Browse Articles
Discover research articles across all indexed journals
Mechanistic study of plastic monomers in gestational diabetes mellitus: A network toxicology and molecular docking approach
Plastics are widely used in various fields such as food packaging, textile fibers, building materials, and transportation. Although the relationship between plastic additives and diseases has been reported, there is limited research on the association between plastic monomers (PM) and gestational diabetes mellitus (GDM). This study aims to investigate the link between environmental PM and GDM. By employing advanced network toxicology and molecular docking techniques, we successfully elucidated the molecular mechanisms by which PM may induce GDM. Utilizing databases such as PubChem, SEA, Super-PRED, SwissTargetPrediction, PharmMapper, Gene Cards, and OMIM, we identified potential targets associated with the disease. Further analysis using STRING and Cytoscape software helped determine the core targets most significantly related to these metabolic disorders. Additionally, Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses were conducted using the David database to characterize these core targets. Finally, molecular docking with CB-Dock2 was used to validate the binding affinity of PM to these target proteins. Our findings suggest that PM may potentially induce GDM by modulating the insulin signaling pathway through STAT3, AKT1, and TP53. In summary, this work provides novel insights into the mechanisms by which environmental pollutants may trigger GDM, thereby laying a theoretical foundation for disease prevention and treatment. It offers valuable references for the safety evaluation of plastics, urging food safety regulatory agencies to strengthen oversight and encouraging the public to reduce plastic usage.
Spatio-temporal variation and dynamic scenario simulation of ecological risk in a typical artificial Oasis in Northwestern China
A histogram transformer approach using attention-based 3D residual network for human action recognition
This paper proposes a lightweight video action recognition framework that integrates 3D Convolutional Neural Networks (CNNs), the Histogram Transformer Block (HTB), and the Split-Attention Residual Block (SAB), while also introducing Spatiotemporal Tensor Factorization (ST-Factor) technology in an innovative manner. The method first incorporates the HTB module into each computational unit of the AR3D backbone network to leverage local statistical features for improve the granularity of spatiotemporal modeling. Next, the SAB module is introduced into the residual path to utilize dynamic channel re-weighting for optimizing feature selection across dimensions. Finally, the ST-Factor decouples the 4D convolution kernels into independent spatial (H × W) and temporal (T × C) operations, which significantly reducing computational redundancy. Experiments on the UCF101/HMDB51 datasets demonstrate that the proposed method not only maintains real-time inference speed but also outperforms existing state-of-the-art (SOTA) methods in recognition accuracy, providing a new paradigm for video understanding research.
AI-driven smart grid optimization for hospital energy systems integrating renewable generation, predictive maintenance, and resilient infrastructure
Correction: An exploratory study of psychological and decision-making outcomes among family members of critically ill patients
Multi-omics analysis reveals shared diagnostic and therapeutic targets in endometriosis and recurrent implantation failure
Abstract Endometrial receptivity is essential for successful pregnancy, and endometriosis is widely recognized as a disruptor of this process. Poor endometrial receptivity is also a key factor contributing to recurrent implantation failure. Although some molecular mechanisms related to endometrial receptivity have been identified, their specific roles in endometriosis and recurrent implantation failure remain unclear. This study aimed to elucidate the shared molecular mechanisms affecting endometrial receptivity in endometriosis and recurrent implantation failure using multi-omics data analysis. We sourced datasets from the NCBI GEO database and employed weighted gene co-expression network analysis to identify gene modules associated with these conditions, followed by gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. Single-cell sequencing analysis and immunofluorescence were used for expression analysis. We identified 3690 and 4892 upregulated genes and 2675 and 5065 downregulated genes in endometriosis and recurrent implantation failure, respectively. Functional enrichment analysis and validation identified 15 hub genes including SRPRB, SLC35B1, and SLC25A6. Receiver operating characteristic curve analysis demonstrated that these genes are associated with high diagnostic accuracy. Single-cell sequencing analysis indicated that these genes are predominantly expressed in basal epithelial cells, with RBM3 being particularly prominent. This study provides new insights into the molecular mechanisms underlying endometrial receptivity and identifies potential targets for the diagnosis and treatment of endometriosis and recurrent implantation failure.
Accelerated Stochastic Conjugate Gradient for a class of convex optimization
The conjugate gradient method is widely recognized as a foundational technique for large-scale unconstrained optimization. In this work, we introduce an Accelerated Stochastic Conjugate Gradient (ASCG) algorithm, specifically designed for a class of convex empirical risk minimization problems. The proposed ASCG method integrates a variance-reduced gradient estimator-inspired by modern stochastic variance reduction techniques-to control noise and improve stability in the optimization process. Moreover, the ASCG algorithm incorporates a novel acceleration mechanism via a deflation factor on the step size, which is shown to achieve faster practical convergence compared to the baseline stochastic FR method. We provide a rigorous theoretical analysis demonstrating that ASCG achieves an expected linear convergence rate under strong convexity assumptions and attains a superior reduction in function values compared to non-accelerated stochastic counterparts. Extensive numerical experiments on four widely-used benchmark datasets confirm that ASCG consistently outperforms state-of-the-art stochastic optimization methods.
Improved YOLOv9-based remote sensing image detection method
A comparative framework for convergence analysis of perturbation series techniques in nonlinear fractional quadratic differential equations
This study tackles the challenge of obtaining highly accurate approximate solutions for nonlinear fractional differential equations, which often lack exact solutions due to their inherent complexity. A unified perturbation framework is proposed based on homotopy topology theory, enabling multiple formulations depending on the number of convergence-control parameters. Through dynamic adjustment of these parameters, the Homotopy Method achieves enhanced precision, particularly for fractional-order models exhibiting long-memory behavior. Numerical results clearly demonstrate that increasing the number of convergence parameters leads to significantly improved accuracy. Supported by detailed graphs and tables, the proposed approach proves to be a flexible, robust, and reliable tool for solving nonlinear fractional differential equations.
A framework for real-time image detection of bioaerosols
Abstract Starting from the elastic light scattering and the induced fluorescence emission from airborne particles and extending to the counting and differentiation of the associated signals, a framework is proposed to describe the detection of bioaerosols using an image sensor. For validation, monodisperse NaCl particles were generated to mimic abiotic particles, while monodisperse 1% mass riboflavin particles were generated to mimic biotic particles. By challenging a prototype sensor with these particles, the exposure time, particle speed, and signal-to-noise ratio were shown to be critical parameters for detection. Additionally, it was displayed that the induced fluorescence emission can be isolated from the elastic light scattering by using a well-selected long-pass filter. Furthermore, correlating the color of the captured signals to an induced fluorescence contribution was shown to be a potential avenue of differentiation between biotic and abiotic particles. It is predicted that this color differentiation method can distinguish between a near-continuous range of visible induced fluorescence emission wavelengths, giving the ability to distinguish individual fluorophores from one another using simple filtering and a single detector. This framework will be used to further optimize the image-based bioaerosol sensor evaluated here.
Mitigating semantic label divergence in federated learning: Obfuscated encoding and alert filtering for security monitoring
Federated learning (FL) is emerging as a key approach for collaborative machine learning (ML) in distributed information systems where direct data sharing is infeasible due to policy constraints. In security operations center (SOC) settings, we study FL for the classification of network intrusion detection system (IDS) alerts —structured event records emitted by sensors (e.g., Snort/Suricata)—where consistent interpretation of event data is critical for reliable ML-based decision support. However, differences in labeling criteria across organizations often lead to semantic inconsistencies, undermining the accuracy and generalizability of FL models. This paper presents two key contributions that mitigate this issue without requiring raw data exchange. First, we propose Keyed Feature Hashing (KFH) , a key-dependent obfuscated encoding scheme that enables consistent vectorization of heterogeneous IDS alerts across entities while reducing the risk of model inversion. Second, we introduce a filtering mechanism that leverages KFH representations to identify and exclude alerts likely to be misclassified due to inter-entity label discrepancies. Experiments using a large-scale real-world dataset collected from 14 organizations demonstrate that our method improves classification F1-score by up to 13.36% while maintaining over 99% alert coverage. These contributions enhance the trustworthiness of FL-based decision models in distributed, label-divergent environments.
A painting art rendering system by deep learning framework and machine translation
Correction: Nuclei Segmentation and Classification from Histopathology Images using Federated Learning for End-Edge Platform
Medication review improves pain management and quality of life in chronic pain: a pilot randomized controlled study
A methodology for calculating the rarity of diverse proteins based on functional specificity and thermodynamic stability
A key question in protein studies is the proportion of amino acid sequences that correspond to functional proteins, often called protein rarity. This issue underlies the relationship between mutations and disease, theories on the origin of proteins, and strategies for engineering new proteins. Recent literature has detailed how to employ estimates of protein rarity to evaluate the required biasing of functional sequences in sequence space to allow for evolutionary paths to connect distinct proteins. One challenge in addressing rarity has been an imprecise definition of function and a lack of consistency in methodology. This study introduces a new methodology, referred to as PRISM, to evaluate protein rarity based on the impact of mutations on stability. PRISM offers a suite of methods that are simpler than traditional approaches while providing accurate upper-bound rarity estimates. The specific method applied is determined by the protein’s function and available empirical data on how accumulating mutations affect its stability and performance. PRISM is applied to several proteins, and the accuracy of the methods is demonstrated by comparing the results to rarity estimates from previous studies. The calculated rarities align with previous research that concludes functional sequences are often exceedingly rare. The application of PRISM is outlined for research in protein engineering, protein evolution, and pathology.
Hypothermia association with all-cause mortality in critically ill patients with sepsis based on the MIMIC-IV database
Sestrin3 confers resistance to recombinant human arginase in small cell lung cancer by activating Akt/mTOR/ASS1 axis
Drug resistance is a major obstacle in the clinical management of small cell lung cancer (SCLC), we have proved the promising anticancer effect of recombinant human arginase (rhArg, BCT-100) in SCLC in vitro and in vivo . In order to promote the clinical application of recombinant human arginase, it is necessary to explore the underlying resistant mechanisms of BCT-100 in SCLC. Here, we cultured and obtained the acquired drug-resistant SCLC cell line (H446-BR), which displayed different cellular phenotypes (enhanced migration ability) compared with the parental cell line (H446). sestrin3 (SESN3) was confirmed with high expression in resistant cell line. Knockdown SESN3 could re-sensitize resistant cells to BCT-100 treatment and reverse the aggressive feature of H446-BR. The Akt-mTOR signal pathway and ASS1, which were highly expressed in resistant cells, were down-regulated after silencing SESN3. MK-2206 and rapamycin suppressed the expression of ASS1 in H446-BR cell. In xenograft model, BCT-100 has little anti-tumor effect on H446-BR compared with H446 as well as H446-BR silenced sestrin3. Collectively, these results elucidate SESN3 plays an essential role in resistant mechanism, which will provide a valuable source of information for translational research.
AI-driven tongue image analysis for diagnosing and predicting coronary artery disease
A multimodal framework for attenuation of piston and planar waves in impedance-lined ducts
This study presents a multimodal formulation to investigate the scattering of acoustic waves in duct systems lined with locally reacting liners. The primary objective is to analyze the attenuation of piston-driven and planar acoustic radiations by employing impedance-based boundary conditions that accurately model liner behavior. A multimodal framework is developed to solve the governing boundary value problems by projecting acoustic fields onto orthogonal basis functions, with eigenvalues and eigenvectors used to characterize the modal propagation. The proposed method is validated against benchmark configurations, including rigid-walled ducts and ducts with impedance boundaries, and cross-compared with traditional mode-matching techniques. Numerical results demonstrate the effectiveness of the liners in attenuating acoustic energy, particularly at low frequencies, and confirm the convergence of the multimodal approach across a range of excitation conditions. The formulation is further applied to a reactive silencer geometry containing impedance-lined cavities, highlighting the liner’s influence on wave scattering and overall noise reduction performance. This work provides a comprehensive modeling framework for evaluating liner treatments in complex acoustic systems and contributes to the design of efficient noise-control devices in ducts and silencers.