Browse Articles
Discover research articles across all indexed journals
Biologists pinpoint how common virus triggers multiple sclerosis
Identification of key clinical features and development of a machine learning model for screening-based detection of diabetic retinopathy
Abstract Diabetic retinopathy (DR) is a major cause of preventable visual impairment. Routinely collected clinical information may help prioritize patients for retinal assessment where immediate fundus examination is not universally available. Cross-sectional identification of DR at screening must, however, be distinguished from prediction of future disease progression. To develop and internally validate machine-learning models based on routine tabular clinical variables for identifying prevalent DR at community screening. This cross-sectional study included 1,741 adults with type 2 diabetes from a community screening program in Changzhi, China. The primary outcome was prevalent DR at screening, defined as mild non-proliferative DR or worse in one prespecified study eye. Fundus photographs were used only to establish the ophthalmologist-graded reference outcome; raw images and image-derived features were not used as predictors. Twenty-two routine clinical predictors were analyzed. Participants were divided at the patient level into a training set ( n = 1,219) and a held-out test set ( n = 522). Elastic-net logistic regression, random forest, and gradient boosting were compared using 10-fold cross-validation within the training set. Test-set evaluation included the area under the receiver operating characteristic curve (AUC), area under the precision-recall curve (AUPRC), Brier score, calibration, threshold-dependent performance, permutation importance, and a complete-known-case sensitivity analysis. DR was identified in 470 of 1,741 participants (27.0%). The random forest showed the highest training-set cross-validated AUC (mean 0.896, standard deviation 0.031) and achieved a test-set AUC of 0.889 (95% confidence interval 0.852–0.925), AUPRC of 0.834, and Brier score of 0.101. Elastic-net logistic regression and gradient boosting achieved test-set AUCs of 0.872 and 0.887, respectively, compared with 0.666 for an HbA1c-only logistic model. Observed DR prevalence was 9.5%, 45.6%, and 93.3% in the exploratory probability strata < 0.3, 0.3 to < 0.6, and > = 0.6. Hypertension, diabetic nephropathy, and diabetes duration ranked highest by test-set permutation importance. In the complete-known-case sensitivity analysis, the random forest AUC was 0.865. A random forest model based on routinely available clinical variables showed good internal discrimination for identifying prevalent DR at screening. Calibration was not perfect, and the model does not predict future DR progression. External validation, recalibration, and prospective evaluation are required before clinical implementation.
Smart packet prioritization in cognitive radio networks for smart agriculture
Abstract The quick evolution in the field of Smart Agriculture, along with the Internet of Things (IoT), has created the need for developing wireless communication paradigms with the ability to handle heterogeneous data for different applications with different degrees of priority. This paper proposes an autonomous packet priority management framework for IEEE 802.11af-based CRNs to meet the heterogeneous data demands of Smart Agriculture and IoT. By employing a Dueling Double Deep Q-Network (D3QN) with a new dynamic aging threshold, the model avoids data starvation while ensuring reliable transmission of high-priority agricultural data over TVWS. Simulation results demonstrate that the D3QN framework achieves significantly better performance compared to standard and DQN-based models, especially in highly congested conditions (λ > 8). The proposed scheme reduces average delay for high-priority packets by 16.1% compared to the Baseline and by 10.5% compared to standard DQN under high-load conditions, and achieves approximately 14.8% reduction in relative energy consumption per high-priority packet, while maintaining comparable throughput. These results demonstrate a more robust and energy-efficient solution for real-time intelligent farming environments.
GIS–AHP–based land suitability assessment for sustainable agricultural planning on Qeshm Island
Abstract Sustainable agricultural planning necessitates a holistic approach that considers ecological, economic, and social dimensions, balancing agricultural yield with long-term ecological sustainability. In the face of intensifying environmental pressures, such as climate change, land degradation, and biodiversity loss, agriculture must adopt adaptive land use strategies to ensure long-term viability. This paper explores how integrating landscape ecology principles into agricultural planning on Qeshm Island can enhance landscape resilience and efficiency to guide land use strategies that preserve biodiversity and sustain ecosystem services. A land suitability model was applied by converting socio-ecological factors (soil, slope, geology, water accessibility, land use and land cover, and sensitive habitats) into standardized raster layers, weighting them with AHP, and integrating them through a GIS-based weighted overlay. Suitability map was evaluated through accuracy assessment and sensitivity analysis to ensure reliability. The results showed high and very high suitability areas (18,628 ha) aligned with optimal conditions, while moderate zones (13,935 ha) required interventions such as erosion control and water management. Low-suitability (1,408 ha) and unsuitable areas (112,188 ha) faced significant constraints as steep topography and erosion-prone soils, limiting agricultural potential. The findings underscore the importance of interdisciplinary collaboration (e.g., agronomy, social, ecology, and geography) and the adoption of innovative planning tools to achieve sustainable agricultural landscapes.
Changes in the epidemiological patterns of SARS-CoV-2 in South Korea from 2020 to 2022: indirect evidence for the impact of non-pharmaceutical interventions
Highly fragmented European wetlands with uneven restoration needs
Impact of PM2.5 on common fruit species around Ghorahi Cement Factory, Dang, Nepal
Effects of aged and pristine microplastics on physical, chemical, and biological soil properties
Quality of life and psychological distress in newly diagnosed treatment‑naïve cancer patients and their determinants at a Tertiary Hospital in Ethiopia
Abstract Newly diagnosed cancer patients often experiencing significant psychological disruption and a decline in quality of life (QoL), which may lead to interruptions in treatment. Nevertheless, the psychosocial status of these patient remains inadequately understood in resource-limited settings. Therefore, the present study aimed to evaluate psychosocial vulnerability, specifically psychological distress and QoL among newly diagnosed, treatment naïve patients, as well as the factors associated with these outcomes. A cross-sectional study was conducted among 231 newly diagnosed cancer patients between June and December 2024 at the oncology Department of Jimma University Medical Center (JUMC). Psychological distress and QoL were assessed using the National Comprehensive Cancer Network (NCCN) distress thermometer (DT, version 1.2024) and the EORTC QLQ-C30, respectively. Data were analyzed using SPSS version 26. Multivariable logistic regression analysis was performed, and the strength of associations between dependent and independent variables was expressed as adjusted odds ratio (AORs) with 95% confidence intervals (CIs). Statistical significance was set at p-value < 0.05. More than half of the patients (51.95%: 95%CI 46%, 59%) experienced moderate-to-severe psychological distress. Higher odds of distress were observed among rural residents (AOR: 2.45), patients with co-morbidities (AOR: 3.59), those reporting low social support (AOR: 6.09), individuals with advanced-stage disease (AOR: 4.82), patients receiving palliative treatment intent (AOR: 2.32), those with ECOG performance status ≥ 2 (AOR: 4.15), and malnourished patients (AOR: 5.68). Notably, perceived low social support, advanced disease stage and malnutrition were common factors associated with both psychological distress and lower QoL scores. This study highlights the substantial burden of psychosocial vulnerability among newly diagnosed cancer patients, with more than half experiencing moderate-to-severe psychological distress. Significant differences in mean QoL scores were observed between early-and advanced-stage patients. Advanced stage disease, malnutrition, and low perceived social support were consistently associated with lower QoL scores. These findings underscore the urgent need to integrate psychosocial care with physical health care in oncology, particularly to address unrecognized psychosocial challenges in resource limited settings.
Mathematics formula found on Maya wall rivals insights of ancient masters
Quantitative correlation of spectroscopic signatures with ligand–protein interactions in anti-cancer drug Afinitor: an integrated experimental–computational study
Hepatitis B prevention practices and associated factors among medical and health sciences students at university of Gondar, Northwest Ethiopia
Evaluating conditioned place preference in calves 0 and 20 days after hot-iron disbudding
Abstract Hot-iron disbudding, a routine procedure that prevents horn bud growth through cauterization, is painful for calves. The resulting burns remain sensitive to touch for weeks, but it is unclear whether calves experience ongoing, non-evoked pain during healing. We evaluated conditioned place preference for analgesia in 44 calves disbudded or sham-disbudded 6 h (Day 0) or 20 days (Day 20) before testing (n = 11/treatment). Calves were conditioned to associate the effects of a lidocaine cornual nerve block with the location (left or right side of the home pen) and pattern of a visual stimulus (solid black or black-and-white striped board), and a control injection of saline with the contrasting stimulus. Calves received 6 conditioning trials over 3 days, alternating between lidocaine- and saline-paired stimuli. On the fourth day, calves were provided both stimuli for 5 min and preference was assessed. Contrary to our predictions, we found no evidence of a preference for either stimulus at 0 or 20 days after disbudding or sham-disbudding. Given that pain is well established in the hours after disbudding and some studies report behavioral changes weeks later, we suggest the conditioned place preference paradigm, as applied in this study, would benefit from further refinement.
Automated seaweed species classification using deep learning and large language models
Abstract Seaweeds are foundational components of marine ecosystems and hold significant value in global aquaculture. Accurate and efficient identification of seaweed species is critical for biodiversity monitoring, ecological research, and sustainable management. This study evaluates a dual-approach framework for automated seaweed classification. First, we implemented a standard convolutional neural network (CNN) based on the EfficientNet-B0 architecture, utilizing transfer learning from ImageNet. The model was trained on an augmented dataset of 3,440 images derived from an original collection of 800 field images (stratified 80/10/10 split) covering 43 seaweed species. The CNN achieved a baseline classification accuracy of 89%. Second, we present a proof-of-concept study using a vision-language model (VLM), specifically Claude 3.5 Sonnet , to explore semantic reasoning in taxonomic classification. The VLM achieved a raw accuracy of 70%, which was further improved to an effective accuracy of 92% on a selected subset through a human-in-the-loop (HITL) validation system. While the CNN provides a robust and rapid classification tool, the VLM-HITL approach offers an interpretable semantic alternative for validating challenging specimens. This work contributes to the development of scalable, automated biodiversity monitoring tools for marine science.
‘Dark comet’ unmasked by its mysterious motions
A high-throughput 3D conjunctival spheroid model for standardized in vitro testing
Abstract New Approach Methodologies (NAMs) based on human cells are increasingly needed to improve the physiological relevance, reproducibility, and ethical acceptability of preclinical ocular surface research. We developed a reproducible and scalable three-dimensional in vitro conjunctival spheroid model using primary human conjunctival epithelial cells and conjunctival fibroblasts as a human-relevant, scaffold-free test system. Agarose-based microwell arrays were fabricated via a combination of custom-made high-resolution 3D printing and polydimethylsiloxane (PDMS) replica molding, enabling the formation of uniform microwells suitable for spheroid culture and parallel production of size-controlled microtissues. Primary human conjunctival epithelial cells and fibroblasts were isolated from donor tissue obtained during routine ophthalmic surgeries and expanded under defined culture conditions. Fibroblast spheroids were first generated within agarose microwells at controlled seeding densities, resulting in stable and size-controlled aggregates. Subsequently, conjunctival epithelial cells were seeded onto pre-formed fibroblast spheroids to establish bilayered conjunctival spheroids that mimic native tissue organization. Spheroid development, morphology, and viability were systematically characterized using optical microscopy, live/dead assays, histological staining, immunofluorescence, and advanced imaging techniques including scanning electron microscopy (SEM), cryo-SEM, and transmission electron microscopy (TEM). Quantitative image analysis demonstrated consistent spheroid size and shape over time, while viability assays confirmed high cell survival. Histological and ultrastructural analyses revealed organized cellular architecture and extracellular matrix deposition, indicative of functional tissue-like constructs. By combining primary human cells, scaffold-free assembly, and microwell-based scalability, this platform contributes to the implementation of the 3Rs and provides a fit-for-purpose NAM for dry eye disease-related research, conjunctival inflammation studies, and early-stage preclinical compound testing.
ASC-YOLOv8n: Enhanced multi-scale feature fusion for accurate detection of four cigar appearance defects
Abstract The appearance defects of cigars can significantly compromise their overall quality, with detection currently relying mainly on manual inspection, a process that is time-consuming and inefficient. The ASC-YOLOv8n model has been proposed for high-precision automated defect detection in full-leaf handmade cigars, targeting defects such as green spots, holes, breaks, and tail breaks during production. This model incorporates several key improvements: it integrates an ASC (Adaptive Spatial Context) module to provide a more hierarchical receptive field, enhancing its ability to detect intricate defect patterns; replaces the conventional C2f module with a more advanced Fusion module, which combines multiple feature maps to improve feature extraction capabilities; and employs a novel WIoU (Weighted Intersection over Union) localization loss function, significantly refining defect localization precision. Experimental results show that the ASC-YOLOv8n model achieves a performance boost, with the mean average precision at an IoU threshold of 0.5 (mAP@0.5) increasing by 1.7% points to 94.00%, outperforming the baseline YOLOv8n model. This demonstrates the model’s effectiveness in accurately identifying critical cigar defects, making it a reliable and robust solution for intelligent cigar inspection, and contributing to enhanced quality control in cigar production.
Synthesis and microwave absorption property of polyaniline incorporated with different weight ratios of MoSe2/MMT/rGO
Seasonal transport, retention, and river-to-ocean transfer dynamics of floating macro-debris in the Umgeni River and estuarine lagoon, South Africa
Abstract Riverine floating macroplastics are considered to contribute a substantial portion of marine plastic pollution, yet the spatiotemporal, geomorphological and hydrological factors that influence river debris passage are poorly understood. To address this, we conducted three field campaigns in the Umgeni River, South Africa, during hydrologically distinct periods (dry season, wet season and recessive season). We deployed GPS drifters ( n = 66 + 4 pilot), each with a density of 0.94 g/cm 3 to mimic floating macroplastics (>5 cm) across five tributaries and the main river. We analysed the drifter trajectories during each campaign to investigate seasonal daily transport rates, retention zones and episodes, and river-to-ocean emission rates. Observed drifter trajectories showed modest seasonal differences in mean daily transport and retention counts. Our pilot drifters captured the 1:50–100 year return period flood occurring in April 2022 showing substantial flushing downstream (0.58 km day − 1 ) to the Indian Ocean. Mean retention durations showed a notable decrease from the dry period (mean = 508 h) towards both the wet and recessive periods (mean = 230 h and mean = 40 h; respectively), indicating increased debris mobility during the wet season. Drifters were retained frequently in upstream river sections along meanders and vegetated banks, which accounted for 48% of all retention events but the estuary emerged as the dominant sink with long-term retention (with mean retention >10 days) thus limiting drifter exports into the ocean. Our study presents quantitative insights of floating macroplastic transport across interconnected riverine domains over three seasons. The results provide novel empirical data on seasonal plastic mobility, retention zones and highlight episodic nature of debris flushing during floods. These findings support river plastic transport modelling and targeted debris cleanup and mitigation policy frameworks.