Browse Articles

Mitigating shoulder spoofing vulnerabilities in mobile payment systems: a security framework

Pathogenic bacterial species and the microbiome of cat fleas (Ctenocephalides felis) inhabiting flea-infested homes

Background Ctenocephalides felis is a common ectoparasite of dogs and cats and can transmit a variety of pathogens including Bartonella and Rickettsia species. These bacteria, along with the known endosymbiont Wolbachia , are well-documented members of the C. felis microbiome, but species-level information is limited. Additionally, little is known about the variation in the C. felis microbiome in fleas from different sources and when different sequencing methods are applied to the same samples. Objective This study aimed to characterize the flea microbiome using both short-read (V3/V4) and long-read (full-length) 16S rRNA gene sequencing, determine whether long-read sequencing improves species-level identification especially in known pathogenic genera, and evaluate differences in microbial composition between fleas collected from cats, dogs, and environmental traps. Methods Fleas were collected from cats, dogs, and traps in flea-infested homes in Florida, pooled by source, and sequenced using short- (V3/V4) and long-read (full-length) 16S rRNA gene sequencing. Microbial prevalence and abundance were compared across sequencing approaches. Community composition was evaluated for differences between sources and houses. Candidate members of the flea microbiome were identified based on a combination of prevalence, abundance, and statistical signatures of potential contaminant origin. For Rickettsia and Bartonella , species-level taxonomic assignments were refined using a phylogenetic approach. Results Wolbachia , Rickettsia , and Bartonella were the most prevalent and abundant taxa. Spiroplasma was identified as a fourth core member of the flea microbiome. Long-read sequencing enabled better, but not perfect, species-level classification of Bartonella and Rickettsia compared to short-read sequencing. Important relationships between specific ASVs and flea sources were identified, for example fleas from cats harbored higher abundances of B. clarridgeiae and B. henselae than fleas from traps.

Investigation of the relationship between carotid intima-media thickness and angiopoietin-like factor 3 levels in metabolic dysfunction-associated steatotic liver disease

Graph former-CL: A novel graph transformer with contrastive learning framework for enhanced drug-drug interaction prediction

Drug-drug interactions (DDI) represent a significant clinical challenge in modern healthcare, contributing to over 125,000 deaths annually in the United States alone. Current computational approaches face substantial limitations in capturing long-range molecular dependencies and generalizing to novel drug combinations. Traditional Graph Neural Networks (GNNs) suffer from over-smoothing and locality bias, while sequence-based methods fail to adequately represent three-dimensional molecular structures. To address these limitations, we propose Graph Former-CL, a novel deep learning framework that synergistically combines Graph Transformer architecture with contrastive learning for DDI prediction. Our approach features four key innovations: (1) a hierarchical Graph Transformer with position-aware multi- head self-attention to capture both local and global molecular patterns, (2) a domain-specific contrastive learning module with molecular augmentation strategies, (3) a cross-modal fusion mechanism integrating SMILES sequences with graph representations, and (4) an adaptive pooling strategy for multi-scale molecular representation. Comprehensive evaluation on four benchmark datasets demonstrates superior performance, with Graph Former-CL achieving 98.2% accuracy on DrugBank and 89.4% on TWOSIDES, both representing statistically significant improvements (p < 0.001) over state-of-the-art methods. Notably, the framework achieves 85.6% accuracy for novel drugs in inductive settings, demonstrating robust generalization capabilities essential for real-world clinical applications.

Complex network analysis of mobility dynamics in Seoul during the COVID-19 pandemic 2020–2022

In vitro assessment of antibacterial and biocompatibility properties of a poly-ε-lysine and hyaluronic acid contact-killing coating to prevent prosthetic joint infection

Prosthetic joint infection (PJI) is a major adverse outcome following total hip and knee arthroplasties. With the rise of antimicrobial resistance, there is a risk of therapeutic insufficiency in treating PJI. This study determined the potential of a poly-ε-lysine and hyaluronic acid (PEL-10/HA144-1) contact-killing coating as a promising prevention of bacterial attachment. A broader development perspective was integrated through a Safe-by-Design approach by adding relevant in vitro tests for implant-host interactions. The PEL-10/HA144-1 coating was deposited on titanium alloy Ti6Al4V (Ti) and ultra-high-molecular-weight-polyethylene (UHMWPE). A combination of the ISO 22196, ASTM E2180-18, and JIS Z 2801 testing standards using Staphylococcus aureus and Escherichia coli demonstrated a bactericidal effect of the coating, up to a 5-log reduction compared to uncoated samples. A bacterial adhesion test showed a decrease in adherent bacteria up to 4-log after 4 h, and up to 5-log after 24 h between coated and uncoated samples. Saos-2 human osteoblast-like cells and L929 mouse fibroblasts exposed to 72 h extracts of the coating for 24 h showed in vitro cell viability of >70%, indicating no cytotoxicity according to ISO 10993−5. Furthermore, while initial osteoblast attachment to the coating appeared challenging, increased proliferation and metabolic activity over time were observed. After 14 and 21 days, no reduction in osteogenic marker expression was found on the coated samples compared to the uncoated samples. Overall, the PEL-10/HA144-1 coating reduced bacterial adhesion, was not cytotoxic to mammalian cells , and supported osteoblast function in vitro , making it a promising technique for future implantable orthopedic applications.

Computational framework and machine learning approach to fractional order soil helminth infections disease model for control mechanism

A novel approach for longitudinal analysis of serum biomarkers of joint metabolism and knee injury in military officers

Purpose To investigate the longitudinal relationships between serum biomarkers of joint metabolism, knee injury, and Knee Injury and Osteoarthritis Outcome Score (KOOS) using novel methodologies. Methods Data were collected from military officers who enrolled as cadets between 2004–2009, with follow-up conducted between 2015–2017. Analyses included 234 officers who had no history of knee ligament/meniscal injury at the time of military academy matriculation, had serum biomarker measurements at matriculation and graduation, demographic data, and KOOS assessment at follow-up. Biomarkers included Collagen Type II (C2C) and Type I and II (C1,2C) collagenase-generated cleavage epitopes, C-terminal propeptide of Type II collagen (CPII), and C- and N-terminal telopeptides of type I collagen (CTX and NTX). Angle-based Joint and Individual Variation Explained (AJIVE) was used to determine demographic determinants of biomarker levels and individual modes of variation specific to biomarker levels at matriculation and graduation, stratified by sex. Results We confirmed known associations of joint metabolism biomarkers with age in both sexes and with smoking in males. Matriculation biomarker data in males suggested a protective biomarker profile characterized by high cartilage synthesis and low cleavage of type I and II collagen in association with healthy KOOS scores at follow-up. CPII measured at matriculation was negatively associated with incident injuries after adjustment for smoking status (p = 0.03, logistic regression), confirming results from AJIVE. Conclusion These exploratory analyses suggest that CPII alone, or in combination with other joint metabolism biomarkers, may help identify individual risk of knee injury.

Balancing noise reduction and neural signature preservation in EEG biometrics

Abstract EEG-based subject identification is an emerging biometric approach with strong potential for secure authentication, but reliable performance requires optimisation of the entire processing pipeline. The key difficulty lies in improving signal quality while preserving the subtle neural signatures that uniquely distinguish individuals . In this study, we propose a complete framework that integrates lenient preprocessing, spectral feature extraction, and ensemble classification. Using the Brain Encoding Dataset(BED), we evaluated three data variants: raw EEG recordings, signals processed with a modified Pre-processing (PREP) pipeline using relaxed thresholds, and expert-curated pre-extracted features. All datasets were analyzed with mel-frequency cepstral coefficients(MFCC), and classification was performed within an ensemble architecture that combined decision trees, random forests, support vector machines, and XGBoost. The experiments covered 21 subjects, 33 sessions, and twelve stimulus conditions including resting state, cognitive tasks, and visual evoked potentials. XGBoost achieved peak accuracy of 98.00% using Visual Evoked Potential Complex stimulation at 10 Hz on cleaned data, representing a 5.3% improvement over raw signals and an 8.4% improvement over pre-extracted features. Statistical validation confirmed that these improvements are robust across all experimental conditions at ( $$p < 0.01$$ ). Cross-session evaluation further demonstrated the expected temporal variability in EEG-based biometrics but showed that the proposed pipeline improves robustness compared with both raw and conventionally processed data, with Rest Closed Eyes emerging as the most stable paradigm. These findings establish a principled framework for EEG-based subject identification and provide practical guidelines for optimizing preprocessing, feature extraction, classification, and stimulus paradigms for real-world deployment with consumer-grade hardware and system approach.

A randomized controlled trial in healthy participants to compare the insulinogenic effects of whey protein and pea protein co-ingested with glucose

Increasing protein content of foods is effective in reducing postprandial hyperglycaemia, but animal protein may exacerbate insulin sensitivity. This single-blind, randomised, crossover study compared the effects of co-ingesting glucose with 10 or 20 g whey protein and glucose with 10 or 20 g pea protein, with a reference product (glucose) on glycaemic and insulinaemic responses in 30 healthy individuals. Blood glucose and plasma insulin were measured at baseline, 15, 30, 45, 60, 90, 120, 150 and 180 minutes after product consumption. The trial was registered with Clinical Trials.gov (NCT04871971). Glucose incremental area under the curve (mmol/l*min) at 180 minutes was significantly reduced (p < 0.001) for glucose with 20 g pea protein (89.8 ± 51.6) and glucose with 20g whey protein (98.5 ± 58.0) compared to glucose (143.2 ± 74.0). Insulin incremental area under the curve at 180 minutes (µU/ml*min) for glucose with 20 g pea protein (4304.56 ± 1896.07) was significantly lower (p < 0.001) than glucose with 20g whey protein (6311.81 ± 3489.12). This study has shown a superior effect of pea protein over whey protein in reducing glycaemic response, without any excessive increase in insulinaemic response.

Dynamic graph convolution with comprehensive pruning and GNN classification for precise lymph node metastasis detection

Abstract Early and accurate detection of lymph node metastases is crucial for improving breast cancer patient outcomes. However, current clinical practices, including CT, PET imaging, and microscopic examination, are time-consuming and prone to errors due to low tissue contrast, varying lymph node sizes, and complex workflows. To address the limitations of existing approaches in lymph node segmentation, feature embedding, and classification, this study proposes a novel framework Graph-Pruned Lymph Node Detection Framework (GPLN-DF) that integrates a Dynamic Graph Convolution (DGC) autoencoder with Node Attribute-wise Attention (NodeAttri-Attention) for accurate lymph node segmentation. This segmentation is further refined using Comprehensive Graph Gradual Pruning (CGP) to reduce unnecessary parameters and computational costs. After segmentation, Hessian-based Locally Linear Embedding (HLLE) is applied for effective feature extraction and dimensionality reduction, preserving the geometric structure of lymph node regions. Finally, a Graph Neural Network (GNN) classifier enhanced with CGP is used to classify the segmented lymph nodes as metastatic or non-metastatic based on the extracted features. This comprehensive framework addresses challenges such as small lymph node size, shape variability, low contrast in medical imaging, and high computational burden. The model was evaluated on the CAMELYON17 dataset, achieving a classification accuracy of 98.65%, surpassing existing models in segmentation precision and classification performance.

Effectiveness of dyadic interventions in improving outcomes for adults with multiple long-term conditions and/or frailty and their informal carers: A systematic review protocol

Aim To synthesise current evidence on the effectiveness of dyadic (pair-based) interventions in improving outcomes for adults with multiple long-term conditions (MLTC) and/or frailty (aged ≥55 years) and their informal carers. Methods The review protocol followed the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P) guidelines, with the protocol registered with PROSPERO (CRD420251144604). MEDLINE, Embase, PsycINFO, CINAHL Plus, CENTRAL, ClinicalTrials.gov will be searched for experimental and quasi-experimental studies examining the effectiveness of community-based dyadic interventions for adults with MLTC (≥2 long-term conditions within an individual) and/or frailty (aged ≥55 years) and their informal carers (spouses/partners, other family members or relatives) published since 2010 and up to September 2025. Dyadic interventions will be defined as pair-based interventions that directly involve informal carers and care recipient adults with MLTC and/or frailty using various techniques targeted at carers, care recipients, or both to change outcomes for at least one member of the carer/care recipient pair (or dyad). Database searches will be followed by a manual search of the reference lists of included studies and lists of papers citing included studies in order to identify additional studies. Two reviewers will independently screen titles and abstracts against the selection criteria and independently screen full texts using Covidence software. Methodological quality will be assessed using the Cochrane Risk of Bias (RoB) 2.0 tool for experimental studies and the Risk Of Bias In Non-randomised Studies of Interventions (ROBINS-I) tool for quasi-experimental studies. Synthesis of evidence will be quantitative. If meta-analysis is not possible, we will follow Cochrane recommendations for quantitative Synthesis Without Meta-Analysis (SWiM) guidance. Conclusion The findings will address the evidence gap in dyadic implementation research in later life and help inform clinical decision-making, policy development and program planning for adults with MLTC and/or frailty and their carers, particularly in primary care and other community health settings.

Overcoming difficulties in segmentation of hyperspectral plant images with small projection areas using machine learning

Abstract Segmentation of hyperspectral image data is a well-established technique in remote sensing. While it is commonly applied to individual field crops, its use for individual trees is less prevalent. Conifers are crucial in forestry, and assessing physiological status, or genetic diversity is required for effective early-age treatment in nurseries and hyperspectral imaging (HSI) combined with high-throughput phenotyping (HTP) offers faster and non-destructive evaluation. NDVI-based thresholding is sufficient for detection of leaves with large projection areas, but needles of conifers present challenges due to spatial resolution constraints and increased proportion of border pixels. This study monitored the offspring of three locally adapted Scots pine ( Pinus sylvestris L.) populations, representing distinct upland and lowland ecotypes. This study presents a hyperspectral image processing pipeline for segmenting and isolating individual Scots pine seedlings. Using a K-means algorithm, 23 hyperspectral centroids were successfully derived and subsequently classified into ten biologically distinct groups. Random forest classification model effectively differentiated Scots pine seedlings based on origin during water stress and recovery periods. This study highlights the potential of hyperspectral imaging and machine learning in evaluating the physiological state of conifer seedlings, demonstrating promising applications in forest tree physiology research and tree breeding.

Evaluating the effect of the SMART intervention in people with recently diagnosed breast cancer who are being treated at a public tertiary hospital in Australia: protocol and statistical analysis plan for a single-blinded, single centre randomised controlled trial

Introduction Adults undergoing treatment for breast cancer (BC) are advised to participate in regular exercise. However, many struggle to exercise consistently due to the side effects of systemic treatments including nausea, fatigue, and pain. In adults with newly diagnosed BC, this trial will evaluate the effectiveness of a new exercise intervention, compared with usual care, on outcomes including health-related quality of life (HRQoL). Materials This randomised controlled trial is underway at an Australian tertiary hospital. The protocol was prospectively registered (Australian New Zealand Clinical Trials RegistryACTRN12623001168640p). Consenting adults with BC diagnosed within the prior six months, with planned chemotherapy and/or endocrine treatment will be randomised to an intervention or control group. Both groups receive usual physiotherapy and medical care. Those allocated to the intervention group are offered participation in the ‘SMART’ intervention (Self-determined, Monitored, Adaptable, Rehabilitation with Telehealth support). This involves 16-weeks of tailored, one-on-one physiotherapy-led exercise sessions including behaviour change techniques and the weekly goal of completing 150 minutes of aerobic exercise and two resistance training sessions. The primary outcome is HRQoL and secondary outcomes include physical assessments (muscle strength, exercise tolerance, body composition), healthcare utilisation, workplace absenteeism, mood, psychological determinants of behaviour change, chemotherapy completion rates and endocrine therapy completion. All outcomes are measured prior to randomisation and 16 weeks following randomisation. Additional assessments of all outcomes (excluding the physical assessment) occur at 8 weeks and 52 weeks following randomisation. Ongoing recruitment for two years from June 2024 is expected to achieve a sample size of 260. No results have been analysed. Discussion If the SMART intervention produces favourable change, this will support its adoption in clinical practice. A greater understanding of factors including BC stage, treatment type or variables relating to the exercise program, that influence the magnitude of exercise-induced change on HRQoL will inform future exercise programs.

Covalent modification of a glutamic acid inspired by HaloTag technology

Abstract For targeted covalent protein modification at low-reactivity aspartates and glutamates, new methods are in high demand. We report a technique inspired by the HaloTag technology, which employs nucleophilic substitution at chloroalkane-functionalised ligands by a specific aspartate residue. Embedding of alkyl bromide warheads into non-covalent inhibitors enables covalent modification of a glutamate in the lipoprotein binding chaperone - phosphodiesterase of retinal rod subunit delta (PDEδ), which shuttles prenylated lipoproteins between cellular membranes and thereby mediates their activity. Its hydrophobic ligand-binding pocket contains p.E88 as the only accessible nucleophile for covalent targeting. We show that a covalent inhibitor, termed DeltaTag, overcomes limitations of non-covalent inhibitors. DeltaTag labels PDEδ at its p.E88 under biologically relevant conditions, modulates mammalian target of rapamycin (mTOR) signalling by disrupting the PDEδ-Rheb (Ras homologue enriched in brain)-mTORC1 (mTOR complex 1) axis and inhibits cancer cell proliferation. This proof-of-concept study demonstrates that the design strategy holds promise for the covalent modification of proteins with lipophilic binding sites that lack accessible reactive amino acids but contain specific carboxylates.

Tibetan herbal medicine improved the health status of calves by regulating the antioxidant ability, inflammatory reaction and microbiota of female Yaks

Correction: Absence of ultimate controller and investment efficiency: Evidence from China

Engineered internal architecture of core-shell lipid nanoparticles promotes efficient mRNA endosomal release

Research on surrounding rock control technology of roof cutting and pressure relieving for roadside filling in gob-side entry retaining of large mining height panel

Households’ poverty and inequality after the COVID-19: Insights from panel data of face-to-face surveys in Southeast Asia

The global COVID-19 pandemic has had catastrophic impacts on global economies and human health. The consequences of the COVID-19 pandemic include the fatalities of millions of people and increased poverty. Given the limited evidence at the micro level and on the heterogeneous impacts of the COVID-19, we use panel data of 2,517 households from Thailand and Vietnam to investigate the correlation and heterogeneous effects of the COVID-19 on household income and poverty and to examine the distributional effects on household income. The results show that the COVID-19 has a negative correlation with household income, while it has a positive correlation with the Gini coefficient of income inequality and poverty. Our results also show that there are unequal effects of the COVID-19 on household income, inequality, and poverty at the micro level. Further, the COVID-19 has the highest negative impact on daily per capita income of households in the 10th and 25th quantile groups. Policy implications are proposed to support households in disadvantageous groups.