Browse Articles
Discover research articles across all indexed journals
Meta-cluster driven ensemble learning with cluster-augmented features for explainable heart disease prediction
Structural characterization and visualization of oligomeric states of the Rhagium mordax antifreeze protein
Secure electronic voting system based on blockchain with dilithium authentication and zero-knowledge proof verification
House price prediction using a hybrid GRU–MLP based on binary whale optimization algorithm and ant colony optimization for hyperparameter tuning
Abstract Accurate house price prediction is essential for real estate valuation, investment planning, and intelligent property decision-support systems. This study proposes an optimized hybrid deep learning framework that integrates a Gated Recurrent Unit and Multilayer Perceptron model with the Binary Whale Optimization Algorithm for feature selection and Ant Colony Optimization for hyperparameter tuning. The proposed framework was evaluated using a publicly available Kaggle house price regression dataset containing 500 housing records with structural, locational, and amenity-related attributes. The dataset was divided into training, validation, and testing subsets using a 70:20:10 ratio, and leakage-free normalization was applied using only the training data. Experimental results show that the proposed BWOA–ACO–GRU–MLP model outperformed standalone GRU, MLP, CNN, LSTM, and BiLSTM models. It achieved an MSE of 0.0146, MAE of 0.1051, RMSE of 0.1208, MAPE of 0.0112, MedAE of 0.0969, and an R 2 of 99.04%. These results demonstrate that combining feature selection, hyperparameter optimization, and hybrid neural regression improves prediction accuracy and model stability for house price estimation. The proposed framework provides a reliable data-driven approach for smart real estate valuation applications.
Resolving FRET signal degeneracy and population heterogeneity via Bayesian nonparametrics
The effects of using a multi-modal rest cabin versus a waitlist control on worker well-being: A pragmatic randomized controlled trial
Abstract Workplace rest cabins are a promising yet understudied approach to supporting worker well-being. This pragmatic randomized controlled trial examined whether a multi-modal rest cabin could improve general and work-related well-being among workers of a Canadian insurance company. Participants were randomly assigned to either a one-month waiting-list control condition or a cabin-use condition. A total of 80 participants were included in the final analytic sample. Surveys were completed at baseline (T0) and again one month later (T1); cabin users also completed follow-up surveys two (T2) and four months (T3) after completing T0. Outcomes included positive and negative affect, depression, anxiety, flourishing, burnout (namely cognitive weariness, emotional exhaustion, and physical fatigue), absenteeism, presenteeism (namely troublesome symptoms at work and impaired productivity), job satisfaction, mindful attention awareness, insomnia, and psychological detachment. Cabin-use frequency was extracted from reservation logs. Linear mixed-effects models evaluated i) differential change between conditions from T0 to T1, ii) maintenance of effects within the experimental group across Times 1 to 3 relative to T0, and iii) whether use frequency was associated with changes in well-being from T0 to T1. Compared with the control group, cabin users experienced protective effects against depression, cognitive weariness, and physical fatigue, and protective effects approaching statistical significance against a deterioration in positive affect and mindful attention awareness. When examining maintenance of effects at follow-ups (versus baseline) in the experimental group, positive affect, cognitive weariness, and mindful attention awareness improved, or trended towards improvement, at T3, whereas physical fatigue improved, or trended towards improvement, at each time point. Workers who used the rest cabin more frequently experienced improved well-being across multiple indicators (positive affect, physical fatigue, insomnia, impaired productivity, mindful attention awareness); however, those who used the cabin four times in the one month showed greater improvement in physical fatigue, insomnia, and mindful attention awareness. Effects for all other outcomes were neither significant nor trending. A multi-modal rest cabin may buffer against declines in well-being and appears to offer accumulating benefits with more sustained use, suggesting usefulness as a personalized micro-break intervention in organizational settings. Trial registration: Retrospectively registered. Registration number: NCT07322354; Date: 2026-01-07.
Fairness and transparency in ML: a methodological framework
Octameric dGTPase assemblies mediate broad anti-phage defense
Abstract Deoxyguanosine triphosphatases (dGTPases) are nucleotide-depleting enzymes known to play a role in antiviral defense. While their enzymatic mechanism is established, the structural and functional diversity of dGTPases remains poorly understood. Here, we report a systematic analysis of dGTPase homologs across bacteria, archaea, and eukaryotes, revealing their widespread distribution and association with diverse immune-related domains. Through integrative bioinformatics and structural mining, we identify a class of bacterial dGTPases that assemble into stable octameric and higher-order oligomeric structures. Using cryo-electron microscopy, we resolve the octameric and 16-mer assemblies of a representative Vibrio dGTPase (Vdg) and further captured filamentous forms. Functional assays demonstrate that octamer formation is essential and sufficient for antiviral activity, while higher-order assemblies are dispensable. We also identify dAMP as an allosteric regulator, underscoring the functional versatility of dGTPases. Our findings provide insights into the modular architecture, oligomerization-driven activation, and immune function of bacterial dGTPases, and broaden our understanding of nucleotide depletion-based antiviral strategies.
Acoustic wave speed in aerogels across material classes: density scaling from theory and experiments
Abstract Understanding the propagation of elastic waves in ultra-light porous solids is essential for linking their microstructure to macroscopic properties and mechanical performance. In this work, we present first an experimental study of longitudinal sound velocity in aerogels spanning a wide range of chemical structures and mesoscopic architectures, including polyurethane, polylactic acid, polyimide, flexible organo-silica, classical silica, and phenolic (resorcinol-formaldehyde) aerogels. Compression wave velocities were measured and correlated with bulk density to establish scaling relations across aerogel material families. While classical elasticity implies direct coupling between sound velocity in isotropic solids and elastic moduli, we then examine the extent to which such continuum relations remain valid in aerogels, whose structure is governed by hierarchical porosity, nanoscale connectivity, and bending-dominated network mechanics. Finally, we suggest measuring the Young modulus of aerogels from the acoustic wave speed as a non-destructive testing alternative.
Low-dose baricitinib plus danazol in primary immune thrombocytopenia: a randomized, controlled phase 2 trial
Adaptive trajectory tracking control for autonomous vehicles based on heading angle deviation
Scalable high resolution ancestry deconvolution for genomic data
Abstract As genome-wide association studies and genetic risk prediction models extend to globally diverse and admixed biobanks, accurate, scalable ancestry deconvolution, also called local ancestry inference (LAI), has become crucial. LAI assigns ancestry to each genomic segment within an individual, enabling studies of population history and ancestry-associated haplotypic effects. Existing LAI methods scale poorly to biobank-scale data, to the distant past, and to large numbers of ancestries. Here, we introduce several independent LAI methods implemented in the Gnomix software suite, achieving higher accuracy and faster computational performance than all existing approaches and with portable models that can be shared without exposing individual-level training data. Gnomix is paired with Gnofix, a swift, scalable phase correction counterpart. We demonstrate performance on worldwide whole-genome data from humans and canids, leveraging high-resolution accuracy to localise ancient New World haplotypes in the Xoloitzcuintli, dating back over 100 generations. Code is available at https://github.com/AI-sandbox/gnomix .
Spatiotemporal dynamics and driving mechanisms of the coupling coordination between transportation carbon emissions and regional economy: a case study of Sichuan Province
Retarding moisture-induced chemical degradation of Yttrium Tellurides by tailoring grain boundary chemistry
Abstract Grain boundary engineering has been extensively applied to improve thermoelectric performance, but its potential to enhance chemical stability remains underexplored. Here, we demonstrate that modifying grain boundary chemistry can effectively suppress the chemical degradation of Y 2 Te 3 under ambient conditions. Scanning transmission electron microscopy and atom probe tomography reveal that H 2 O preferentially infiltrates along grain boundaries, initiating oxidation of Y 2 Te 3 into Y–O–H phases and causing chemo-mechanical breakdown of the matrix. This process, remarkably, can be retarded by just 1 at.% of Bi incorporation due to its segregation along grain boundaries. Density functional theory calculations reveal the thermodynamic and kinetic origins of Bi segregation, and show how segregated Bi modifies the local electronic and chemical environment of grain boundaries, thereby linking GB chemistry to both chemical stability and thermoelectric performance. These findings establish multifunctional grain boundary engineering as a generalizable strategy for the design of next-generation thermoelectric materials.
Multi timescale predictive energy management for battery life extension in electric vehicles
Abstract Electric-vehicle battery energy management increasingly requires coordinated control of electrical demand, thermal behavior, and long-term degradation to ensure safe and durable operation. Existing predictive and digital-twin-inspired model-based observer battery-management approaches often do not fully integrate fast electro-thermal regulation with slow health-aware supervisory adaptation. To address this gap, this study proposes a multi-timescale health-resilient predictive energy-management framework that combines a fast predictive control layer for real-time traction-demand satisfaction with a slow supervisory layer that updates health-dependent limits, adaptive weights, and operating envelopes using cumulative electro-thermal-aging stress. The framework is evaluated in discrete-time simulation under Urban-Nominal, Highway-Nominal, Aggressive-Hot, and Aged-Battery-Hot scenarios against Rule-Based, Fast-MPC-Only, and Electro-Thermal-MPC strategies. Results show that the proposed controller consistently provides the most favorable battery-preservation tradeoff. Under Urban-Nominal operation, it reduces RMS current to 125.75 A and achieves the lowest cumulative degradation among the predictive controllers. Under Aggressive-Hot operation, it lowers RMS current to 120.44 A and cumulative degradation to $$\:1.4929\times\:{10}^{-5}$$ , while under Aged-Battery-Hot conditions it again yields the lowest degradation and loss-energy trends among the predictive methods. The proposed framework therefore offers a balanced compromise between short-term energy-management performance and long-term battery durability, suggesting its potential usefulness for health-aware EV battery energy-management studies, subject to further experimental and long-horizon validation.
Alumina-Templated ortho-Prenylation of Phenols
DFT and molecular docking investigation of tamoxifen interactions with metal-encapsulated boron nitride nanocages
Abstract Tamoxifen (TMF), a lipid-soluble selective estrogen receptor modulator (SERM), is extensively used in the treatment of breast cancer. In this work, we systematically investigate the covalent and non-covalent interactions of TMF with the perfect B 12 N 12 compared to as well as with calcium- and potassium-encapsulated derivatives (B 12 CaN 12 and B 12 KN 12 ). These interactions are predominantly mediated by the dimethylamino group (-N(CH₃)₂) of TMF and were examined using density functional theory (DFT) calculations at the M06-2X level, incorporating D3 dispersion corrections (M06-2X-D3) and the 6–31 + G** basis set. The results reveal that TMF undergoes strong chemisorption on B 12 KN 12 (-2.27 eV) and B 12 CaN 12 (-1.91 eV), in contrast to weaker adsorption on the pristine B 12 N 12 surface (-1.75 eV). The strong binding of TMF to B₁₂N₁₂ via its dimethylamino group occurs through a synergistic combination of covalent interactions and hydrogen bonding, accompanied by a larger charge transfer from the drug to the cage, leading to a significant increase in dipole moment and a change in the energy gap. Thermodynamic analyses based on Gibbs free energy and enthalpy changes confirm that complex formation is highly stable, spontaneous, and exothermic. Notably, B 12 N 12 exhibits the shortest recovery time, indicating rapid TMF detachment, whereas B 12 CaN 12 and B 12 KN 12 show longer desorption times, favoring sustained release. ADMET predictions suggest moderate intestinal absorption and good membrane permeability, indicating potential for oral bioavailability. Docking studies reveal that B 12 KN 12 enhances TMF binding to EGFR, HER2, and Caspase-8, despite a minor reduction in ERα affinity, supporting the potential of B 12 KN 12 as a nanocarrier for HER2-driven breast cancer treatment.
Highly scalable, dialysis-controlled synthesis of colloidal molecules via regulated self-assembly of polymer-grafted gold nanoparticles
Electrostatic headroom coordinated minimum feasible high-frequency injection strategy for low-speed motor drives using a half bridge modular multilevel converter
Abstract This paper proposes a boundary-based minimum-injection high-frequency balancing strategy for low-speed standard half-bridge modular multilevel converter (MMC) motor drives to reduce the injected high-frequency balancing voltage and current and the associated electrical stress while maintaining the prescribed submodule capacitor-voltage ripple limit. The proposed method preserves the conventional Korn-type high-frequency injection (HFI) balancing path, but determines the commanded injection level from a constrained minimum-injection condition governed by the residual arm-energy demand and the available electrostatic buffering capability of the capacitor stack. An analytical operating boundary between HFI-assisted balancing and capacitor-based self-buffering is established by comparing the residual-energy requirement with the usable capacitor-voltage headroom constrained by modulation feasibility and the upper submodule voltage limit. Accordingly, the injection coefficient and the average capacitor-voltage reference are jointly coordinated so that the capacitor stack buffers the admissible residual energy, whereas the HFI channel supplies only the remaining balancing component required for ripple regulation. Comparative simulations against fixed HFI and a recent adaptive HFI benchmark demonstrate reductions of 33.0% in the mean compensation coefficient, 29.1% in the representative peak-to-peak submodule capacitor-voltage ripple, 13.7% in the representative arm-current RMS, 28.5% in the representative arm-current peak, and 25.3% in the current-squared loss proxy. The common-mode-voltage benefit is mainly reflected in the RMS and accumulated-burden indices rather than in the reduction of every instantaneous peak. These results verify that the proposed strategy reduces the injected high-frequency balancing content and the associated current- and voltage-side stresses under explicit ripple, modulation, and capacitor-headroom constraints.