Browse Articles
Discover research articles across all indexed journals
Toxic chinese herbal medicine recognition in real-world images via multi-scale and attention-enhanced EfficientNetV2
Accurate identification of toxic Chinese herbal medicines is critical for public health and clinical safety. However, real-world herbal images often exhibit complex backgrounds and small, indistinct target regions, posing substantial challenges to automated classification systems. In this study, we present a novel image dataset comprising over 4,000 samples from 47 toxic herb categories, captured under diverse environmental conditions to reflect real-world variability. We benchmark several state-of-the-art convolutional neural networks, including ResNet, ResNeXt, and EfficientNet variants, and identify EfficientNetV2 as the most effective baseline. To further enhance model robustness and discriminative capability, we propose an improved EfficientNetV2 architecture incorporating two lightweight yet effective modules: a Multi-Scale Feature Fusion (MSFF) module to integrate hierarchical features, and a Convolutional Block Attention Module (CBAM) to refine both spatial and channel-wise representations. Experimental results demonstrate that our enhanced model achieves 91.28% Top-1 accuracy, 97.52% Top-5 accuracy, and a 90.27% macro F1-score, significantly outperforming baseline methods. Ablation studies confirm the complementary benefits of MSFF and CBAM, and targeted evaluations on challenging image subsets reveal improved resilience to background clutter and small object localization. The proposed architecture offers a high-accuracy, generalizable, and computationally efficient solution for toxic herbal medicine classification and provides a valuable reference for intelligent traditional medicine recognition applications. The GitHub repository for this project is available at: https://github.com/zhuguohua1992/Toxic-chinese-herbal-medicine-recognition-via-enhanced-EfficientNetV2 .
HOD08 for low-risk pediatric Hodgkin lymphoma
Leveraging haze-aware features for improved image clarity and detection accuracy with an optimized DCNN-YOLOv8 network
Factors associated with patient adherence to medical recommendations during a health crisis
Introduction Patient non-adherence to medical recommendations is associated with significant economic and health consequences, particularly in chronic diseases. While less studied, similar implications are also observed for short-term antibiotic treatment non-adherence. A global pandemic creates a complex medical reality with the potential to impact patient adherence, and it is of interest to investigate its effect on patient adherence regarding common community-acquired infectious diseases. This study focused on primary care clinic visits with a diagnosis of fever or cough and followed the performance of referred laboratory tests and chest X-rays and the purchase of prescribed antibiotics, as adherence outcomes. Methods A comprehensive and unique dataset of clinic visits in Southern Israel was used to assess the impact of the pandemic on patient adherence outcomes. We compared the outcomes between the pandemic period and the preceding and succeeding periods and used multivariate modeling to study the impact of the pandemic period on adherence, subject to patient and visit characteristics. Results In total, 609,823 visits were analyzed. Before the pandemic, adherence to throat cultures surpassed that of chest X-ray and referred laboratory tests, and cephalosporins exhibiting higher adherence rates than other antibiotics. Social restrictions and lockdown periods were found to be associated with a decrease in patient adherence. Adherence to referred tests decreased among all patients during social restrictions for complete blood count (Odds Ratio (OR)=0.79, 95%CI = 0.64,0.96), and throat culture (OR=0.64, 95%CI = 0.43,0.93) and among older patients for chest X-ray. Adherence to prescribed antibiotics dropped during social restrictions for penicillin (OR=0.26, 95%CI = 0.17,0.40) and cephalosporins (OR=0.52, 95%CI = 0.32,0.86), but increased for macrolides (OR=2.16, 95%CI = 1.20,3.88). Rural clinics were associated with lower adherence in all outcomes, and visits by phone call were associated with lower adherence, most prominently in throat culture (OR=0.32, 95%CI = 0.26,0.38). Conclusion This study offers a unique contribution to our understanding of patient adherence to acute infection management strategies during a pandemic. The findings illuminate the significant impact of uncertainty on patient adherence to diagnostic testing and antibiotic regimens. Understanding these multifaceted influences is essential for improving patient outcomes and optimizing antibiotics use and resource availability during public health crises.
Platelets in the crossfire of immune checkpoint therapy
A server-assisted secure blockchain model for residential demand response in smart grids
Abstract The paper proposes a two layered blockchain-based residential demand side management (DSM) scheme that integrates a secure centralized server, (Energy Plus) EnPlus, and a private Ethereum blockchain to deliver both real-time operational efficiency and distributed transaction trust. The architecture also allows secure and transparent peer-to-peer energy trading using (Green Energy Reward) GER Tokens, which are controlled through Solidity-based smart contracts. Computationally demanding tasks of load forecasting, bid-offer matching and policy enforcement are carried out in the EnPlus server. The authorization of prosumers is provided through Public Key Infrastructure where AES-256 encryption is used when data from smart meter collected and uploaded to server and SHA-256 hashing with RSA is used to provide data confidentiality and integrity. The shiftable appliances are scheduled through Firefly Optimization Algorithm in two conditions: DSM without P2P trading and DSM with P2P trading using tokens to reduce the energy costs. The validity of the system is brought out with a case study of 52 prosumers (25 real prosumers and 27 are part of the extended project) within the climatic conditions of Kolkata with each having a 2.5 kW rooftop photovoltaic system. This system offers better savings and reduces peak to average ratio. P2P trading has typical financial advantages and it will lead to local energy independence but there will be little delays in transactions because of the hybrid processing model. The protocol is proven to be highly scalable, privacy preserving, and robust that offers a viable roadmap to secure, decentralized residential energy markets.
Altitudinal and household breeding patterns of the medically important mosquitoes Aedes aegypti, Aedes albopictus and Culex quinquefasciatus in Nepal
Medically relevant mosquitoes have established populations across Nepal, adapting to diverse environmental conditions and altitudes. In addition to arbovirus vectors, the presence of Culex quinquefasciatus , a primary vector of lymphatic filariasis, underscores the need for comprehensive mosquito surveillance. This study investigated the presence, abundance, container preferences, and altitudinal distribution of Aedes aegypti , Aedes albopictus , and Cx. quinquefasciatus across selected districts in Nepal. We also examined the co-occurrence of Aedes species and their association with container type and location. An entomological survey was conducted along an altitudinal transect from 140 meters above sea level (m asl) (Chitwan) to 2340 m asl (Dolakha). Immatures were collected from 1,157 water-holding containers across 636 households using dipper and dropper techniques, while adults were captured using BG-Sentinel traps. Immatures were reared to adulthood for identification, while non-reared specimens were identified via microscopic slides. A total of 1,323 immature and 158 adult mosquitoes were recorded. Among immatures, Ae. albopictus (47.6%) was the most abundant, followed by Ae. aegypti (43.4%). Adults were predominantly Cx. quinquefasciatus (82.9%). Plastic bottles and drums were the most common larval habitats. Aedes species were found across the study transect, from 140 to 2066 m asl. The altitude did not have a statistically significant effect on the number of collected mosquitoes ( p = 0.563). These findings highlight the widespread presence of Aedes and Culex species across different altitudes and breeding habitats. The predominance and co-occurrence of dengue vectors and Cx. quinquefasciatus in nearby households underscore the need for integrated vector surveillance and control strategies in Nepal.
Dose-dense chemotherapy enables elimination of RT for the majority of low-risk pediatric Hodgkin lymphomas: PHC study HOD08
Abstract The Pediatric Hodgkin Consortium hypothesized that by increasing chemotherapeutic dose density for Hodgkin lymphoma (HL) they could increase the complete response (CR) rate among patients with favorable-risk HL after 8 weeks of Stanford V (vinblastine, doxorubicin, vincristine, bleomycin, mechlorethamine, etoposide and prednisone) compared with 8 weeks of VAMP (vinblastine, Adriamycin [doxorubicin], methotrexate, and prednisone). This would translate to a decrease in patients who required radiation therapy (RT) to achieve a cure. The HOD08 study was a phase 2 multicenter, investigator-initiated single-arm trial for patients aged ≤21 years with previously untreated stage 1A or 2A HL without mediastinal bulk or extranodal disease extension and <3 sites of disease. Treatment consisted of a modified 8-week Stanford V regimen. Modified, tailored, field RT was administered only to disease sites achieving less than a CR. The primary objective was to increase CR rate after 8 weeks of chemotherapy by at least 20% (from an estimated 44% to 64%) compared with patients treated on a previous trial (HOD99). HOD08 enrolled 85 patients with HL and 72 were evaluable for the primary objective, of whom 55 (76.4%) achieved a CR at all sites and did not receive RT. The 5-year event-free survival and overall survival rates for the entire cohort were 87.4% (95% confidence interval [CI], 80.4-95.0) and 98.7% (95% CI, 96.2-100), respectively. A dose-dense modified Stanford V regimen reduced the proportion of pediatric patients with low-risk HL who received RT while maintaining excellent outcomes. This trial was registered at www.clinicaltrials.gov as #NCT00846742.
Paired electrolysis enables para-C–H amination of phenols with nitroarenes and mechanistic visualization via multifunctional electrochemical mass spectrometry
Bst2-targeted senotherapy restores visual function by eliminating senescent retinal cells
Abstract Senescent cells contribute to degenerative processes in multiple tissues, including the retina. In the retinal pigment epithelium (RPE), their accumulation is closely associated with retinal aging and disease progression. Eliminating senescent RPE cells has shown therapeutic potential, but conventional senolytics often lack the specificity required to spare non-senescent cells, raising safety concerns. To overcome this, we performed integrated transcriptomic analyses of male mouse-derived RPE cells under natural aging and chemically induced senescence conditions. These analyses identified Bst2 as a membrane-localized marker selectively upregulated in senescent RPE cells, with minimal expression in young controls. Based on this discovery, we developed a modular, antibody-pluggable drug delivery platform–B-Z-PON–comprising mesoporous silica nanoparticles functionalized with a recombinant Fc-binding domain and conjugated with anti-Bst2 antibodies. This nanocarrier selectively accumulates in Bst2-expressing senescent RPE cells, enabling targeted drug delivery and sparing healthy retinal cells. In vivo administration of ABT-263-loaded B-Z-PON in aged and senescence-induced retinal degeneration models resulted in the selective ablation of senescent cells, restoration of RPE function, and improved visual outcomes. Together, our study integrates senescence-specific marker discovery with precision nanomedicine, establishing a versatile platform for targeted senotherapy. These findings offer a promising therapeutic approach for retinal aging disorders, such as age-related macular degeneration.
Enhancing the performance of Magnets photosensors
Abstract Photosensory protein domains, derived from nature, are foundational for optogenetic protein engineering. Tailoring their properties enables their full exploitation for optogenetic regulation in basic research and applied bioengineering applications. Here, we present a simple, yet powerful strategy based on random mutagenesis coupled to high-throughput screening that allowed altering the most fundamental properties of the widely used nMag/pMag photodimerization system: its light sensitivity and activation. Variants were characterized in vivo in bacteria by flow cytometry and during the entire growth curve by spectrofluorometry. We identify mutations that either increase or decrease the light sensitivity at sub-saturating light intensities, while also improving the light activation and dark-to-light fold change. Notably, light sensitivity and activation levels could be changed independently. In addition, we demonstrated that the shapes of the dose-response curves can be finely tuned. This broadens the applicability of the Magnets photosensors for optogenetic regulation strategies.
RPA directly stimulates Mer3 helicase processivity to ensure normal crossover formation in meiosis
Abstract Meiotic crossover formation is critical for generating viable gametes and enhancing genetic diversity. The helicase Mer3 (HFM1 in humans) is a highly conserved factor essential for promoting crossovers and ensuring their proper distribution. Here, we identify replication protein A (RPA) as a direct interactor of budding yeast Mer3. We demonstrate that this interaction is conserved between human HFM1 and RPA. Cross-linking mass spectrometry and structural modelling with AlphaFold2 reveal a conserved and specific Mer3-RPA interface. Single-molecule magnetic tweezers assays demonstrate that direct RPA interaction is required for Mer3 helicase processivity under conditions of low DNA tension. Consistently, a mer3 mutant deficient in RPA binding exhibits reduced crossover frequencies and accumulates unresolved recombination intermediates during budding yeast meiosis. Via genome-wide localisation experiments, we link this effect to weakened recruitment of the mer3 mutant to double-strand break sites. Our findings provide mechanistic insights into coordination of meiotic recombination by the Mer3 helicase through interactions with the canonical DNA repair machinery, highlighting a conserved mechanism underlying crossover control during sexual reproduction.
Loss of polarity by Cdc42 depletion and oncogenic Kras activation in the mouse intestinal epithelia leads to a necrotizing enterocolitis (NEC)-like disease
Abstract Cell polarity is essential for maintaining intestinal epithelial organization and function. Here we show that combined loss of polarity by epithelial loss of Cdc42 with oncogenic Kras expression in mice causes small intestine failure leading to weight loss, inflammation, epithelial necroptosis, and lethality. These phenotypic defects are characterized by a loss of intestinal stem cells, disrupted epithelial architecture, altered hippo signaling, elevated inflammatory cytokines, and activation of necroptotic cell death, that closely resemble necrotizing enterocolitis (NEC). Single-cell transcriptomic analysis reveals a coordinated dysregulation of polarity machinery, inflammatory pathways, and necroptosis program. Suppression of YAP, IL-1, TNFα signaling or necroptosis rescues the intestinal pathology. Similar NEC-like phenotypes arise when Cdc42 loss and oncogenic Kras activation are initiated from intestinal stem cells. These findings provide mechanism insights involving polarity-YAP-IL1/TNFα signaling induced necroptosis for the synergistic effect of hyperactivation of Kras signaling and loss of polarity in disrupting intestinal epithelia.