Search Results (376)

Node classification is a central graph data mining task, yet repeated message passing can over-smooth representations and degrade frozen graph neural network (GNN) predictions after deployment. We present SETTA (Spectral-Energy Test-Time Adaptation), a prediction-level graph test-time adaptation framework that refines frozen outputs without test labels, gradients, parameter updates, or learnable adaptation parameters. SETTA denoises features for semantic-neighbor construction, adds complementary semantic routes while preserving observed edges, monitors a smoothness-energy proxy during diffusion, and accepts refinements through entropy-based gating. Configurations are fixed by a dataset-level protocol or selected using validation data only. Across six mostly homophilic benchmarks with 2708–19,717 nodes, SETTA improved a frozen two-layer GCN on every dataset and achieved the highest mean accuracy among the evaluated methods on five, with gains of 4.61, 3.08, and 2.01 percentage points on Cora, CiteSeer, and PubMed, respectively. Positive mean gains were also observed across all 30 dataset–backbone settings. Ablations and transition analyses indicate that semantic injection is most beneficial on sparse citation graphs and that selective refinement limits harmful changes. The current dense implementation supports benchmark-scale, amortized refinement; scalability and robustness on heterophilic graphs remain open. Full article

During the COVID-19 pandemic, healthcare systems experienced significant disruption, increasing the risk of multidrug-resistant (MDR) pathogen transmission. Acinetobacter baumannii, a critical-priority MDR pathogen, is known for its ability to persist in hospital environments and rapidly acquire resistance. To investigate the genomic characteristics, antimicrobial resistance determinants, and phylogenetic relationships of outbreak-associated Acinetobacter baumannii isolates, whole-genome sequencing (WGS) and comparative genomic analyses on 24 clinical and environmental strains collected during the COVID-19 period were performed. Twenty-four A. baumannii isolates collected between August 2020 and February 2021 from clinical and environmental samples were analyzed by WGS. All isolates displayed an MDR phenotype, with uniform resistance to carbapenems and aminoglycosides, and preserved colistin susceptibility. One environmental strain showed extreme drug resistance. WGS confirmed medium-quality genome assemblies and the clonal spread of a single A. baumannii lineage. Most resistance genes, including OXA-23, ADC-type β-lactamases, and ade efflux pumps, were chromosomally encoded and shared across all isolates. Plasmid-mediated resistance genes were variably distributed. This outbreak of MDR A. baumannii was driven by the clonal dissemination of a genomically stable lineage. Combined genomic and epidemiological analyses underscore the importance of integrated surveillance and environmental decontamination to prevent the spread of MDR pathogens. Full article

Indentation and crush tests are widely used as crash-relevant surrogates for mechanical intrusion in Lithium-Ion Batteries (LIBs), yet their interpretation still relies on macroscopic, binary pass/fail criteria established at the Beginning of Life (BoL). This neglects how calendar and cyclic degradation alter a cell’s mechanical, thermal, and structural state before abuse, leaving the literature on aged cells apparently contradictory. This review establishes an aging-aware interpretive framework that separates three levels commonly conflated: the boundary conditions governing deformation, the study-specific criteria and evidence strength behind a declared electrical event, and the post-trigger hazard progression captured within the available monitoring window. A descriptive synthesis of 22 normalized fresh-to-aged comparisons from 13 primary studies shows no universal aging-driven shift of the reported comparison event toward earlier or later force–displacement states, with force and displacement changing in the same direction in only about half of the cases. An apparent format-associated displacement pattern is substantially attenuated when the pouch-cell displacement ratios are normalized by the respective initial cell thicknesses: the pouch-cell median displacement ratio decreases from 1.07 to approximately 1.00, and the remaining variation differs across aging routes and states of charge, although their individual contributions cannot be separated in the available dataset. Many apparent contradictions likewise stem from non-equivalent mechanical boundary conditions and trigger definitions, and continuity of confinement—especially for pouch cells aged under stack pressure and tested in a relaxed state—is identified as a frequently overlooked interpretive variable, although no controlled aged-cell study has yet isolated its contribution. An earlier-reported trigger does not necessarily imply a more severe outcome. We distill these requirements into a tier-resolved reporting checklist, providing a more rigorous and reusable basis for abuse-test interpretation, model development, and safety qualification of batteries intended for long service life and second-life deployment. Full article

Sea buckthorn (Hippophae rhamnoides L.), known as a “superfruit” due to its complex nutritional profile, exhibits significant variations in bioactive compounds depending on variety and ripening stage. This study evaluates the ripening dynamics of four specific varieties—Dora, Cora, Clara, and Mara—cultivated in the Republic of Moldova to identify the optimal harvest window and technological potential for each variety. Over a seven-week period (August–September), several physicochemical and biochemical parameters were monitored, including pH, titratable acidity, total soluble solids (TSS), total dry matter (TDM), lipid content, color, hardness, vitamin C, total carotenoids content (TCC), and organic acids. Analytical methods such as refractometry, spectrophotometry, and capillary electrophoresis were employed to track these components. A downward trend in TSS and titratable acidity, characteristic of non-climacteric fruits, was observed. The Cora and Mara varieties exhibited the highest productivity and total dry matter content (up to 35.61%), making them suitable for concentrated food products. Clara stood out with the highest lipid accumulation (reaching 2.32%), identifying it as the premium variety for oil extraction. Vitamin C levels generally decreased across all varieties during ripening, with the Mara variety maintaining the highest concentrations (up to 518 mg/100 g). Conversely, total carotenoids showed a peak during full maturity (late August) for most varieties, with Dora reaching the highest values (40.73 mg/100 g). The present study highlights that the optimal harvest time for maximizing bioactive compounds is between late August and early September. Full article

Background: Liver cancers are a major cause of morbidity and mortality in patients where effective, non-invasive treatment options remain limited. Objective: Histotripsy is a non-invasive, non-thermal, image-guided focused ultrasound method of ablation that mechanically disrupts cells and offers a range of potential advantages over other ablation modalities. The lack of physiologically and anatomically relevant animal models of human liver cancer has significantly hindered biomedical device development, including histotripsy. Methods: To address these limitations, we developed a clinically relevant large animal orthotopic, dual-tumor model of human liver cancer and utilized these unique animals to evaluate the safety and efficacy of histotripsy. Here, we utilized immunocompromised pigs with genetic modifications in their IL-2RG and RAG2 genes and orthotopically engrafted human hepatocellular carcinoma (HepG2/C3A) and pancreatic adenocarcinoma (Panc-1) cells within the liver. The models were designed to recapitulate primary and metastatic liver tumor phenotypes. Results: Histotripsy enabled real-time visualization of the treatment by the formation of bubble clouds and accurate targeting of the lesions. Histological analysis confirmed the engraftment of tumor cells and the ablation of targeted tissue. Serum biomarkers demonstrated no significant differences in bilirubin, ALT, ALKP, or CK post-treatment, suggesting that histotripsy treatment was well tolerated with minimal hepatic dysfunction or hepatocellular injury. Conclusions: These findings establish a novel, clinically relevant porcine model of primary and metastatic liver tumors and demonstrate the safety and feasibility of using this model for evaluating histotripsy as a noninvasive modality for precise tumor ablation. Full article

Background/Objectives: Clinical outcomes after traumatic brain injury (TBI) remain difficult to predict, highlighting the need for more sensitive diagnostic and prognostic biomarkers, particularly in polytrauma. This study aimed to identify TBI-specific proteomic signatures in plasma and extracellular vesicles (EV)-enriched fractions of critically injured trauma patients. Methods: Seventy-five severely injured adult trauma patients (ISS ≥ 16) were included: isolated severe TBI (TBI; AIS_head ≥ 4, other AIS ≤ 1, n = 23), polytrauma with TBI (PT-TBI; AIS_head ≥ 4, n = 22), and polytrauma without TBI (PT; AIS_head = 0, n = 30). 24 age- and sex-matched healthy volunteers served as controls. Neat plasma and EV-enriched fractions were profiled using HPLC-MS/MS. Differentially expressed proteins (DEP) were analyzed bioinformatically, and associations with clinical parameters were assessed using Spearman’s correlation. Results: The EV-enriched plasma fraction yielded more DEPs than neat plasma (846 vs. 258). Most DEPs were linked to polytrauma, with plasma reflecting metabolism and EVs showing translation and protein catabolism signatures. Among TBI-context proteins, EV-associated NRCAM and AQR and plasma IGHV1-69D, MASP1, PON1, and IGFBP7 remained significantly associated with TBI after adjustment for confounder. Notably, only EV-associated proteins correlated with injury-related clinical parameters. AQR negatively correlated with GCS (r = −0.51, p < 0.0001), while elevated NRCAM, AQR, and plasma MASP1 were associated with neurological deterioration and neurosurgical intervention. Conclusions: EV-enriched plasma proteomics enhances biomarker discovery in polytrauma, revealing distinct pathways and greater sensitivity than whole plasma. While most changes reflect systemic injury, EV-associated NRCAM and AQR, along with plasma MASP1, show TBI-specific associations with neurological status, highlighting their potential as candidate TBI biomarkers for further study. Full article

Graph Neural Networks (GNNs) have become essential for analyzing graph-structured data, yet their deployment on resource-constrained edge devices is severely limited by high computational complexity and irregular memory access patterns. Here, we introduce DynamiGraph, a specialized FPGA-based overlay accelerator engineered for ultra-low-latency GNN inference in edge computing scenarios. Unlike general-purpose accelerators that incur high resource overhead to support a broad range of operators, DynamiGraph adopts a streamlined architecture focusing exclusively on essential General Matrix Multiplication (GEMM) and Sparse–Dense Matrix Multiplication (SpDMM) kernels. We implement a hardware-native runtime optimization mechanism that dynamically exploits graph sparsity via an edge-centric execution flow, eliminating redundant computations without requiring complex static preprocessing. Experimental results on an AXU2CGA edge platform demonstrate that DynamiGraph achieves sub-millisecond inference latencies on small-scale benchmarks (e.g., Cora) and a peak throughput of 1467 inferences per second. Furthermore, our runtime sparsity exploitation yields over 2000× reductions in floating-point operations compared to dense equivalents. These findings indicate that trading off model generality for architectural specialization and runtime awareness offers an efficient architectural alternative for enabling real-time graph intelligence in power- and bandwidth-limited edge environments. Full article

Background: Corallopyronin A (CorA) is a bacterial RNA polymerase inhibitor that binds a site distinct from rifamycins, but its activity across the ESKAPEE pathogen panel and its translational potential remain incompletely defined. Methods: CorA activity was evaluated against representative ESKAPEE pathogens using broth microdilution assays ± polymyxin B nonapeptide (PMBN). Activity was benchmarked against rifampin (Rif). Resistance, cross-resistance, serum activity, in vivo efficacy in the Galleria mellonella wax moth larva model, and biofilm disruption were assessed. Results: CorA inhibited Acinetobacter baumannii (minimal inhibitory concentration [MIC] = 16–32 µg/mL), including MDR isolates, with susceptibility enhanced 4–32-fold by efflux disruption or membrane permeabilization. In contrast, most other Gram-negative ESKAPEE pathogens required PMBN for activity, while Gram-positive organisms were intrinsically susceptible. Rif was consistently more potent than CorA across the panel. Rif-resistant A. baumannii and Klebsiella pneumoniae remained fully susceptible to CorA, confirming the absence of cross-resistance. Fractional inhibitory concentration (FIC) index analysis revealed pharmacological indifference between CorA and Rif, with no synergy or antagonism detected. CorA activity was abolished in 50% serum conditions and was not restored by PMBN, consistent with serum sequestration; no efficacy was observed in a G. mellonella infection model at doses up to 20 mg/kg despite Rif demonstrating significant protection. Notably, CorA reduced established A. baumannii biofilms by ∼3–4 log colony-forming units per mL (CFU/mL) across concentrations ≥4× MIC after 24 h of treatment. Conclusions: CorA exhibits selective activity against A. baumannii and retains efficacy against Rif-resistant strains. While high serum binding reduces in vitro activity and efficacy was not observed in the G. mellonella model at a lower dose than in other in vivo models, established in vivo activity against other pathogens demonstrates that serum binding does not preclude therapeutic utility. These findings highlight both the translational considerations for CorA against ESKAPEE pathogens and potential niches for its application, including filarial nematodes, biofilm-associated infections and combination strategies. Full article

Personal growth is a central aspect of development and well-being during young adulthood, yet sexually diverse and gender-diverse (SGD; a more inclusive term for LGBTQ+) young adults navigate this process within unique contexts shaped by identity, stress, and varying levels of social support. Despite growing visibility and social recognition of SGD identities in the United States, SGD individuals continue to face prejudice and discrimination, which negatively affects their physical and psychological health. Research indicates that stigmatized and marginalized populations with greater psychosocial resources are better able to cope with identity-related stressors. Yet, scholarship on coping with stigma and discrimination remains largely disconnected from research on social support, personal growth, and thriving within close relationships. The present observational study of SGD young adults (N = 400) examines how identity-affirming support from close others contributes to positive well-being outcomes, specifically personal growth, self-concept clarity, and thriving. Whereas much prior work focuses on how support buffers stress, we examine its role across stressors and opportunities for growth, experienced broadly and in relation to SGD identity. Our findings underscore the critical role that close relationships play in fostering social safety and personal growth for SGD young adults navigating identity development. Full article

Graph Neural Networks (GNNs) often suffer from degree bias in node classification tasks, where prediction performance varies across nodes with different degrees. Several approaches, which adopt Graph Contrastive Learning (GCL), have been proposed to mitigate this bias. However, the limited number of positive pairs and the equal weighting of all positives and negatives in GCL still lead to low-degree nodes acquiring insufficient and noisy information. This paper proposes the Hardness Adaptive Reweighted (HAR) contrastive loss to mitigate degree bias. It adds more positive pairs by leveraging node labels and adaptively weights positive and negative pairs based on their learning hardness. In addition, we develop an experimental framework named SHARP extending HAR to a broader range of scenarios. Both our theoretical analysis and experiments validate the effectiveness of SHARP. Across four datasets, SHARP outperforms baselines in 14 of 16 settings, improving global accuracy on Cora by 3.6% and 4.2% under the GCN and GAT backbones, while at the degree level it raises accuracy for the lowest-degree nodes by over 10%, confirming that the gains are targeted at the low-degree nodes most affected by degree bias. Full article

Accurate ionization energies are essential for understanding the electronic structures of atoms and molecules and benchmarking quantum-chemical methods. We report the calculated ionization energies of the H, C, N, O, F, P, and S atoms using several quantum-chemical approaches, aiming at reproducing the experimental values within chemical accuracy. The methods include the electron propagator approximations OVGF and P3+, the coupled-cluster methods CCSD(T), CCSDT, and IP-EOM-CCSD, and the composite methods G3 and CBS-QB3. The CCSD(T), CCSDT, G3, and CBS-QB3 methods, together with the DFT method with B2PLYP density functional and several post-Hartree–Fock methods, were used in conjunction with the energy-difference approach. The coupled-cluster calculations were combined with the aug-cc-pVXZ-DK, aug-cc-pVXZ, and ANO-RCC basis sets, all-electron correlation, DKH2 scalar relativistic corrections, atomic spin–orbit corrections, and extrapolation to the complete basis set (CBS) limit. The OVGF and P3+ methods do not achieve chemical accuracy on average, while CCSD(T) and CCSDT combined with the aug-cc-pVXZ-DK basis set and CBS extrapolation reach mean absolute errors of 0.0030 and 0.0026 eV, respectively, an order of magnitude below the chemical accuracy threshold. CCSD(T)/aug-cc-pVXZ-DK with CBS extrapolation provides the best compromise between accuracy and computational cost and can be used as a reference atomic benchmark for these ionization energies. Full article

Accurate tire–soil traction prediction is critical for agricultural and off-road vehicle design, yet rigorous comparisons of advanced discretization strategies under controlled-slip conditions remain limited. This study compares MM-ALE and Hybrid FE-SPH (H-SPH) discretization in LS-DYNA for SRTT (225/60R16) traction prediction on sandy loam (0.4% gravimetric moisture content) across 5–40% slip ratios. A CT-scan-based tire model using Yeoh visco-hyperelastic rubber (Material_2) was validated against experimental data, achieving CORA scores of 0.989 (radial deflection), 0.999 (loaded radius), 0.947 (footprint area), and 0.985 (contact pressure), outperforming the Mooney–Rivlin formulation (Material_1; CORA = 0.618). Soil moisture content (0.4%, 8%, 14%) was included as a design variable through a Latin Hypercube Sampling framework. Both methods reproduced a monotonic increase in traction; inter-method differences ranged from 29 to 36% at low slip, converging to a 7.8% coefficient of variation at 40% slip. A 27-run full-factorial DOE-I identified normal load as the dominant traction driver (90.1%), followed by velocity (8.6%) and inflation pressure (1.3%). An LHS-based DOE-II revealed moisture content as the primary driver of traction coefficient (67.5%), via a non-monotonic cohesion mechanism peaking at 8% gravimetric moisture. H-SPH reduced runtime by 38% versus MM-ALE. The validated framework provides reusable traction prediction protocols for variable conditions. Full article

Despite their reduced viability, kidneys from donors-after-circulatory-death (DCD) increase the pool of transplantable kidneys. Molecular hydrogen (H₂) is emerging as a gas with therapeutic potential against graft injury. We investigated the effect of H₂ in an ex vivo porcine model of DCD kidney transplantation. Renal arteries of male Yorkshire pigs (n = 6) were clamped in situ for 60 min to induce ischemia, and ureters and arteries were cannulated to mimic DCD kidney injury. Upon nephrectomy, kidneys were flushed with UW solution or H₂-saturated UW solution and then preserved by machine perfusion at 4 °C for 4 h followed by a 4-h reperfusion period with warm autologous blood. Urine and arterial blood samples were collected hourly. H₂ preserved renal architecture, evidenced by significantly reduced tubular necrosis and renal expression of damage markers, which corresponded with the downregulated renal expression of pro-inflammatory genes compared to the UW-only group (p < 0.05). H₂ also markedly reduced levels of serum creatinine, BUN and intrarenal resistance, while flow rate, creatinine clearance and urine output were significantly higher, which positively correlated with Trx-1 and HO-1 expression in comparison with UW only group (p < 0.05). Improvement in renal graft quality and function is associated with Trx-1/HO-1 activation, suggesting preliminary clinical trials in kidney transplantation. Full article

Graph Neural Networks (GNNs) have achieved remarkable progress in community detection, which is an essential topic in network analysis with the aim of dividing a network into multiple subgraphs to mine potential information. However, most existing GNN-based community detection approaches adopt static loss function weights during the training process. In this paper, we propose an unsupervised end-to-end community detection framework and define an adaptive loss weighting layer within this framework, named QALW, which is capable of learning an optimal combination of loss weights during model training to balance reconstruction loss and clustering loss. Experimental results obtained based on three real-world benchmark datasets (Cora, Citeseer, and Pubmed) demonstrate that QALW achieves effective and stable community detection performance compared with eight representative baseline methods. In particular, QALW improves ACC by 5.7% over the strongest baseline on the Cora dataset, and achieves ACC values of 64.9%, 61.7%, and 63.9% on Cora, Citeseer, and Pubmed, respectively. Furthermore, the results verify that the proposed dynamic scheduling mechanism effectively alleviates gradient conflicts and enables more stable optimization than fixed-weight strategies. Overall, QALW demonstrates promising competitiveness and good robustness for unsupervised community detection in attributed networks. Full article

Graph neural networks (GNNs) are highly effective on relational data, yet their performance degrades sharply when graph topology is poisoned before training. Existing defenses usually assume a fixed attack pattern and a fixed graph structure, which makes them brittle when the poisoned graph changes across attacks, perturbation budgets, or deployment conditions. We propose GADD, a game-inspired adversarial distillation framework for robust graph defense. GADD first constructs multiple positive and negative graph views through a homophily-aware graph sampling scheme, allowing the model to learn from both purified and high-risk subgraphs. It then trains a heterogeneous group of student GNNs online, where each student receives global class-distribution knowledge from its peers and local structural knowledge through an adversarial cyclic distillation objective. Finally, GADD replaces uniform ensembling with an entropy-regularized adaptive aggregation rule that assigns graph-adaptive weights according to confidence and inter-model agreement. On Cora, CiteSeer, and PubMed, GADD consistently improves robustness against both Meta and Nettack attacks while preserving clean accuracy. Under the strongest Meta and Nettack settings in the main benchmark, GADD improves the best competing baseline by up to 2.99 and 3.42 percentage points, respectively. Additional ablations show that graph sampling, adversarial distillation, and adaptive aggregation all contribute materially to the final robustness gains. Full article