Inn Med (Heidelb). 2025 May;66(5):533-539. doi: 10.1007/s00108-025-01892-7. Epub 2025 Apr 24.

ABSTRACT

Rare diseases, defined in the European Union as conditions affecting fewer than five per 10,000 inhabitants, often manifest themselves in childhood, but are playing an increasingly important role in internal medicine due to the significantly improved long-term prognosis and a number of diseases that primarily occur in adulthood. Although noteworthy structures already exist nationally and internationally (networks, registers, databases, self-help groups), awareness of these diseases in daily routine and knowledge of the partly divergent care structures must be improved. There are no specific treatments for many of these diseases, but drugs are increasingly being developed-particularly in oncology-that are subject to special orphan drug status.

PMID:40272470 | DOI:10.1007/s00108-025-01892-7

Orphanet J Rare Dis. 2025 Apr 24;20(1):195. doi: 10.1186/s13023-025-03711-6.

ABSTRACT

BACKGROUND: Chanarin-Dorfman syndrome (CDS) is a multisystemic autosomal recessive rare disorder. CDS is caused by variants in the abhydrolase domain containing 5 (ABHD5) encoding gene (CGI-58), which ultimately leads to excessive lipid storage, and therefore a high abundance of cellular lipid droplets (LDs). Although the molecular etiology of the disease was described many years ago, no treatment for CDS is currently available.

RESULTS: To further characterize the molecular basis of the disease and to uncover new treatment avenues, we used skin fibroblasts originating from a young patient diagnosed with CDS due to a homozygous nonsense mutation. We show that dysfunctional ABHD5 does not only affect LDs, but also influences other metabolic-related organelles; the mitochondria and peroxisomes. Additionally, we found that expressing functional ABHD5 in CDS patient cells reduced LD number. Finally, we developed and applied a high content-based drug repurposing screen based on a collection of ∼2500 FDA approved compounds, yielding several compounds that affected LD total area and size.

CONCLUSIONS: Our findings enhance the understanding of the dysfunction underlying CDS and propose new avenues for the treatment of CDS patients.

PMID:40275410 | DOI:10.1186/s13023-025-03711-6

Sci Rep. 2025 Apr 24;15(1):14315. doi: 10.1038/s41598-025-98689-3.

ABSTRACT

Congenital factor (F) VII deficiency is caused by mutations in the F7 gene. The p.Q160R variant manifests with bleeding episodes due to reduced FVII activity and antigen in patient plasma, most likely caused by protein misfolding and intracellular retention. As current replacement therapy is expensive and requires frequent intravenous injections, there is an unmet need for new and less invasive therapeutic strategies. Drug repurposing allows for rapid, more cost-effective discovery and implementation of new treatments, and identification of pharmacological enhancers of FVII variant activity would be of clinical importance. High-throughput screening of > 1800 FDA-approved drugs identified the orally available histone deacetylase inhibitor abexinostat and the inhaled surfactant tyloxapol as enhancers of FVII p.Q160R variant activity. The positive hits were verified in an in vitro cell model transiently expressing wild type or variant FVII and ex vivo in patients' plasma. Both drugs showed a dose-response effect on FVII antigen and activity levels in conditioned cell medium and on FVII activity in patients' plasma. In conclusion, the efficacy of the FDA-approved drugs abexinostat and tyloxapol in enhancing FVII variant activity constitute a proof of principle for high-throughput identification of drugs that may be feasible for novel treatment of FVII deficiency.

PMID:40274887 | DOI:10.1038/s41598-025-98689-3

Bioorg Chem. 2025 Apr 10;160:108464. doi: 10.1016/j.bioorg.2025.108464. Online ahead of print.

ABSTRACT

Pancreatic cancer has a poor prognosis with limited therapeutic options, necessitating novel treatment strategies. While B3GALT5 enzyme overexpression has been reported in pancreatic cancer cases, effective mechanisms to suppress its activity remain unexplored. In this study, we utilized bioinformatics and in silico studies to evaluate the relationship between B3GALT5 enzyme and pancreatic cancer. Through molecular docking analysis, FDA-approved drugs 6-AZA-UTP and itraconazole were identified as potential B3GALT5 enzyme inhibitors. Biological evaluation on MIA PaCa-2 and AsPC-1 pancreatic cancer cell lines demonstrated that both compounds significantly reduced cell viability. Flow cytometry analysis revealed that both drugs effectively suppressed B3GALT5 enzyme activation by decreasing SSEA-3 expression. Furthermore, both compounds exhibited potent anti-tumor effects by inhibiting cell adhesion, colony formation, and migration while inducing apoptosis in pancreatic cancer cells. Notably, both drugs demonstrated favorable ADMET profiles with no carcinogenic or toxic effects. Our investigations revealed that 6-AZA-UTP and itraconazole can effectively suppress B3GALT5 enzyme activity, resulting in tumor suppression and metastasis inhibition. These findings suggest that either 6-AZA-UTP or itraconazole can inhibit B3GALT5 enzyme activity and may serve as promising therapeutic options for pancreatic cancer treatment through drug repurposing strategy.

PMID:40273705 | DOI:10.1016/j.bioorg.2025.108464

Curr Pharm Teach Learn. 2025 Apr 23;17(7):102359. doi: 10.1016/j.cptl.2025.102359. Online ahead of print.

ABSTRACT

INTRODUCTION: Longitudinal curriculum has been suggested for improving pharmacogenomics education, however the outcome of such curriculum design has yet to be reported. Here we evaluated the effectiveness of a simple longitudinal curriculum consisting of didactic lecturing and laboratory-based teaching in two sequential semesters towards pharmacogenomics education.

METHODS: Four pharmacogenomics lectures were offered to professional year 3 (PY3) pharmacy students during the fall semester. During the following spring semester, students participated in two laboratories followed by an implementation project. Knowledge attainment was assessed through an exam following the fall lectures. Students' perception about their clinical pharmacogenomics skills were collected by electronic questionnaire before, immediately after, and 3 months after the fall lectures and the spring laboratories. Statistical analysis was performed using one-way ANOVA followed by pairwise t-test.

RESULTS: The average exam score in Fall 2023 was 79 % (54 %-96 %). Students' perception in a 1-5 Likert scale improved from 1.35 to 3.63 immediately following the lectures (p < 0.0001) but dropped to 1.94 after three months (p < 0.0001). In contrast, after two laboratories in Spring 2024, students' perception improved from 1.94 to 3.67 immediately following the laboratories (p < 0.0001), and importantly, remained high at 3.55 three months later (p = 0.36).

CONCLUSIONS: Combination of didactic lecturing and laboratory-based teaching offered in two sequential semesters is conducive to maintaining student's positive perception about their clinical pharmacogenomics skills. Our curriculum design is simple to implement and has the potential to improve long-term retention of pharmacogenomics knowledge.

PMID:40273885 | DOI:10.1016/j.cptl.2025.102359

Drugs Aging. 2025 Apr 24. doi: 10.1007/s40266-025-01207-3. Online ahead of print.

ABSTRACT

Cystic fibrosis (CF) is an inherited condition that leads to multiorgan dysfunction, especially in the respiratory, gastrointestinal, and reproductive tracts, with associated conditions including persistent pulmonary infection, liver disease, pancreatic insufficiency, and infertility. Historically, people with CF (pwCF) suffered a shortened lifespan due to complications of the condition, namely respiratory. The emphasis on center-based, multidisciplinary care and the widespread introduction of cystic fibrosis transmembrane conductance regulator (CFTR) modulator therapy has resulted in pwCF living longer and healthier lives. Now they may encounter some of the health and social issues associated with growing older, which previously were not a typical experience for this population. In this article, we review relevant health issues for the aging CF population, including complications that arise from the condition itself, issues encountered due to treatment, and general conditions associated with aging that may manifest earlier or differently in pwCF. We discuss the recommendations for screening and treatment of relevant conditions, and considerations for the integration of healthcare professionals across disciplines into the care of this population.

PMID:40274760 | DOI:10.1007/s40266-025-01207-3

Best Pract Res Clin Haematol. 2025 Mar;38(1):101602. doi: 10.1016/j.beha.2025.101602. Epub 2025 Mar 7.

ABSTRACT

Innovation in cystic fibrosis (CF) supportive care, including implementing new antimicrobial agents, improved physiotherapy, and highly effective modulators therapy, has advanced patient survival into the 4th and 5th decades of life. However, even with these remarkable improvements in therapy, CF patients continue to suffer from pulmonary infection and other visceral organ complications associated with long-term deficient cystic fibrosis transmembrane conductance regulator (CFTR) expression. Human mesenchymal stem cells (MSCs) have been utilized in tissue engineering based upon their capacity to provide structural components of mesenchymal tissues. An alternative role of MSCs, however is their versatile utilization as cell-based infusion powerhouses due to the unique capacity to deliver milieu specific soluble biologic factors, promoting immune supportive antimicrobial and anti-inflammatory potency. MSCs derived from umbilical cord blood, bone marrow, adipose and other tissues can be expanded in ex vivo using good manufacturing procedure facilities for a safe, unique therapeutic to reduce and limit CF infection and facilitate the resolution of multi-organ inflammation. In our efforts, we conducted extensive preclinical development and validation of an allogeneic derived bone marrow derived MSC product in preparation for a clinical trial in CF. In this process, potency models were developed to ensure the functional capacity of the MSC product to provide clinical benefit. In vitro, murine in vivo and patient tissue ex vivo potency models were utilized to follow MSC anti-infective and anti-inflammatory potency associated with the CFTR deficient environment. We showed in our "First in CF" clinical trial that the allogeneic MSCs obtained from healthy volunteer bone marrow samples were safe. The advent of improved CF care measures and exciting new small molecules has changed the survival and morbidity phenotype of patients with CF, however, there are CF patients who cannot tolerate or have genotypes that are non-responsive to modulators. Additionally, even with the small molecule therapy, CF patients are living longer, but without genetic correction, with the CF disease manifestation aggravated by the continuance of pre-existing CFTR-associated clinical issues such as ongoing inflammation. MSCs secrete bio-active factors that enhance and protect tissue function and can promote "self-immune" regulation. These properties can provide therapeutic support for the traditional and changing face of CF disease clinical complications. Further, MSC-derived bio-active factors can directly mitigate colonizing pathogens' survival by producing antimicrobial peptides (AMPs) which change the pathogen surface and increase host recognition, elimination, and sensitivity to antibiotics. Herein, we review the potential of MSC therapeutics for treating many facets of CF, emphasizing the potential for providing great additive therapeutics for managing morbidity and quality of life.

PMID:40274338 | DOI:10.1016/j.beha.2025.101602

Respir Med. 2025 Apr 22:108109. doi: 10.1016/j.rmed.2025.108109. Online ahead of print.

ABSTRACT

BACKGROUND: Methicillin-resistant Staphylococcus aureus (MRSA) commonly occur in cystic fibrosis (CF) individuals with advanced disease, but their role in younger CF children is unknown. This study aimed to investigate clinical, functional, and radiological outcomes of CF children with early MRSA colonization.

METHODS: This retrospective cohort study compared CF individuals with MRSA isolation in cultures from respiratory secretions during the first 5 years of life (MRSA group) to age-matched controls. Data from the Brazilian CF Patient Registry and electronic medical records were used. Nutritional outcomes, lung function results between 6 and 7 years, and the first chest CT scan results were comparatively analysed. A linear regression model verified associations between MRSA identification before age 5 and potential confounders with lung function results.

RESULTS: The MRSA group (n=32) had greater oral antibiotic exposure but similar hospital admission rates compared to controls (n=49). The proportion of cultures with methicillin-sensitive S. aureus was greater in the control group (48% vs 29.4%, p=0.009). The MRSA group had lower FEV1 (80.02% vs 92.51%, p=0.023) and higher bronchiectasis scores on chest CT scans (1.9 vs 0.38, p= 0.031). MRSA identification before age 5 was significantly associated with an average 10.1% (95%CI -19.518 - -0.586, p= 0.038) decrease of FEV1.

CONCLUSIONS: Early MRSA identification was associated with increased exposure to antibiotics, higher bronchiectasis scores and lower FEV1 values at age 6. While these findings cannot define a relationship of causality, they provide insight into the associations between early MRSA infection and CF lung disease.

PMID:40273997 | DOI:10.1016/j.rmed.2025.108109

Anim Microbiome. 2025 Apr 24;7(1):39. doi: 10.1186/s42523-025-00410-2.

ABSTRACT

BACKGROUND: The gut microbiome plays a crucial role in health and disease, influencing digestion, metabolism, and immune function. Traditional microbiome analysis methods are often expensive, time-consuming, and require specialized expertise, limiting their practical application in clinical settings. Evolving artificial intelligence (AI) technologies present opportunities for developing alternative methods. However, the lack of transparency in these technologies limits the ability of clinicians to incorporate AI-driven diagnostic tools into their healthcare systems. The aim of this study was to investigate an AI approach that rapidly predicts different bacterial genera and bacterial groups, specifically butyrate producers, from digital images of fecal smears of rhesus macaques (Macaca mulatta). In addition, to improve transparency, we employed explainability analysis to uncover the image features influencing the model's predictions.

RESULTS: By integrating fecal image data with corresponding metagenomic sequencing information, the deep learning (DL) and machine learning (ML) algorithms successfully predicted 16 individual bacterial genera (area under the curve (AUC) > 0.7) among the 50 most abundant genera in rhesus macaques (Macaca mulatta). The model was successful in predicting functional groups, major butyrate producers (AUC 0.75) and a mixed group including fermenters and short-chain fatty acid (SCFA) producers (AUC 0.81). For both models of butyrate producers and mixed fermenters, the explainability experiments revealed no decline in the AUC when random noise was added to the images. Increased blurring led to a gradual decline in the AUC. The model's performance was robust against the impact of fecal shape from smearing, with a stable AUC maintained until patch 4 for all groups, as assessed through scrambling. No significant correlation was detected between the prediction probabilities and the total fecal weight used in the smear; r = 0.30 ± 0.3 (p > 0.1) and r = 0.04 ± 0.36 (p > 0.8) for the butyrate producers and mixed fermenters, respectively.

CONCLUSION: Our approach demonstrated the ability to predict a wide range of clinically relevant microbial genera and microbial groups in the gut microbiome based on digital images from a fecal smear. The models proved to be robust to the smearing method, random noise and the amount of fecal matter. This study introduces a rapid, non-invasive, and cost-effective method for microbiome profiling, with potential applications in veterinary diagnostics.

PMID:40275402 | DOI:10.1186/s42523-025-00410-2

BMC Public Health. 2025 Apr 24;25(1):1520. doi: 10.1186/s12889-025-22705-4.

ABSTRACT

BACKGROUND: Understanding the complex interplay between life course exposures, such as adverse childhood experiences and environmental factors, and disease risk is essential for developing effective public health interventions. Traditional epidemiological methods, such as regression models and risk scoring, are limited in their ability to capture the non-linear and temporally dynamic nature of these relationships. Deep learning (DL) and explainable artificial intelligence (XAI) are increasingly applied within healthcare settings to identify influential risk factors and enable personalised interventions. However, significant gaps remain in understanding their utility and limitations, especially for sparse longitudinal life course data and how the influential patterns identified using explainability are linked to underlying causal mechanisms.

METHODS: We conducted a controlled simulation study to assess the performance of various state-of-the-art DL architectures including CNNs and (attention-based) RNNs against XGBoost and logistic regression. Input data was simulated to reflect a generic and generalisable scenario with different rules used to generate multiple realistic outcomes based upon epidemiological concepts. Multiple metrics were used to assess model performance in the presence of class imbalance and SHAP values were calculated.

RESULTS: We find that DL methods can accurately detect dynamic relationships that baseline linear models and tree-based methods cannot. However, there is no one model that consistently outperforms the others across all scenarios. We further identify the superior performance of DL models in handling sparse feature availability over time compared to traditional machine learning approaches. Additionally, we examine the interpretability provided by SHAP values, demonstrating that these explanations often misalign with causal relationships, despite excellent predictive and calibrative performance.

CONCLUSIONS: These insights provide a foundation for future research applying DL and XAI to life course data, highlighting the challenges associated with sparse healthcare data, and the critical need for advancing interpretability frameworks in personalised public health.

PMID:40275204 | DOI:10.1186/s12889-025-22705-4

BMC Pediatr. 2025 Apr 24;25(1):324. doi: 10.1186/s12887-025-05628-2.

ABSTRACT

BACKGROUND: Congenital heart disease (CHD) affects approximately 1% of newborns and is a leading cause of mortality in early childhood. Despite the importance of early detection, current screening methods, such as pulse oximetry and auscultation, have notable limitations, particularly in identifying non-cyanotic CHD. (AI)-assisted electrocardiography (ECG) analysis offers a cost-effective alternative to conventional CHD detection. However, most existing models have been trained on older children, limiting their generalizability to infants and young children. This study developed an AI model trained on real-world ECG data for the detection of hemodynamically significant CHD in children under five years of age.

METHODS: ECG data was retrospectively collected from 1,035 patients under five years old at Chang Gung Memorial Hospital, Taoyuan, Taiwan (2013-2020). Based on ECG findings, patients were categorized into the following groups: normal heart structure (NOR), non-significant right heart disease (RHA), significant right heart disease (RHB), non-significant left heart disease (LHA), and significant left heart disease (LHB). ECG signals underwent preprocessing using continuous wavelet transformation and segmentation into 2-s intervals for data augmentation. Transfer learning was applied using three pre-trained deep learning models: ResNet- 18, InceptionResNet-V2, and NasNetMobile. Model performance was evaluated in terms of accuracy, sensitivity, specificity, F1 score, and area under the receiver operating characteristic curve (AUC).

RESULTS: Among the tested models, the model based on ResNet-18 demonstrated the best overall performance in predicting clinically significant CHD, achieving accuracy of 73.9%, an F1 score of 75.8%, and an AUC of 81.0% in differentiating significant from non-significant CHD. InceptionResNet-V2 performed well in detecting left heart disease but was computationally intensive. The proposed AI model significantly outperformed conventional ECG interpretation by pediatric cardiologists (accuracy 67.1%, sensitivity 71.6%).

CONCLUSIONS: This study highlights the potential of AI-assisted ECG analysis for CHD screening in young children. The ResNet-18-based model outperformed conventional ECG evaluation, suggesting its feasibility as a supplementary tool for early CHD detection. Future studies should focus on multi-center validation, inclusion of more CHD subtypes, and integration with other screening modalities to improve diagnostic accuracy and clinical applicability.

PMID:40275174 | DOI:10.1186/s12887-025-05628-2

Sci Rep. 2025 Apr 24;15(1):14321. doi: 10.1038/s41598-025-98963-4.

ABSTRACT

In the process of industrial production, bearing fault detection has always been a hot issudza20000528@163.comsolved. At present, the problem of less fault data samples in the field of fault detection has caused great trouble to the research of deep learning. In the application of industrial fault detection, which is difficult to obtain massive data, it is easy to lead to the lack of fitting of neural network training and many generalization problems. To solve the above problems, this paper proposes an improved and more efficient method of few-shot supervised learning, which is called the Temporal Classification Prototype Network (TCPN). This model is designed to maintain both training efficacy and generalization capabilities under conditions of data scarcity. Initially, Fourier transform is employed to accentuate the frequency domain characteristics of the fault section in the bearing signal before it is input into the model, thereby enabling the subsequent model to concentrate on distinguishing between normal and fault signals. Subsequently, discrete data sample points are transformed into points within the feature space via our Enhanced Temporal Convolutional Network(ETCN). In our investigation, we utilize the features of the support set as anchors within the feature space and employ similarity measures as the basis for classification, thus developing a more effective comparative learning classifier known as the ContractSim Classifier (CSC). Within the CSC, the model learns the data features of the query set, which are then back-propagated to refine our model. The proposed TCPN model has been evaluated across four standard bearing datasets, corroborating its few-shot learning proficiency through k-shot experiments. In comparative model experiments, our TCPN outperforms baseline models, while the ablation study confirms the rationality and robustness of our module integration.

PMID:40275051 | DOI:10.1038/s41598-025-98963-4

Sci Rep. 2025 Apr 24;15(1):14249. doi: 10.1038/s41598-025-98766-7.

ABSTRACT

Music genre categorization has been considered to be an essential task within the context of music data recovery. Genres serve as categories or labels that enable the classification of music based on shared attributes, including musical style, instrumentation, cultural origins, historical context, and other distinctive elements. The purpose of classifying music genres is to automatically assign music pieces to one or more predefined genres. The present research suggests a new method for music genre identification via integrating deep learning models with a metaheuristic algorithm. The proposed model uses a pre-trained Zeiler and Fergus Network (ZFNet) to extract high-level features from audio signals, while an Extreme Learning Machines (ELM) is utilized for efficient classification. Furthermore, the model incorporates a newly developed metaheuristic algorithm called the Modified Electric Eel Foraging Optimization (MEEFO) algorithm to optimize the ELM parameters and enhance overall performance. To evaluate the effectiveness of the model, it has been tested on two widely recognized benchmark datasets, namely GTZAN and Ballroom, and the results are contrasted with some advanced models, comprising MusicRecNet, Parallel Recurrent Convolutional Neural Network (PRCNN), RNN-LSTM, ResNet-50, VGG-16, Deep Neural Network (DNN). The outcomes demonstrated that the suggested system surpassed several existing methods regarding precision, recall, and accuracy.

PMID:40275047 | DOI:10.1038/s41598-025-98766-7

Sci Rep. 2025 Apr 24;15(1):14266. doi: 10.1038/s41598-025-98423-z.

ABSTRACT

In the field of deep learning-based object detection, RGB-T salient object detection (SOD) networks show significant potential for cross-modal information fusion. However, existing methods still face considerable challenges in complex scenes. Specifically, current cross-modal feature fusion approaches fail to exploit the complementary information between modalities fully, resulting in limited robustness when handling diverse inputs. Furthermore, inadequate adaptation to multi-scale features hinders accurately recognizing salient objects at different scales. Although some feature decoding strategies attempt to mitigate noise interference, they often struggle in high-noise environments and lack flexible feature weighting, further restricting fusion capabilities. To address these limitations, this paper proposes a novel salient object detection network, TCAINet. The network integrates a Channel Attention (CA) mechanism, an enhanced cross-modal fusion module (CAF), and an adaptive decoder (AAD) to improve both the depth and breadth of feature fusion. Additionally, diverse noise addition and augmentation methods are applied during data preprocessing to boost the model's robustness and adaptability. Specifically, the CA module enhances the model's feature selection ability, while the CAF and AAD modules optimize the integration and processing of multimodal information. Experimental results demonstrate that TCAINet outperforms existing methods across multiple evaluation metrics, proving its effectiveness and practicality in complex scenes. Notably, the proposed model achieves improvements of 0.653%, 1.384%, 1.019%, and 5.83% in Sm, Em, Fm, and MAE metrics, respectively, confirming its efficacy in enhancing detection accuracy and optimizing feature fusion. The code and results can be found at the following link:huyunfei0219/TCAINet.

PMID:40275036 | DOI:10.1038/s41598-025-98423-z

Sci Rep. 2025 Apr 24;15(1):14258. doi: 10.1038/s41598-025-99174-7.

ABSTRACT

Classifying river morphology is crucial for fluvial geomorphology and hydrology. River morphology reflects hydrodynamic and sedimentary processes, providing critical insights into the diversity of global river systems. This study establishes a global framework for river morphology classification based on remote sensing and topographic data. Using the Global Inland Water Dynamics Characterization dataset and the global digital elevation model ASTER GDEM V3, a river spatial image decomposition process was developed, dividing global river data into tens of thousands of image blocks containing dynamic imagery and elevation information. A ResNet-50 deep neural network was employed to construct an image-elevation fusion classification model, classifying global rivers into five major types: meandering rivers, braided rivers, straight rivers, anastomosing rivers, and anabranching rivers. These types were further divided into 17 subtypes to capture finer morphological variations. The spatial distribution patterns and morphological features of these river types were analyzed, providing a comprehensive understanding of the global distribution of river planforms. This framework advances the knowledge of river systems at a global scale and lays the foundation for future studies in fluvial geomorphology and hydrology.

PMID:40275033 | DOI:10.1038/s41598-025-99174-7

Sci Rep. 2025 Apr 24;15(1):14282. doi: 10.1038/s41598-025-98565-0.

ABSTRACT

Glioma, a malignant intracranial tumor with high invasiveness and heterogeneity, significantly impacts patient survival. This study integrates multi-omics data to improve prognostic prediction and identify therapeutic targets. Using single-cell data from glioblastoma (GBM) and low-grade glioma (LGG) samples, we identified 55 distinct cell states via the EcoTyper framework, validated for stability and prognostic impact in an independent cohort. We constructed multi-omics datasets of 620 samples, integrating transcriptomic, copy number variation (CNV), somatic mutation (MUT), Microbe (MIC), EcoTyper result data. A scRNA-seq enhanced Self-Normalizing Network-based glioma prognosis model achieved a C-index of 0.822 (training) and 0.817 (test), with AUC values of 0.867, 0.876, and 0.844 at 1, 3, and 5 years in the training set, and 0.820, 0.947, and 0.936 in the test set. Gradient attribution analysis enhanced the interpretability of the model and identified key molecular markers. The classification into high- and low-risk groups was validated as an independent prognostic factor. HDAC inhibitors are proposed as potential treatments. This study demonstrates the potential of integrating scRNA-seq and multi-omics data for robust glioma prognosis and clinical decision-making support.

PMID:40275021 | DOI:10.1038/s41598-025-98565-0

Sci Rep. 2025 Apr 24;15(1):14255. doi: 10.1038/s41598-025-98180-z.

ABSTRACT

At present, the internet of things (IoT) plays a vital part in the growth of programmed electrical power stations while presenting magnificent chances, particularly cybersecurity. In IoT networks, security is now required owing to the higher amount of data to be handled. IoT cybersecurity aims to decrease the cybersecurity threat for users and organizations over protecting IoT assets and privacy. Therefore, identifying numerous anomalies or cyberattacks in a network and constructing an effectual intrusion detection system (IDS) becomes more significant. Artificial intelligence (AI), mostly machine learning (ML) and deep learning (DL), has been employed to construct a data-driven intelligent IDS. This paper presents a multi-head attention-driven intrusion detection with improved white shark optimization algorithm (MHAID-IWSOA) methodology in IoT networks. The main intention of the MHAID-IWSOA methodology relies on enhancing the cybersecurity detection and migration model using advanced optimization algorithms. Initially, the data pre-processing applies min-max scaling to transform input data into a beneficial format. Besides, the sand cat swarm optimization (SCSO) model is used for the feature selection (FS) process. The proposed MHAID-IWSOA model employs the bidirectional gated recurrent unit with multi-head attention (BiGRU-MHA) technique for attack detection and classification. Finally, the improved white shark optimization (IWSO) technique optimally alters the hyperparameter value of the BiGRU-MHA technique and results in superior classification performance. The experimental evaluation of the MHAID-IWSOA model is performed on the Edge-IIoT dataset. The extensive comparison analysis of the MHAID-IWSOA model illustrated a superior accuracy outcome of 98.28% over existing techniques.

PMID:40274990 | DOI:10.1038/s41598-025-98180-z

Sci Rep. 2025 Apr 24;15(1):14300. doi: 10.1038/s41598-025-99009-5.

ABSTRACT

Breast cancer is the most prevalent cancer and the second cause of cancer related death among women in the United States. Accurate and early detection of breast cancer can reduce the number of mortalities. Recent works explore deep learning techniques with ultrasound for detecting malignant breast lesions. However, the lack of explanatory features, need for segmentation, and high computational complexity limit their applicability in this detection. Therefore, we propose a novel ultrasound-based breast lesion classification framework that utilizes two-dimensional variational mode decomposition (2D-VMD) which provides self-explanatory features for guiding a convolutional neural network (CNN) with mixed pooling and attention mechanisms. The visual inspection of these features demonstrates their explainability in terms of discriminative lesion-specific boundary and texture in the decomposed modes of benign and malignant images, which further guide the deep learning network for enhanced classification. The proposed framework can classify the lesions with accuracies of 98% and 93% in two public breast ultrasound datasets and 89% in an in-house dataset without having to segment the lesions unlike existing techniques, along with an optimal trade-off between the sensitivity and specificity. 2D-VMD improves the areas under the receiver operating characteristics and precision-recall curves by 5% and 10% respectively. The proposed method achieves relative improvement of 14.47%(8.42%) (mean (SD)) in accuracy over state-of-the-art methods for one public dataset, and 5.75%(4.52%) for another public dataset with comparable performance to two existing methods. Further, it is computationally efficient with a reduction of [Formula: see text] in floating point operations as compared to existing methods.

PMID:40274985 | DOI:10.1038/s41598-025-99009-5

J Comput Aided Mol Des. 2025 Apr 25;39(1):19. doi: 10.1007/s10822-025-00596-2.

ABSTRACT

Ubiquitin-specific proteases (USPs) are crucial regulators of protein degradation pathways, influencing diverse cellular processes and disease mechanisms. Among them, USP30 plays a pivotal role in mitochondrial quality control and has been implicated in idiopathic pulmonary fibrosis (IPF), a chronic lung disease for which current therapies merely slow disease progression. The high flexibility of USP30's catalytic site, coupled with its dependence on covalent interaction with the catalytic cysteine presents significant challenges in discovering suitable small molecule inhibitors. In this study, we identified three non-covalent small molecule inhibitors for USP30 using molecular modeling, X-ray crystallography, and virtual screening. These findings offer valuable insights and novel chemical starting points for further medicinal chemistry optimization.

PMID:40274689 | DOI:10.1007/s10822-025-00596-2

J Ethnopharmacol. 2025 Apr 22:119866. doi: 10.1016/j.jep.2025.119866. Online ahead of print.

ABSTRACT

ETHNOPHARMACOLOGICAL RELEVANCE: Saengmaek-san (SMS) is a herbal prescription comprising Liriope platyphylla, Panax ginseng, and Schisandra chinensis. In traditional Korean medicine (TKM), SMS has been used to treat a condition known as the dual deficiency of qi and yin in the lungs, a syndrome characterized by the depletion of vitality and body fluids, often resulting from heat exhaustion. SMS has primarily been used to promote fluid production, alleviate dry cough, and relieve progressive dyspnea.

AIM OF THE STUDY: The current study was planned to explore the efficacy and underlying mechanisms of SMS in managing idiopathic pulmonary fibrosis.

MATERIALS AND METHODS: In mice with bleomycin-induced pulmonary fibrosis, the SMS water extract was administered at doses of 50, 150, and 450 mg/kg twice daily for 14 days. The extent of pulmonary fibrosis was assessed using the Ashcroft scale in stained lung tissues. The levels of transforming growth factor-β, α-smooth muscle actin (α-SMA), and collagen accumulation were also evaluated. Bronchoalveolar lavage fluid (BALF) was collected to measure the total cell counts, white blood cell ratios, and cytokine levels (IL-6 and IL-10).

RESULTS: We observed statistically significant and potential anti-fibrotic effects in the SMS 450 mg/kg treatment group in terms of preventing body weight loss, decreasing Ashcroft scale, and reducing macrophage and granulocyte counts in BALF, as well as reducing α-SMA and collagen production. Additionally, an increase was observed in the levels of anti-inflammatory cytokine IL-10.

CONCLUSIONS: SMS demonstrated potential as a therapeutic candidate for idiopathic pulmonary fibrosis by exerting anti-inflammatory effects and reducing collagen deposition.

PMID:40274032 | DOI:10.1016/j.jep.2025.119866