionnaires were collected from 209 participants through a web-based survey for the subjective experiment, while EEG data were obtained from 35 participants in the objective experiment. Next, by integrating acoustic indicators, subjective questionnaire responses, and objective EEG data, this study constructs the $C_K$ index using principal component analysis. $C_K$ provides a single, interpretable score of emotional impact, where lower values indicate greater calm. Results show that natural sounds consistently outperformed the other three sound types, showing the highest comfort (3.54) and pleasure (3.40) ratings on a five-point Likert scale, as well as the strongest physiological response with a parietal alpha power of 18.44 μV²/Hz. The calculated $C_K$ values for natural, social, construction, and traffic sounds were 8.73, 9.91, 10.20, and 10.29, respectively. This study contributes to quantifying the emotional impacts of urban sounds and refining noise mitigation priorities using the $C_K$ index. Full article

To accurately understand and investigate carbon fluxes in cropland ecosystems, this study adopted a machine learning ensemble model for estimation. Focusing on the Jinzhou station of the ChinaFLUX, we integrated eddy covariance carbon flux observations with multi-source satellite remote sensing data to construct a machine learning-based cropland carbon flux estimation model. For environmental driver selection, a strategy combining correlation analysis with ecological mechanism understanding was employed to screen LST, NDVI, and NDMI as model input variables, effectively avoiding multicollinearity issues. Using footprint-weighted integrated data from 2005 to 2014 for model training and validation, a Stacking ensemble model was constructed with the RF model serving as the meta-learner to stack the predictions of RF, CART, and GBM. The ensemble model further reduced the prediction error (RMSE = 39.82), maintaining an R² > 0.9 in most years and effectively improving predictive performance during anomalous years where single models underperformed. Based on these findings, the model was applied to analyze the spatiotemporal evolution of NEE in Jinzhou croplands from 2005 to 2014. The analysis revealed that while the region functioned overall as a carbon sink, it exhibited significant spatiotemporal heterogeneity. Spatially, the distribution followed a pattern of “strong intensity in the northeast and center, and weak intensity in the northwest and southwest.” Temporally, the sink intensity underwent significant interannual oscillations characterized by a “strengthening–weakening–re-strengthening–declining” trajectory. The high-precision prediction method proposed in this study is of great significance for revealing spatiotemporal variations in carbon sources/sinks, guiding green agricultural development, and supporting relevant policy formulation. Full article

In the past two decades, important progress has allowed a better understanding of how light signals are perceived by plants, not only as a source of energy for photosynthesis but also as environmental cues that modulate growth, development, and stress responses. These advances open up promising prospects for light-based treatments in agriculture. This review synthesizes recent scientific findings on the application of specific wavelengths (from ultraviolet to infrared) to improve crop yield, quality, and resilience. The analysis focuses on controlled environment agriculture, where most experimental data have been generated and where the integration of lighting strategies is technically more feasible compared to open-field settings. Preharvest: we explore how spectral quality, intensity, and duration can be used to modulate plant growth, photosynthesis, defense pathways, and the accumulation of nutritional compounds. Postharvest: the focus shifts to how light can help maintain visual and nutritional quality, regulate ripening, limit pathogen development, and extend shelf-life. The review emphasizes plant photoreceptors and signal transduction pathways, as well as technical parameters such as spectrum selection, application timing, and lighting configuration. A selection of recent patents illustrates how fundamental research is being translated into deployable, energy-efficient lighting technologies for sustainable crop management. Full article

Business Intelligence (BI) comprises methods and technologies for collecting, organizing, and analyzing data to support managerial decision-making. This study presents a systematic literature review with a two-tier scope: first, identifying the most widely adopted BI tools across organizational contexts, and second, examining the specific application of Microsoft Power BI within the civil construction sector. The review followed the PRISMA guidelines and was complemented by the snowball sampling technique. A total of 81 articles published between 2015 and 2025 were analyzed to identify the most used tools, main application sectors, benefits, and challenges in BI adoption. The analysis combines descriptive bibliometric techniques with qualitative content analysis to examine publication trends, tools, application domains, and reported challenges. Results indicate that Power BI, Tableau, and Qlik Sense are the most frequent BI tools, with Power BI standing out for its integration with diverse data sources such as spreadsheets, databases, management software, and cloud platforms, enabling the creation of dashboards. The civil construction, business management, and manufacturing industries show the highest adoption rates, mainly for cost control, performance monitoring, and sustainability indicators. Reported benefits include operational efficiency, process automation, and improved decision-making. However, gaps remain regarding data standardization, interoperability, technological infrastructure, and user resistance. As a contribution, this review advances the existing literature by explicitly distinguishing general BI tool adoption from the sector-specific use of Power BI in civil construction, systematically classifying application domains and revealing limitations in maturity that remain underexplored in prior reviews. Full article

The Prespa Lakes system, shared between Greece, the Republic of North Macedonia, and Albania, forms a significant transboundary, large-scale integrated freshwater ecosystem subject to multiple anthropogenic and natural pressures. This study focuses on the Greek part of the Prespa Lakes system with particular emphasis on the identification of the ecological and hydrological impacts of the contributing stream inflows on the lakes by examining the spatial variability in physicochemical and biological conditions and conducting water balance and isotopic analyses. Based on our results, streams draining into Lesser Prespa Lake exhibited more pronounced hydrological and physicochemical fluctuations than the Agios Germanos River connected to Great Prespa Lake, while ecological status classifications of all studied streams ranged from high to moderate. Furthermore, moderate ecological status conditions (mainly observed at the downstream stations) were closely associated with adjacent anthropogenic pressures, including agricultural drainage, livestock activities, irrigated croplands, and wastewater discharges. In addition, although both lakes were classified as mesotrophic, field data indicated greater transparency loss in Lesser Prespa than in Great Prespa Lake. Regarding the stream influences on Lesser Prespa Lake’s water quality, nutrient loads induced changes in lake concentrations by roughly one month. Total nitrogen showed moderate stream–lake correlations (R = 0.61) and a strong negative correlation for total phosphorus (R = −0.94), suggesting substantial nutrient retention and processes within the lake. Water balance analysis revealed an annual water deficit for both Lesser and Great Prespa, with the latter exhibiting a markedly stronger and systematic long-term decline in water level. In the Lesser Prespa, seasonal fluctuations in water volume were primarily driven by excess rainfall, while stream inflows contributed minimally. Conversely, correlation analysis for Great Prespa identified surface inflow from the Ag. Germanos catchment as the dominant driver of water storage variability, surpassing direct rainfall, with strong correlations in both wet (R = 0.79) and dry (R = 0.88) periods. Isotopic compositions (δ¹⁸O, δ²H) did not differ significantly between the two lakes, indicating common recharge sources and strong evaporative imprints, while stream isotopic signatures highlighted spatial and seasonal variability in hydrological inputs. Seasonal and spatial variations were proved to be strongly influenced by both natural hydrological dynamics and anthropogenic pressures within the basin, while these findings reinforce the importance and the necessity of adopting holistic, cross-border management strategies that maintain the ecological integrity and the long-term sustainability of the Prespa Lakes ecosystem. Full article

To investigate the dynamic behavior of smart composite structures with embedded heat sources over a wide temperature range, this study employed thermoplastic polypropylene as the matrix, combined with glass/carbon fiber prepregs and Ni80Cr20 alloy heating wires, and fabricated functional laminated specimens with integrated heating elements via a prepreg molding process. Using a self-developed variable-temperature cantilever beam vibration testing system, the evolution of natural frequencies and damping ratios from room temperature to 140 °C was systematically examined. Results indicate that temperature-induced thermal softening of the polypropylene matrix reduces the effective bending stiffness of the composites, leading to a decline in natural frequencies across all modes. For example, the first-order natural frequency of the sample decreased from approximately 30.8 Hz at room temperature to about 28.3 Hz at 140 °C, representing a reduction of approximately 8.12%. The second-order reduction reached about 8.99%, and the third-order reduction was approximately 9.65%. Carbon fiber-reinforced specimens exhibited relatively smaller frequency reductions due to the high modulus of the fibers. Concurrently, elevated temperatures enhance molecular chain mobility and interfacial viscoelastic dissipation at the fiber–matrix interface, causing a sharp increase in damping ratios at high temperatures (>100 °C). For instance, the damping ratio of the first-order mode increased significantly from approximately 1.02% at room temperature to about 2.9% at 140 °C. By comparatively analyzing carbon fiber and glass fiber systems, the study elucidated the distinct mechanisms underlying the “fiber-dominated” stiffness retention effect and the “resin/interface-dominated” damping dissipation effect under thermal influence. These findings provide critical experimental data and theoretical references for the active thermal regulation of structural performance in thermoplastic composite structures with integrated heat sources, thereby mitigating damage caused by external disturbances. Full article