Stem Cell Res Ther. 2025 Apr 20;16(1):194. doi: 10.1186/s13287-025-04313-6.

ABSTRACT

BACKGROUND: Idiopathic pulmonary fibrosis is a progressive lung disorder, presenting clinically with symptoms such as shortness of breath and hypoxemia. Despite its severe prognosis and limited treatment options, the pathogenesis of idiopathic pulmonary fibrosis remains poorly understood. This study aims to investigate the therapeutic potential of mesenchymal stromal cells in treating idiopathic pulmonary fibrosis, focusing on their ability to modulate regulatory T cells through the low tumor necrosis factor superfamily member 4 (TNFSF4) pathway. The goal is to identify mesenchymal stromal cells subtypes with optimal immunomodulatory effects to enhance regulatory T cells functions and ameliorate fibrosis.

METHODS: We identified the immune characteristics of idiopathic pulmonary fibrosis by mining and analyzing multiple public datasets and detecting regulatory T cells in the blood and lung tissues of idiopathic pulmonary fibrosis patients. An extensive examination followed, including assessing the impact of mesenchymal stromal cells on regulatory T cells proportions in peripheral blood and lung tissue, and exploring the specific role of TNFSF4 expression in regulatory T cells modulation. Whole-genome sequencing and cluster analysis were used to identify mesenchymal stromal cells subtypes with low TNFSF4 expression.

RESULTS: Mesenchymal stromal cells characterized by TNFSF4 expression (TNFSF4low-MSCs) demonstrated enhanced ability to regulate regulatory T cells subpopulations and exhibited pronounced anti-fibrotic effects in the bleomycin-induced idiopathic pulmonary fibrosis mouse model. These mesenchymal stromal cells increased regulatory T cells proportions, reduced lung fibrosis, and improved survival rates. TNFSF4-tumor necrosis factor receptor superfamily member 4 (TNFRSF4) signaling was identified as a critical pathway influencing regulatory T cells generation and function.

CONCLUSIONS: Our findings underscore the pivotal role of TNFSF4 in mesenchymal stromal cells mediated regulatory T cells modulation and highlight the therapeutic potential of selecting mesenchymal stromal cells subtypes based on their TNFSF4 expression for treating idiopathic pulmonary fibrosis. This approach may offer a novel avenue for the development of targeted therapies aimed at modulating immune responses and ameliorating fibrosis in idiopathic pulmonary fibrosis.

TRIAL REGISTRATION: Our study involved collecting 10 mL of peripheral blood from idiopathic pulmonary fibrosis patients, and the Medical Ethics Committee of Nanjing Drum Tower Hospital approved our study protocol with the approval number 2023-675-01.

PMID:40254578 | DOI:10.1186/s13287-025-04313-6

J Ethnopharmacol. 2025 Apr 19:119763. doi: 10.1016/j.jep.2025.119763. Online ahead of print.

NO ABSTRACT

PMID:40254530 | DOI:10.1016/j.jep.2025.119763

Ann Clin Microbiol Antimicrob. 2025 Apr 20;24(1):27. doi: 10.1186/s12941-025-00793-9.

ABSTRACT

BACKGROUND: Long COVID is a complex, heterogeneous syndrome affecting over four hundred million people globally. There are few recommendations, and no formal training exists for medical professionals to assist with clinical evaluation and management of patients with Long COVID. More research into the pathology, cellular, and molecular mechanisms of Long COVID, and treatments is needed. The goal of this work is to disseminate essential information about Long COVID and recommendations about definition, diagnosis, treatment, research and social issues to physicians, researchers, and policy makers to address this escalating global health crisis.

METHODS: A 3-round modified Delphi consensus methodology was distributed internationally to 179 healthcare professionals, researchers, and persons with lived experience of Long COVID in 28 countries. Statements were combined into specific areas: definition, diagnosis, treatment, research, and society.

RESULTS: The survey resulted in 187 comprehensive statements reaching consensus with the strongest areas being diagnosis and clinical assessment, and general research. We establish conditions for diagnosis of different subgroups within the Long COVID umbrella. Clear consensus was reached that the impacts of COVID-19 infection on children should be a research priority, and additionally on the need to determine the effects of Long COVID on societies and economies. The consensus on COVID and Long COVID is that it affects the nervous system and other organs and is not likely to be observed with initial symptoms. We note, biomarkers are critically needed to address these issues.

CONCLUSIONS: This work forms initial guidance to address the spectrum of Long COVID as a disease and reinforces the need for translational research and large-scale treatment trials for treatment protocols.

PMID:40254579 | DOI:10.1186/s12941-025-00793-9

J Neuroendocrinol. 2025 Apr 20:e70033. doi: 10.1111/jne.70033. Online ahead of print.

ABSTRACT

Evidence from both clinical and preclinical studies indicates that females experience a faster progression to drug addiction and more severe addiction-related health effects compared with males. Estradiol (E2) plays a critical role in these sex differences. Recently, we demonstrated that E2 significantly exacerbates adverse health effects, such as respiratory distress and weight loss, in ovariectomized (OVX) female rats during withdrawal from extended-access fentanyl self-administration. To uncover the mechanisms behind E2-enhanced toxicity, we investigated proteomic changes in the plasma of fentanyl-withdrawn OVX rats under both E2 and non-E2 presentation conditions.Plasma samples were collected following extended-access fentanyl self-administration during protracted withdrawal, when adverse health effects were most pronounced. Using liquid chromatography coupled with electrospray ionization tandem mass spectrometry (LC-ESI MS/MS) we conducted proteomic analysis in OVX rats comparing the effect of fentanyl withdrawal, with or without E2, to drug-naïve control rats.We found a significant effect of fentanyl withdrawal on plasma proteomes within OVX rats. Fentanyl withdrawal was associated with a significant change in 15 plasma proteins including B-factor, properdin (Cfb), apolipoprotein E (ApoE), complement 4, precursor (C4), C-reactive protein (Crp), zinc-alpha-2-glycoprotein precursor (Azgp1), and serine peptidase inhibitor 3L (Serinpa3l). The addition of E2 was associated with enhanced proteomic changes. Bioinformatic gene ontology enrichment analysis indicates that fentanyl withdrawal can disrupt the expression of proteins associated with immunity, lipid transport, and components of the extracellular matrix. We identify protein changes in plasma that may contribute to adverse health outcomes in females, with and without E2, during fentanyl withdrawal. These findings support the development of targeted strategies addressing health risks during opioid use disorder in women.

PMID:40254411 | DOI:10.1111/jne.70033

Cancer Lett. 2025 Apr 18:217733. doi: 10.1016/j.canlet.2025.217733. Online ahead of print.

ABSTRACT

Inflammation is a widely recognized key contributor to KRAS-driven lung adenocarcinoma (LUAD). Tumor-associated macrophages (TAM) are an integral part of the tumor microenvironment and create a supportive niche that sustains inflammation-driven tumorigenesis. In the present study, we unravel a dual role of sphingosine kinase 1 (SPHK1) in KRAS-driven LUAD. While SPHK1 promotes tumorigenesis in in vitro experimental models, it paradoxically suppresses tumorigenesis in in vivo models of KRAS-mutated LUAD. Mechanistically, tumor-intrinsic loss of SPHK1 leads to disrupted lipid homeostasis, increased inflammation and infiltration by TAM, ultimately driving tumor progression. Thus, our study suggests that clinically targeting the SPHK1/S1P axis could potentially result in increased tumor progression, possibly by rewiring the tumor microenvironment toward a more inflammatory and pro-tumorigenic state.

PMID:40254091 | DOI:10.1016/j.canlet.2025.217733

Mol Cell Proteomics. 2025 Apr 18:100974. doi: 10.1016/j.mcpro.2025.100974. Online ahead of print.

ABSTRACT

Relative quantitation, used by most MS-based proteomics laboratories to determine protein fold-changes, requires samples being processed and analyzed together for best comparability through minimizing batch differences. This limits the adoption of MS-based proteomics in population-wide studies, and the detection of subtle but relevant changes in heterogeneous samples. Absolute quantitation circumvents these limitations and enables comparison of results across laboratories, studies, and longitudinally. However, high costs of the essential stable isotope labeled (SIL) standards prevents widespread access and limits the number of quantifiable proteins. Our new approach, called "SysQuan", repurposes SILAC mouse tissues/biofluids as system-wide internal standards for matched human samples to enable absolute quantitation of, theoretically, two-thirds of the human proteome using 157,086 shared tryptic peptides, of which 73,901 with lysine on the c-terminus. We demonstrate that SysQuan enables quantification of 70% and 31% of the liver and plasma proteomes, respectively. We demonstrate for 14 metabolic proteins that abundant SIL mouse tissues enable cost-effective reverse absolute quantitation in, theoretically, 1000s of human samples. Moreover, 10,000s of light/heavy doublets in untargeted SysQuan datasets enable unique post-acquisition absolute quantitation. SysQuan empowers researchers to replace relative quantitation with affordable absolute quantitation at scale, making data comparable across laboratories, diseases and tissues, enabling completely novel study designs and increasing reusability of data in repositories.

PMID:40254065 | DOI:10.1016/j.mcpro.2025.100974

Age Ageing. 2025 Mar 28;54(4):afaf095. doi: 10.1093/ageing/afaf095.

ABSTRACT

BACKGROUND: Ageing increases the risk of treatment-related toxicities (TRT) in patients with cancer. This systematic review provided an overview of existing prediction models for TRT in this population and evaluated their predictive performances.

METHODS: A systematic search was conducted in MEDLINE (Ovid), Embase, PubMed, CINAHL and CENTRAL (Cochrane Central Register of Controlled Trials) databases for studies developing severe TRT prediction models in older cancer patients published between 1 January 2000 and 31 October 2023. The included models were summarised and assessed using the Prediction Model Risk of Bias Assessment Tool (PROBAST).

RESULTS: Out of the 6192 studies identified through literature searching, 12 studies involving 90 819 participants met the inclusion criteria. About 15 prediction models (9 (60%) for diverse cancer types; 6 (40%) for specific cancer types) were analysed. The models included between 4 and 11 variables. The most common predictors were physical function (n = 12, 80%), performance status (n = 5, 33.3%) and the MAX2 index (n = 5, 33.3%). About 2 models (13.3%) had external validation, 9 (60.0%) had internal validation and 6 (40.0%) lacked any validation. All studies were assessed to have a high risk of bias according to the PROBAST criteria.

CONCLUSION: This systematic review demonstrated that existing prediction models for TRT exhibited moderate discrimination ability in older patients with cancer, with significant heterogeneity in clinical settings and predictive variables. Standardised procedures for developing and validating prediction models are essential to improve the prediction of severe TRT in this vulnerable population.

PMID:40253686 | DOI:10.1093/ageing/afaf095

Eur J Med Genet. 2025 Apr 17;75:105016. doi: 10.1016/j.ejmg.2025.105016. Online ahead of print.

ABSTRACT

A systematic review of case reports, case series, and case-control studies was conducted to quantify the diagnostic delay in 84 rare genetic diseases where neuropsychiatric symptoms may be primary or part of the early clinical presentation. Data abstracted from 1221 published articles encompassing 1838 individual cases revealed a mean diagnostic delay of 9.3 ± 8.7 years, with no significant improvement in time to diagnosis over the 65-year period from 1958 to 2023. Subanalysis of the most recent 10 years, 2014-2023, revealed no change in diagnostic delay, even when stratifying by genetic and other diagnostic tests. Neuropsychiatric symptoms were reported in 68 % of the included cases. Following a definitive diagnosis and optimized management of the underlying rare genetic disease, 66 % of individuals experienced an improvement in their neuropsychiatric symptoms. Despite increasing access to, and substantial advancement in, genetic and other testing, diagnostic delays remain lengthy for individuals affected by these rare genetic diseases. This often results in suboptimal management of the associated neuropsychiatric symptoms. Thus, earlier implementation of genetic testing and other diagnostic tools may reduce these delays, improving patient outcomes and alleviating the burden of diagnostic uncertainty.

PMID:40252994 | DOI:10.1016/j.ejmg.2025.105016

Methods. 2025 Apr 17:S1046-2023(25)00103-3. doi: 10.1016/j.ymeth.2025.04.008. Online ahead of print.

ABSTRACT

The global COVID-19 pandemic has drastically affected daily life, emphasizing the urgent need for early and accurate detection to provide adequate medical treatment, especially with limited antiviral options. Chest X-ray imaging has proven crucial for distinguishing COVID-19 from other respiratory conditions, providing an essential diagnostic tool. Deep learning (DL)-based models have proven highly effective in image diagnostics in recent years. Many of these models are computationally intensive and prone to overfitting, especially when trained on limited datasets. Additionally, conventional models often fail to capture multi-scale features, reducing diagnostic accuracy. This paper proposed a highly efficient convolutional neural network (CNN) called EffiCOVID-Net, incorporating diverse feature learning units. The proposed model consists of a bunch of EffiCOVID blocks that incorporate several layers of convolution containing (3×3) filters and recurrent connections to extract complex features while preserving spatial integrity. The performance of EffiCOVID-Net is rigorously evaluated using standard performance metrics on two publicly available COVID-19 chest X-ray datasets. Experimental results demonstrate that EffiCOVID-Net outperforms existing models, achieving 98.68% accuracy on the COVID-19 radiography dataset (D1), 98.55% on the curated chest X-ray dataset (D2), and 98.87% on the mixed dataset (DMix) in multi-class classification (COVID-19 vs. Normal vs. Pneumonia). For binary classification (COVID-19 vs. Normal), the model attains 99.06%, 99.78%, and 99.07% accuracy, respectively. Integrating Grad-CAM-based visualizations further enhances interpretability by highlighting critical regions influencing model predictions. EffiCOVID-Net's lightweight architecture ensures low computational overhead, making it suitable for deployment in resource-constrained clinical settings. A comparative analysis with existing methods highlights its superior accuracy, efficiency, and robustness performance. However, while the model enhances diagnostic workflows, it is best utilized as an assistive tool rather than a standalone diagnostic method.

PMID:40252941 | DOI:10.1016/j.ymeth.2025.04.008

Comput Med Imaging Graph. 2025 Apr 17;123:102560. doi: 10.1016/j.compmedimag.2025.102560. Online ahead of print.

ABSTRACT

Self-supervised learning has shown potential in enhancing deep learning methods, yet its application in brain magnetic resonance imaging (MRI) analysis remains underexplored. This study seeks to leverage large-scale, unlabeled public brain MRI datasets to improve the performance of deep learning models in various downstream tasks for the development of clinical decision support systems. To enhance training efficiency, data filtering methods based on image entropy and slice positions were developed, condensing a combined dataset of approximately 2 million images from fastMRI-brain, OASIS-3, IXI, and BraTS21 into a more focused set of 250 K images enriched with brain features. The Momentum Contrast (MoCo) v3 algorithm was then employed to learn these image features, resulting in robustly pretrained models specifically tailored to brain MRI. The pretrained models were subsequently evaluated in tumor classification, lesion detection, hippocampal segmentation, and image reconstruction tasks. The results demonstrate that our brain MRI-oriented pretraining outperformed both ImageNet pretraining and pretraining on larger multi-organ, multi-modality medical datasets, achieving a ∼2.8 % increase in 4-class tumor classification accuracy, a ∼0.9 % improvement in tumor detection mean average precision, a ∼3.6 % gain in adult hippocampal segmentation Dice score, and a ∼0.1 PSNR improvement in reconstruction at 2-fold acceleration. This study underscores the potential of self-supervised learning for brain MRI using large-scale, tailored datasets derived from public sources.

PMID:40252479 | DOI:10.1016/j.compmedimag.2025.102560

Microb Pathog. 2025 Apr 17:107586. doi: 10.1016/j.micpath.2025.107586. Online ahead of print.

ABSTRACT

Tuberculosis is an infectious disease caused by Mycobacterium tuberculosis (Mtb) and is a growing public health problem worldwide. Within the innate immune response, we highlight the secretion of the antimicrobial peptide LL-37, which is crucial for Mtb elimination in infected cells. Previous reports have shown that CEBPα activation induces LL-37 independently of its main inducer, vitamin D, under endoplasmic reticulum (ER) stress. In this study, we report that infection with Mtb causes ER stress in pulmonary epithelial cells and macrophages. The stress induces the activation of CEBPα, which in turn promotes the LL-37 expression. Furthermore, the participation of CEBPα is necessary for the correct clearance of Mtb in an in vitro infection model. We identify candidate drugs (mycophenolic acid, indapamide, and glibenclamide) capable of activating CEBPα and promoting LL-37 through in silico assays. The effect of the drugs was corroborated by gene and protein expression analysis. Finally, we observed that treatment with these drugs improves bacterial clearance in infected cells. Our results lead us to suggest CEBPα as a potential therapeutic target as an adjuvant in the standard treatment of tuberculosis, seeking a reduction in treatment time, and thus a lower appearance of drug resistance.

PMID:40252936 | DOI:10.1016/j.micpath.2025.107586

Comput Biol Chem. 2025 Apr 15;118:108471. doi: 10.1016/j.compbiolchem.2025.108471. Online ahead of print.

ABSTRACT

Triple-negative breast cancer (TNBC) is an aggressive subtype characterized by the overexpression of poly-ADP ribose polymerase 1 (PARP1), a key enzyme in DNA repair. Targeting PARP1 with inhibitors presents a promising therapeutic strategy, particularly given the limited treatment options for TNBC. This study employed in silico methodologies to evaluate the pharmacokinetic and inhibitory potential of FDA-approved drugs and compounds derived from Onosma. dichroantha root extracts against PARP1. Virtual screening and molecular docking identified Midazolam, Olaparib, Beta-sitosterol, and 1-Hexyl-4-nitrobenzene as top candidates, exhibiting strong binding affinities of -10.6 kcal/mol, -9.9 kcal/mol, -6.83 kcal/mol, and -5.53 kcal/mol respectively. Molecular dynamics simulations (MDS) over 100 nanoseconds revealed that Beta-sitosterol formed the most stable complex with PARP1, demonstrating minimal structural deviations and robust hydrogen bonding. The Molecular Mechanics-Poisson-Boltzmann Surface Area (MM-PBSA) analysis further confirmed Beta-Sitosterol and Olaparib superior binding free energy (ΔGbind= -175.43 kcal/mol and -180.8 kcal/mol respectively), highlighting its potential as a potent PARP1 inhibitor. ADMET (Absorption, Distribution, Metabolism, Excretion, and Toxicity) profiling indicated that Beta-Sitosterol adheres to Lipinski's Rule of Five, with high intestinal absorption (95.88 %) and blood-brain barrier permeability (0.824), despite low water solubility. Protein-protein interaction analysis identified key PARP1-associated proteins, including CASP3, CASP7, and XRCC1, suggesting broader therapeutic implications. These findings underscore the potential of Beta-Sitosterol as a novel PARP1 inhibitor for TNBC treatment, combining computational validation with favorable pharmacokinetic properties. The study also highlights the utility of drug repurposing and plant-derived compounds in developing targeted therapies for TNBC, paving the way for further preclinical and clinical investigations.

PMID:40252254 | DOI:10.1016/j.compbiolchem.2025.108471

Commun Chem. 2025 Apr 19;8(1):118. doi: 10.1038/s42004-025-01519-w.

NO ABSTRACT

PMID:40253564 | DOI:10.1038/s42004-025-01519-w

J Mol Biol. 2025 Apr 17:169160. doi: 10.1016/j.jmb.2025.169160. Online ahead of print.

ABSTRACT

There is growing research on the allosteric behaviour of proteins, including studies on allosteric mutations that contribute to human diseases and the development of allosteric drugs. Allostery also plays a key role in drug metabolism, an essential factor in drug development. However, population specific variations, particularly in 3D protein structures, remain understudied. This study focuses on CYP3A4, a key enzyme responsible for metabolizing over 50% of FDA-approved drugs and often linked to adverse drug reactions. Given the vast genetic diversity of Africa, we investigated 13 CYP3A4 alleles from African populations using post-molecular dynamics analyses, with 12 being single variations and one containing a double variation. Except for one, all allele variations were located away from the active site, suggesting allosteric effects. Our comparative analyses of reference and variant structures, through hydrogen bond interactions, dynamic residue network analysis and substrate channel dynamics, revealed notable differences at both global and residue levels. The *32-I335T variant showed the largest changes compared to the reference structure, while *3-M445T (near normal metabolizer) exhibited the least change, with other variants falling in between. The *32-I335T variant showed a distorted conformation in the radius of gyration, a distinct kink in the I helix with specific hydrogen bonds and altered channel patterns. The *12-L373F variant, associated with reduced metabolism of midazolam and quinine, showed increased rigidity in its vicinity, potentially interfering with catalytic activity. Our findings align with clinical and wet lab data, suggesting that our approaches could be applied to analyse variants without clinical evidence.

PMID:40252954 | DOI:10.1016/j.jmb.2025.169160

Epilepsy Behav. 2025 Apr 18;168:110434. doi: 10.1016/j.yebeh.2025.110434. Online ahead of print.

ABSTRACT

BACKGROUND: Drug-resistant epilepsy (DRE) presents a significant clinical challenge since many patients fail to respond adequately to pharmacological treatments, resulting in persistent seizures and a great decline in quality of life. This highlights the urgent need for alternative or adjunctive therapeutic strategies. Carbohydrate-restricted diets have emerged as promising adjunctive treatments for epilepsy. However, while the efficacy of these diets has been well-established in pediatric populations, their effectiveness on adult DRE patients remains underexplored. This study aims to evaluate and compare the efficacy of various carbohydrate-restricted diets in treating DRE among all age groups, providing valuable insights into their potential clinical applications.

METHODS: PRISMA guidelines for network meta-analysis were followed. Randomized controlled trials (RCTs) comparing the efficacy of different carbohydrate-restricted diets in DRE patients, and published in PubMed, Embase, Cochrane, and Web of Science up to 22 December 2023 were screened. The primary outcomes were >50 %, >90 % seizure frequency reduction from the baseline and seizure freedom. Secondary outcomes included compliance and adverse events. Random-effects models with a Bayesian-based approach were employed to estimate between-group comparisons, with results presented as odds ratios (OR) and 95 % credible intervals (CrI).

RESULTS: A total of 17 RCTs involving 1468 DRE patients were included. The diets evaluated were the ketogenic diet (KD), modified Atkins diet (MAD), and low glycemic index treatment (LGIT). For >50 % and >90 % seizure reduction from baseline, all three diets resulted in significant efficacy compared to the normal diet. Notably, MAD was the only diet that demonstrated a statistically significant association with seizure freedom (OR 7.36, 95 % CrI 2.21-60.36), compared to the normal diet, while its lower compliance (OR 0.39, 95 % CrI 0.18-0.76) was likely influenced by the inclusion of adult subjects. Adverse effects were reported across all three diets with similar profiles, highlighting the need for individualized monitoring.

CONCLUSIONS: This meta-analysis indicated that in RCTs, the included diet therapies were overall equivalent in efficacy and side effects, with the MAD showing a higher chance of seizure freedom. Compliance was lower with the MAD, but this was likely due to a preponderance of adult studies using this therapy.

PMID:40252525 | DOI:10.1016/j.yebeh.2025.110434

Nat Commun. 2025 Apr 19;16(1):3729. doi: 10.1038/s41467-025-59134-1.

ABSTRACT

Disease tolerance mitigates organ damage from non-resolving inflammation during persistent infections, yet its underlying mechanisms remain unclear. Here we show, in a Pseudomonas aeruginosa pneumonia mouse model, that disease tolerance depends on the mitochondrial metabolite itaconate, which mediates cooperative host-pathogen interactions. In P. aeruginosa, itaconate modifies key cysteine residues in TCA cycle enzymes critical for succinate metabolism, inducing bioenergetic stress and promoting the formation biofilms that are less immunostimulatory and allow the bacteria to integrate into the local microbiome. Itaconate incorporates into the central metabolism of the biofilm, driving exopolysaccharide production-particularly alginate-which amplifies airway itaconate signaling. This itaconate-alginate interplay limits host immunopathology by enabling pulmonary glutamine assimilation, activating glutaminolysis, and thereby restrain detrimental inflammation caused by the inflammasome. Clinical sample analysis reveals that P. aeruginosa adapts to this metabolic environment through compensatory mutations in the anti-sigma-factor mucA, which restore the succinate-driven bioenergetics and disrupt the metabolic synchrony essential for sustaining disease tolerance.

PMID:40253414 | DOI:10.1038/s41467-025-59134-1

J Cyst Fibros. 2025 Apr 18:S1569-1993(25)00766-0. doi: 10.1016/j.jcf.2025.03.670. Online ahead of print.

ABSTRACT

AIM: Cystic fibrosis (CF) care increasingly demands flexible and personalised approaches, particularly with the growing role of telemedicine in disease management. This study aimed to evaluate the use of home-based spirometry and antibiotic monitoring for assessing lung function trends and treatment patterns in individuals with CF.

METHOD: Individuals aged 0-25 years from seven Swedish CF centres participated in 12 months of routine CF care, digitally recording antibiotic usage and performing home spirometry (aged ≥5 years). Home spirometry sessions were graded according to ATS/ERS criteria, with A-C representing high-quality sessions. Longitudinal FEV1 trends from home and hospital spirometry were analysed using linear mixed effects models, adjusting for clinical stability.

RESULTS: Of 126 invited participants, 110 were enrolled and followed for a median (range) duration of 12 months (9-17). A total of 779 usable home spirometry sessions were conducted, with 388 sessions (50 %) from 80 out of 95 (84 %) participants aged ≥5 years graded as high-quality. Mean (95 % CI) FEV1 was (86-92 %) for home spirometry and 88 % (85-90 %) for hospital spirometry. After adjusting for clinical stability and including only high-quality home spirometry data, the mean difference was 0.6 % (-3.8 %-5.0 %, p=0.78). The mean annual rate of FEV1 decline was -0.48 % (-1.29-0.32 %) for home spirometry and -0.18 % (-0.77-0.41 %) for hospital spirometry, with no statistically significant difference.

CONCLUSION: High-quality home spirometry measurements, adjusted for clinical stability and antibiotic usage, may provide lung function levels and trends closely comparable to hospital spirometry.

PMID:40253216 | DOI:10.1016/j.jcf.2025.03.670

Theor Appl Genet. 2025 Apr 19;138(5):103. doi: 10.1007/s00122-025-04894-z.

ABSTRACT

Enhancing early selection through genomic estimated breeding values is pivotal for reducing generation intervals and accelerating breeding programs. Recently, deep learning (DL) approaches have gained prominence in genomic prediction (GP). Here, we introduce a novel DL framework for GP based on Elastic Net feature selection and bidirectional encoder representations from transformer's embedding and multi-head attention pooling (EBMGP). EBMGP applies Elastic Net for the selection of features, thereby diminishing the computational burden and bolstering the predictive accuracy. In EBMGP, SNPs are treated as "words," and groups of adjacent SNPs with similar LD levels are considered "sentences." By applying bidirectional encoder representations from transformers embeddings, this method models SNPs in a manner analogous to human language, capturing complex genetic interactions at both the "word" and "sentence" scales. This flexible representation seamlessly integrates into any DL network and demonstrates a marked improvement in predictive performance for EBMGP and SoyDNGP compared to the widely used one-hot representation. We propose multi-head attention pooling, which can adaptively assign weights to features while learning features from multiple subspaces through multi-heads for a high level of semantic understanding. In a comprehensive comparative analysis across four diverse plant and animal datasets, EBMGP outperformed competing models in 13 out of 16 tasks, achieving accuracy gains ranging from 0.74 to 9.55% over the second-best model. These results underscore EBMGP's robustness in genomic prediction and highlight its potential for deep learning applications in life sciences.

PMID:40253568 | DOI:10.1007/s00122-025-04894-z

Sci Rep. 2025 Apr 19;15(1):13606. doi: 10.1038/s41598-025-98138-1.

ABSTRACT

To enhance bearing fault diagnosis performance under various operating conditions, this paper proposes a hybrid approach based on generative adversarial networks (GANs), transfer learning, wavelet transform time-frequency representations, asymmetric convolutional networks, and the multi-head attention mechanism (MAC-MHA). Firstly, GANs are utilized to generate new bearing fault data to meet the model's training requirements. Then, wavelet transform is applied to convert the bearing vibration signals into time-frequency representations, capturing the temporal evolution of frequency components. Next, an improved asymmetric convolutional network (MAC-MHA), combined with the multi-head attention mechanism, is employed to enhance the focus on key time-frequency features, further improving fault diagnosis accuracy. Considering the differences in operating conditions, transfer learning techniques are applied to facilitate knowledge transfer from the source domain to the target domain, thereby enhancing the model's generalization ability. Experimental results demonstrate the effectiveness and robustness of the proposed method under various operating conditions. Finally, the proposed hybrid fault diagnosis approach is validated using the PADERBORN and CWRU datasets.

PMID:40253550 | DOI:10.1038/s41598-025-98138-1

Sci Rep. 2025 Apr 19;15(1):13591. doi: 10.1038/s41598-025-96336-5.

ABSTRACT

Artificial intelligence (AI) has the potential to improve healthcare and patient safety and is currently being adopted across various fields of medicine and healthcare. AI and in particular computer vision (CV) are well suited to the analysis of minimally invasive surgical simulation videos for training and performance improvement. CV techniques have rapidly improved in recent years from accurately recognizing objects, instruments, and gestures to phases of surgery and more recently to remembering past surgical steps. Lack of labeled data is a particular problem in surgery considering its complexity, as human annotation and manual assessment are both expensive in time and cost, and in most cases rely on direct intervention of clinical expertise. In this study, we introduce a newly collected simulated Laparoscopic Surgical Performance Dataset (LSPD) specifically designed to address these challenges. Unlike existing datasets that focus on instrument tracking or anatomical structure recognition, the LSPD is tailored for evaluating simulated laparoscopic surgical skill performance at various expertise levels. We provide detailed statistical analyses to identify and compare poorly performed and well-executed operations across different skill levels (novice, trainee, expert) for three specific skills: stack, bands, and tower. We employ a 3-dimensional convolutional neural network (3DCNN) with a weakly-supervised approach to classify the experience levels of surgeons. Our results show that the 3DCNN effectively distinguishes between novices, trainees, and experts, achieving an F1 score of 0.91 and an AUC of 0.92. This study highlights the value of the LSPD dataset and demonstrates the potential of leveraging 3DCNN-based and weakly-supervised approaches to automate the evaluation of surgical performance, reducing reliance on manual expert annotation and assessments. These advancements contribute to improving surgical training and performance analysis.

PMID:40253514 | DOI:10.1038/s41598-025-96336-5