Search Results (147)

Background/Objectives: To explore the protective effects of 7-Ketolithocholic acid (7-KLCA) against renal fibrosis and its mechanism, focusing on its interaction with farnesoid X receptor (FXR). Methods: In vitro, TGF-β-induced fibrosis in HK-2/NRK-49F cells and LPS-induced inflammation in HK-2 cells were detected by CCK-8, Western blot, and qPCR. In vivo, unilateral ureteral obstruction (UUO) and adenine (Ade)-induced mouse models were treated with a low/high-dose 7-KLCA or losartan. Renal injury was evaluated via H&E/Masson staining, serum creatinine (SCR), and blood urea nitrogen (BUN) levels. The 7-KLCA-FXR interaction was verified by molecular docking, CETSA, and DARTS. FXR downstream genes and related proteins were measured by WB and qPCR. Results: 7-KLCA inhibited the expression of fibrotic proteins (fibronectin, collagen-I) and reduced the LPS-induced release of inflammatory factors (IL-1β, IL-6). In mice, it alleviated renal swelling, collagen deposition, and tubular damage, while lowering serum SCR and BUN levels. Mechanistically, 7-KLCA stably bound to the FXR ligand-binding domain, enhanced its thermal stability and degradation resistance. It upregulated FXR and its downstream genes SHP and FGF15, thereby inhibiting the activation of TGF-β/Smad and Wnt/β-catenin pathways. Conclusions: This is the first study to clarify the molecular mechanism through which 7-KLCA targets FXR and dually suppresses the key pro-fibrotic pathways TGF-β/Smad and Wnt/β-catenin, thereby exerting anti-renal fibrosis effects. Full article

Plasma membrane Ca²⁺-ATPases (PMCA) are activated by calmodulin (CaM) via a C-terminal calmodulin-binding domain, CaMBD. Although specific mutations in this domain have been linked to disease, the broader impact of alternative substitutions across the interface remains unexplored. We applied an integrative in silico workflow to test six substitutions within CaMBD positions 1–18, L5R, N6I, I8T, V14E/D, and F18S, across PMCA isoforms 1–4. CaMBD sequences were aligned across isoforms, and candidates for substitutions were selected by conservation and nucleotide feasibility, prioritizing conserved or co-evolutionarily relevant sites, with substitutions possible by single-nucleotide change. PolyPhen-2 screened the impact of the substitutions on the protein functionality, the DisGeNET database was used to contextualize ATP2B genes with clinical phenotypes, and structural models plus binding free energy changes were estimated with AlphaFold3, FoldX, and MutaBind2. Effects were isoform and subregion dependent, with the strongest weakening toward the CaMBD C-terminus. V14E/D and F18S showed the largest and consistent predicted destabilization, consistent with disruption of conserved hydrophobic anchors. I8T and L5R had mixed outcomes depending on isoform, while N6I presented various scenarios with no clear effect. PolyPhen-2 classified most tested substitutions as damaging. Gene-disease evidence linked ATP2B to neurological, endocrine, and oncologic phenotypes, consistent with roles in Ca²⁺ homeostasis. Overall, CaMBD appears highly sensitive to perturbation, with distal positions 14–18 particularly vulnerable to substitutions that can destabilize CaM binding and potentially impair PMCA-mediated Ca²⁺ clearance in susceptible tissues. Full article

Reliable detection of X-ray tire defects is essential for safety and quality assurance in manufacturing. However, low contrast and high noise make traditional methods unreliable. This paper presents DyReCS-YOLO, a dynamic re-parameterized channel-shuffle network based on YOLOv8. The model introduces a C2f_DyRepFusion module combining dynamic convolution and a shuffle-and-routing mechanism, enabling adaptive kernel adjustment and efficient cross-channel interaction. Experiments on an industrial X-ray tire dataset containing 8326 images across 58 defect categories demonstrate that DyReCS-YOLO achieves an mAP@0.5 of 0.741 and mAP@0.5:0.95 of 0.505, representing improvements of 4.5 and 2.8 percentage points over YOLOv8-s, and 9.2 and 7.7 percentage points over YOLOv11-s, respectively. The precision increases from 0.698 (YOLOv8-s) and 0.668 (YOLOv11-s) to 0.739, while maintaining real-time inference at 189.5 FPS, meeting industrial online detection requirements. Ablation results confirm that the combination of dynamic convolution and channel shuffle improves small-defect perception and robustness. Moreover, DyReCS-YOLO achieves an mAP@0.5 of 0.975 on the public MT defect dataset, verifying its strong cross-domain generalization. Full article

Medical anomaly detection is challenged by limited labeled data and domain shifts, which reduce the performance and generalization of deep learning (DL) models. Hybrid convolutional neural network–Vision Transformer (CNN–ViT) architectures have shown promise, but they often rely on large datasets. Multistage transfer learning (MTL) provides a practical strategy to address this limitation. In this study, we evaluated parallel hybrids, where convolutional neural network (CNN) and Vision Transformer (ViT) features are fused after independent extraction, and sequential hybrids, where CNN features are passed through the ViT for integrated processing. Models were pretrained on non-wrist musculoskeletal radiographs (MURA), fine-tuned on the MURA wrist subset, and evaluated for cross-domain generalization on an external wrist X-ray dataset from the Al-Huda Digital X-ray Laboratory. Parallel hybrids (Xception–DeiT, a data-efficient image transformer) achieved the strongest internal performance (accuracy 88%), while sequential DenseNet–ViT generalized best in zero-shot transfer. After light fine-tuning, parallel hybrids achieved near-perfect accuracy (98%) and recall (1.00). Statistical analyses showed no significant difference between the parallel and sequential models (McNemar’s test), while backbone selection played a key role in performance. The Wilcoxon test found no significant difference in recall and F1-score between image and patient-level evaluations, suggesting balanced performance across both levels. Sequential hybrids achieved up to 7× faster inference than parallel models on the MURA test set while maintaining similar GPU memory usage (3.7 GB). Both fusion strategies produced clinically meaningful saliency maps that highlighted relevant wrist regions. These findings present the first systematic comparison of CNN–ViT fusion strategies for wrist anomaly detection, clarifying trade-offs between accuracy, generalization, interpretability, and efficiency in clinical AI. Full article

Offshore wind turbines are rapidly scaling in size, which amplifies the need for credible integrated load analyses that consistently resolve the coupled dynamics among rotor–nacelle–tower systems and their support substructures. This study presents a comprehensive ultimate limit state (ULS) load assessment for a fixed jack-up-type substructure (hereafter referred to as K-wind) coupled with the IEA 15 MW reference wind turbine. Unlike conventional monopile or jacket configurations, the K-wind concept adopts a self-installable triangular jack-up foundation with spudcan anchorage, enabling efficient transport, rapid deployment, and structural reusability. Yet such a configuration has never been systematically analyzed through full aero-hydro-servo-elastic coupling before. Hence, this work represents the first integrated load analysis ever reported for a jack-up-type offshore wind substructure, addressing both its unique load-transfer behavior and its viability for multi-MW-class turbines. To ensure numerical robustness and cross-code reproducibility, steady-state verifications were performed under constant-wind benchmarks, followed by time-domain simulations of standard prescribed Design Load Case (DLC), encompassing power-producing extreme turbulence, coherent gusts with directional change, and parked/idling directional sweeps. The analyses were independently executed using two industry-validated solvers (Deeplines Wind v5.8.5 and OrcaFlex v11.5e), allowing direct solver-to-solver comparison and establishing confidence in the obtained dynamic responses. Loads were extracted at the transition-piece reference point in a global coordinate frame, and six key components (Fx, Fy, Fz, Mx, My, and Mz) were processed into seed-averaged signed envelopes for systematic ULS evaluation. Beyond its methodological completeness, the present study introduces a validated framework for analyzing next-generation jack-up-type foundations for offshore wind turbines, establishing a new reference point for integrated load assessments that can accelerate the industrial adoption of modular and re-deployable support structures such as K-wind. Full article

Object detection in challenging imaging domains like security screening, medical analysis, and satellite imaging is often hindered by signal degradation (e.g., noise, blur) and spatial ambiguity (e.g., occlusion, extreme scale variation). We argue that many standard architectures fail by fusing multi-scale features prematurely, which amplifies noise. This paper introduces the Enhance-Fuse-Align (E-F-A) principle: a new architectural blueprint positing that robust feature enhancement and explicit spatial alignment are necessary preconditions for effective feature fusion. We implement this blueprint in a model named SecureDet, which instantiates each stage: (1) an RFCBAMConv module for feature Enhancement; (2) a BiFPN for weighted Fusion; (3) ECFA and ASFA modules for contextual and spatial Alignment. To validate the E-F-A blueprint, we apply SecureDet to the highly challenging task of X-ray contraband detection. Extensive experiments and ablation studies demonstrate that the mandated E-F-A sequence is critical to performance, significantly outperforming both the baseline and incomplete or improperly ordered architectures. In practice, enhancement is applied prior to fusion to attenuate noise and blur that would otherwise be amplified by cross-scale aggregation, and final alignment corrects mis-registrations to avoid sampling extraneous signals from occluding materials. Full article

Muscle growth is a critical determinant of meat yield and quality in livestock. Although follistatin (FST) is recognized as a key regulator of skeletal muscle development and fat metabolism, its specific function in geese remains largely unexplored. In this study, we identified two transcript variants of goose FST (gFST) in Zhedong White geese: gFST-X1 (1125 bp), encoding a 343-amino acid protein with a 28-amino acid signal peptide and four conserved domains, and gFST-X2, which contains a 243 bp insertion within the gFST-X1 transcript. RT-qPCR analysis revealed that gFST mRNA expression varied across tissues from female embryos (25 days), adults (70 days), and laying geese (270 days), as well as in skeletal muscle satellite cells (SMSCs) at embryonic day 16 (E16d). Overexpression of gFST in SMSCs resulted in 3596 differentially expressed genes (DEGs), including 2247 upregulated and 1349 downregulated genes (padj < 0.01). Key stemness markers (PAX7, PAX3) and myogenic regulators (MYOG, MYOD, MYF5) were significantly downregulated, whereas genes associated with lipid metabolism (PPARG, FABP5, ACSL5) and myosin-related processes (MYO1D, MYO1F, MYO1E) were markedly upregulated (padj < 0.01). Functional enrichment analysis linked these DEGs to the TGF-β, PPAR signaling, fatty acid metabolism, and Notch signaling pathways. These transcriptomic findings were further validated by qRT-PCR. Collectively, our results demonstrate the dual regulatory role of gFST in skeletal muscle development and provide new mechanistic insights into muscle development in geese. Full article

The medical imaging domain frequently encounters the dual challenges of annotation scarcity and class imbalance. A critical issue lies in effectively extracting information from limited labeled data while mitigating the dominance of head classes. The existing approaches often overlook in-depth modeling of sample relationships in low-dimensional spaces, while rigid or suboptimal dynamic thresholding strategies in pseudo-label generation are susceptible to noisy label interference, leading to cumulative bias amplification during the early training phases. To address these issues, we propose a semi-supervised medical image classification framework combining labeled data-contrastive learning with class-adaptive pseudo-labeling (CLCP-MT), comprising two key components: the semantic discrimination enhancement (SDE) module and the class-adaptive pseudo-label refinement (CAPR) module. The former incorporates supervised contrastive learning on limited labeled data to fully exploit discriminative information in latent structural spaces, thereby significantly amplifying the value of sparse annotations. The latter dynamically calibrates pseudo-label confidence thresholds according to real-time learning progress across different classes, effectively reducing head-class dominance while enhancing tail-class recognition performance. These synergistic modules collectively achieve breakthroughs in both information utilization efficiency and model robustness, demonstrating superior performance in class-imbalanced scenarios. Extensive experiments on the ISIC2018 skin lesion dataset and Chest X-ray14 thoracic disease dataset validate CLCP-MT’s efficacy. With only 20% labeled and 80% unlabeled data, our framework achieves a 10.38% F1-score improvement on ISIC2018 and a 2.64% AUC increase on Chest X-ray14 compared to the baselines, confirming its effectiveness and superiority under annotation-deficient and class-imbalanced conditions. Full article

Background/Objectives: The limited availability of labeled data, particularly in the medical domain, poses a significant challenge for training accurate diagnostic models. While deep learning techniques have demonstrated notable efficacy in image-based tasks, they require large annotated datasets. In data-scarce scenarios—especially involving rare diseases—their performance deteriorates significantly. Meta-learning offers a promising alternative by enabling models to adapt quickly to new tasks using prior knowledge and only a few labeled examples. This study aims to evaluate the effectiveness of representative meta-learning models for thoracic disease classification in chest X-rays. Methods: We conduct a comparative evaluation of four meta-learning models: Prototypical Networks, Relation Networks, MAML, and FoMAML. First, we assess five backbone architectures (ConvNeXt, DenseNet-121, ResNet-50, MobileNetV2, and ViT) using a Prototypical Network. The best-performing backbone is then used across all meta-learning models for fair comparison. Experiments are performed on the ChestX-ray14 dataset under a 2-way setting with multiple k-shot configurations. Results: Prototypical Networks combined with DenseNet-121 achieved the best performance, with a recall of 68.1%, an F1-score of 67.4%, and a precision of 0.693 in the 2-way, 10-shot configuration. In a disease-specific analysis, Hernia obtains the best classification results. Furthermore, Prototypical and Relation Networks demonstrate significantly higher computational efficiency, requiring fewer FLOPs and shorter execution times than MAML and FoMAML. Conclusions: Prototype-based meta-learning, particularly with DenseNet-121, proves to be a robust and computationally efficient approach for few-shot chest X-ray disease classification. These findings highlight its potential for real-world clinical applications, especially in scenarios with limited annotated medical data. Full article

Understanding the atomistic basis of multi-layer mechanisms employed by broadly reactive neutralizing antibodies of the SARS-CoV-2 spike protein without directly blocking receptor engagement remains an important challenge in coronavirus immunology. Class 4 antibodies represent an intriguing case: they target a deeply conserved, cryptic epitope on the receptor-binding domain yet exhibit variable neutralization potency across subgroups F1 (CR3022, EY6A, COVA1-16), F2 (DH1047), and F3 (S2X259). The molecular basis for this variability is not fully understood. Here, we employed a multi-modal computational approach integrating atomistic and coarse-grained molecular dynamics simulations, binding free energy calculations, mutational scanning, and dynamic network analysis to elucidate how these antibodies engage the receptor-binding domain (RBD) of the SARS-CoV-2 spike protein and influence its function. Our results reveal that neutralization efficacy arises from the interplay of direct interfacial interactions and allosteric effects. Group F1 antibodies (CR3022, EY6A, COVA1-16) primarily operate via classic allostery, modulating flexibility in RBD loop regions to indirectly interfere with the ACE2 receptor binding through long-range effects. Group F2 antibody DH1047 represents an intermediate mechanism, combining partial steric hindrance—through engagement of ACE2-critical residues T376, R408, V503, and Y508—with significant allosteric influence, facilitated by localized communication pathways linking the epitope to the receptor interface. Group F3 antibody S2X259 achieves potent neutralization through a synergistic mechanism involving direct competition with ACE2 and localized allosteric stabilization, albeit with potentially increased escape vulnerability. Dynamic network analysis identified a conserved “allosteric ring” within the RBD core that serves as a structural scaffold for long-range signal propagation, with antibody-specific extensions modulating communication to the ACE2 interface. These findings support a model where Class 4 neutralization strategies evolve through the refinement of peripheral allosteric connections rather than epitope redesign. This study establishes a robust computational framework for understanding the atomistic basis of neutralization activity and immune escape for Class 4 antibodies, highlighting how the interplay of binding energetics, conformational dynamics, and allosteric modulation governs their effectiveness against SARS-CoV-2. Full article

Early detection of breast cancer via mammography enhances patient survival rates, prompting this study to assess object detection models—Cascade R-CNN, YOLOv12 (S, L, and X variants), RTMDet-X, and RT-DETR-X—for detecting masses and calcifications across four public datasets (INbreast, CBIS-DDSM, VinDr-Mammo, and EMBED). The evaluation employs a standardized preprocessing approach (CLAHE, cropping) and augmentation (rotations, scaling), with transfer learning tested by training on combined datasets (e.g., INbreast + CBIS-DDSM) and validating on held-out sets (e.g., VinDr-Mammo). Performance is measured using precision, recall, mean Average Precision at IoU 0.5 (${\mathrm{mAP}}_{50}$), and F1-score. YOLOv12-L excels in mass detection with an ${\mathrm{mAP}}_{50}$ of 0.963 and F1-score up to 0.917 on INbreast, while RTMDet-X achieves an ${\mathrm{mAP}}_{50}$ of 0.697 on combined datasets with transfer learning. Preprocessing improves ${\mathrm{mAP}}_{50}$ by up to 0.209, and transfer learning elevates INbreast performance to an ${\mathrm{mAP}}_{50}$ of 0.995, though it incurs 5–11% drops on CBIS-DDSM (0.566 to 0.447) and VinDr-Mammo (0.59 to 0.5) due to domain shifts. EMBED yields a low ${\mathrm{mAP}}_{50}$ of 0.306 due to label inconsistencies, and calcification detection remains weak (${\mathrm{mAP}}_{50}$ < 0.116), highlighting the value of high-capacity models, preprocessing, and augmentation for mass detection while identifying calcification detection and domain adaptation as key areas for future investigation. Full article

Helicobacter pylori (Hp), a Class I carcinogen infecting over 50% of the global population, is increasingly resistant to conventional antibiotics. This study presents an AI-engineered probiotic strategy targeting urease, a key Hp virulence factor. A humanized single-domain antibody (UreBAb), previously identified and selected in our laboratory, was synthesized commercially and modeled using AlphaFold2, with structural validation conducted via SAVES 6.0. Molecular docking (PyMOL/ClusPro2) and binding energy analysis (InterProSurf) identified critical urease-active residues: K40, P41, K43, E82, F84, T86, K104, I107, K108, and R109. Machine learning-guided optimization using mCSA-AB, I-Mutant, and FoldX prioritized four mutational hotspots (K43, E82, I107, R109), leading to the generation of nine antibody variants. Among them, the I107W mutant exhibited the highest activity, achieving 65.6% urease inhibition—a 24.95% improvement over the wild-type antibody (p < 0.001). Engineered Escherichia coli Nissle 1917 (EcN) expressing the I107W antibody significantly reduced gastric HP colonization by 4.42 log10 CFU in the treatment group and 3.30 log10 CFU in the prevention group (p < 0.001 and p < 0.05, respectively), while also suppressing pro-inflammatory cytokine levels. Histopathological (H&E) analysis confirmed that the I107W antibody group showed significantly enhanced mucosal repair compared to wild-type probiotic-treated mice. Notably, 16S rRNA sequencing revealed that intestinal microbiota diversity and the abundance of core microbial species remained stable across different ethnic backgrounds. By integrating AI-guided antibody engineering with targeted probiotic delivery, this platform provides a transformative and microbiota-friendly strategy to combat antibiotic-resistant Hp infections. Full article

Human methyltransferase-like protein 16 (METTL16) installs N⁶-methyladenosine on U6 small nuclear RNA (snRNA) and other RNAs. Multiple X-ray crystal structures of METTL16 have been published; however, we do not yet fully understand the structure–function relationships of specific residues. We designed 38 mutants, including seven cancer-associated mutants, and used electrophoretic mobility shift assays and single-turnover kinetic assays to better understand the functional roles of specific domains and amino acid residues in binding to U6 snRNA, formation of the METTL16•U6 snRNA•S-adenosylmethionine (SAM) complex, and the rate of methylation. While point mutations in the methyltransferase domain mildly weaken the binding of METTL16 to U6 snRNA, the C-terminal vertebrate conserved regions (VCRs), particularly the arginine-rich region (R382 to R388), mediate cooperative binding and contribute more to RNA binding. All METTL16 K-loop mutants displayed tighter SAM binding, suggesting that the K-loop blocks SAM binding. In addition, residues E133 and F227 are critical for stabilizing SAM binding. Mutations in the ₁₈₄NPPF₁₈₇ catalytic core and R282A abolished methyltransferase activity. Two METTL16 somatic cancer-associated mutants (G110C and R241Dfs*2) displayed reduced methylation activity. This mutational analysis expands our understanding of how specific domains and residues contribute to substrate-binding activity and methylation of U6 snRNA catalyzed by METTL16. Full article

In the rapidly evolving landscape of digital marketing and electronic commerce, short-form content—particularly on platforms like Twitter (now X)—has become pivotal for real-time branding, community engagement, and product promotion. The rise of Non-Fungible Tokens (NFTs) and Web3 ecosystems further underscores the need for domain-specific, engagement-oriented social media content. However, automating the generation of such content while balancing linguistic quality, semantic relevance, and audience engagement remains a substantial challenge. To address this, we propose RL-TweetGen, a socio-technical framework that integrates instruction-tuned large language models (LLMs) with reinforcement learning (RL) to generate concise, impactful, and engagement-optimized tweets. The framework incorporates a structured pipeline comprising domain-specific data curation, semantic classification, and intent-aware prompt engineering, and leverages Parameter-Efficient Fine-Tuning (PEFT) with LoRA for scalable model adaptation. We fine-tuned and evaluated three LLMs—LLaMA-3.1-8B, Mistral-7B Instruct, and DeepSeek 7B Chat—guided by a hybrid reward function that blends XGBoost-predicted engagement scores with expert-in-the-loop feedback. To enhance lexical diversity and contextual alignment, we implemented advanced decoding strategies, including Tailored Beam Search, Enhanced Top-p Sampling, and Contextual Temperature Scaling. A case study focused on NFT-related tweet generation demonstrated the practical effectiveness of RL-TweetGen. Experimental results showed that Mistral-7B achieved the highest lexical fluency (BLEU: 0.2285), LLaMA-3.1 exhibited superior semantic precision (BERT-F1: 0.8155), while DeepSeek 7B provided balanced performance. Overall, RL-TweetGen presents a scalable and adaptive solution for marketers, content strategists, and Web3 platforms seeking to automate and optimize social media engagement. The framework advances the role of generative AI in digital commerce by aligning content generation with platform dynamics, user preferences, and marketing goals. Full article

This paper mainly studies the existence of sign-changing solutions for the following $p\left(x\right)$-biharmonic Kirchhoff-type equations: $-\left(a+b{\int }_{{\mathbb{R}}^N}{\displaystyle \frac{1}{p\left(x\right)}}{|\Delta u|}^{p\left(x\right)}\mathrm{d}x\right){\Delta }_{p\left(x\right)}^{2}u+V\left(x\right){|u|}^{p\left(x\right)-2}u$ = $K\left(x\right)f\left(u\right),x\in {\mathbb{R}}^N$, where ${\Delta }_{p\left(x\right)}^{2}u=\Delta \left({|\Delta u|}^{p\left(x\right)-2}\Delta u\right)$ is the $p\left(x\right)$ biharmonic operator, $a,b>0$ are constants, $N\ge 2$, $V\left(x\right),K\left(x\right)$ are positive continuous functions which vanish at infinity, and the nonlinearity f has subcritical growth. Using the Nehari manifold method, deformation lemma, and other techniques of analysis, it is demonstrated that there are precisely two nodal domains in the problem’s least energy sign-changing solution $u_b$. In addition, the convergence property of $u_b$ as $b\to 0$ is also established. Full article