Sci Rep. 2025 Feb 15;15(1):5594. doi: 10.1038/s41598-025-90261-3.

ABSTRACT

Fine particulate matter 2.5 (PM2.5) is a prevalent atmospheric pollutant that is closely associated with asthma. Elderly patients have a high incidence of asthma with a long course of illness. Our previous studies revealed that exposure to PM2.5 diminishes lung function and exacerbates lung damage in elderly rats. In the present study, we investigated whether PM2.5 exposure influences susceptibility to allergic asthma in elderly rats. Brown-Norway elderly rats were treated with ovalbumin (OVA) for different durations before and after PM2.5 exposure. The results from pulmonary function tests and histopathology indicated that early exposure to allergens prior to PM2.5 exposure increased susceptibility to airway hyperresponsiveness and led to severe lung injury in elderly asthmatic rats. Cytokine microarray analysis demonstrated that the majority of cytokines and chemokines were upregulated in OVA-treated rats before and after PM2.5 exposure. Cytological examination showed no change in eosinophil (EOS) counts, yet the amounts of neutrophils (NEU), white blood cells (WBC), lymphocytes (LYM), and monocytes (MON) in the lung lavage fluid of OVA-treated rats were significantly higher than those in control rats before and after PM2.5 exposure, suggesting that PM2.5 affects noneosinophilic asthma in elderly rats. ELISA results from the plasma and lung lavage fluid revealed that the levels of IgG1, IgE, IgG2a and IgG2b were significantly elevated in OVA-treated rats, whereas the level of IgG2b in the lung lavage fluid was significantly lower in rats treated with OVA prior to PM2.5 exposure compared to those treated afterward. A non-targeted metabolomic analysis of plasma identified 202 metabolites, among which 31 metabolites were differentially abundant. Ten metabolites and 11 metabolic pathways were uniquely detected in OVA-treated rats before PM2.5 exposure. Specifically, there were positive or negative correlations between the levels of Th2-associated cytokines (IL-4, IL-5, and IL-13) and six metabolites in the OVA-treated group before PM2.5 exposure, whereas the levels of IL-4 and IL-5 were negatively correlated with five metabolites in the OVA-treated group after PM2.5 exposure. Our findings suggest that PM2.5 exposure could influence the susceptibility of allergic asthma in response to allergens in elderly rats, potentially through changes in plasma metabolites.

PMID:39955443 | DOI:10.1038/s41598-025-90261-3

Sci Rep. 2025 Feb 15;15(1):5643. doi: 10.1038/s41598-025-89667-w.

ABSTRACT

Monoclonal antibodies (mAbs) are considered one of the most game-changing products of the biopharmaceutical industry. The introduction of several diverse and complex formats consisting of several polypeptide chains and engineered with multiple antigen-binding domains has made the manufacturability process particularly challenging, especially in the context of assessing expression levels and yields of the formats. Here we present the largest and most diversified CHO transcriptomics analysis consisting of data derived from 892 different monoclonal cell lines, producing 11 different mAbs with various non-standard, highly complex formats. We apply three robust feature selection methods, one traditional differential expression analysis and two machine learning approaches to identify genes correlated to high product titer and quality. Cnpy3 gene is identified as a novel gene biomarker, showing a very strong negative correlation (Pearson r2 = 0.94) to the overall format productivity. These results were validated by a hold-out data set from cell lines expressing two different antibody formats. Additionally, the expression of Cnpy3 gene is positively correlated to the structural complexity of the examined mAbs. As complexity increases, cellular stress escalates leading to reduced productivity, implicating Cnpy3 as a strong CHO cell lines stress indicator. Thus, we conclude that Cnpy3 gene has the potential to be used as a screening biomarker for assessing format manufacturability and selecting formats and pools with a high potential to deliver subsequent higher productivity rates, resulting in a substantially smarter cell line and process development.

PMID:39955392 | DOI:10.1038/s41598-025-89667-w

Nat Commun. 2025 Feb 15;16(1):1671. doi: 10.1038/s41467-025-56760-7.

ABSTRACT

Cryogenic electron microscopy (cryo-EM), X-ray crystallography, and nuclear magnetic resonance (NMR) contribute structural data that are interchangeable, cross-verifiable, and visualizable on common platforms, making them powerful tools for our understanding of protein structures. Unfortunately, atomic force microscopy (AFM) has so far failed to interface with these structural biology methods, despite the recent development of localization AFM (LAFM) that allows extracting high-resolution structural information from AFM data. Here, we build on LAFM and develop a pipeline that transforms AFM data into 3D-density files (.afm) that are readable by programs commonly used to visualize, analyze, and interpret structural data. We show that 3D-LAFM densities can serve as force fields to steer molecular dynamics flexible fitting (MDFF) to obtain structural models of previously unresolved states based on AFM observations in close-to-native environment. Besides, the .afm format enables direct 3D or 2D visualization and analysis of conventional AFM images. We anticipate that the file format will find wide usage and embed AFM in the repertoire of methods routinely used by the structural biology community, allowing AFM researchers to deposit data in repositories in a format that allows comparison and cross-verification with data from other techniques.

PMID:39955301 | DOI:10.1038/s41467-025-56760-7

Trends Biotechnol. 2025 Feb 14:S0167-7799(25)00006-X. doi: 10.1016/j.tibtech.2025.01.006. Online ahead of print.

ABSTRACT

Valorization of lignocellulosic biomass for sustainable production of high-value chemicals is challenged by the complexity of lignin, a phenolic biopolymer. Beyond the classical lignin monomers derived from p-coumaryl, coniferyl, and sinapyl alcohol, grass lignins incorporate substantial amounts of monolignol p-coumarates that are produced by p-COUMAROYL-CoA:MONOLIGNOL TRANSFERASE (PMT). Here, the CRISPR/Cas9-mediated mutation of ZmPMT1 in maize enabled the design of biomass depleted in p-coumaroylated lignin and enriched in guaiacyl lignin. Lignin-first biorefining of stem biomass from zmpmt1 mutants by reductive catalytic fractionation (RCF) generated a lignin oil depleted in carboxylates and enriched in guaiacyl-derived alcohols, which are desirable substrates for bio-based polyurethane synthesis. The reported lignin engineering in maize is a promising strategy for designing a dual-purpose crop, providing both food and feed, along with a renewable feedstock for the production of plant-based chemicals.

PMID:39955231 | DOI:10.1016/j.tibtech.2025.01.006

Metab Eng. 2025 Feb 13:S1096-7176(25)00012-6. doi: 10.1016/j.ymben.2025.02.003. Online ahead of print.

ABSTRACT

Driven by the urgent need to reduce the reliance on fossil fuels and mitigate environmental impacts, microbial cell factories capable of producing value-added products from renewable resources have gained significant attention over the past few decades. Notably, non-model yeasts with unique physiological characteristics have emerged as promising candidates for industrial applications, particularly for the production of organic acids. Among them, Issatchenkia orientalis stands out for its exceptional natural tolerance to low pH and high osmotic pressure, traits that are critical for overcoming the limitations of conventional microbial organisms. The acid tolerance of I. orientalis enables organic acid production under low pH conditions, bypassing the need for expensive neutral pH control typically required in conventional processes. Organic acids produced by I. orientalis, such as lactic acid, succinic acid, and itaconic acid, are widely used as building blocks for bioplastics, food additives, and pharmaceuticals. This review summarizes the key findings from systems biology studies on I. orientalis over the past two decades, providing insights into its unique metabolic and physiological traits. Advances in genetic tool development for this non-model yeast are also discussed, enabling targeted metabolic engineering to enhance its production capabilities. Additionally, case studies are highlighted to illustrate the potential of I. orientalis as a platform organism. Finally, the remaining challenges and future directions are addressed to further develop I. orientalis into a robust and versatile microbial cell factory for sustainable biomanufacturing.

PMID:39954846 | DOI:10.1016/j.ymben.2025.02.003

J Allergy Clin Immunol Pract. 2025 Feb 13:S2213-2198(25)00164-3. doi: 10.1016/j.jaip.2025.01.039. Online ahead of print.

ABSTRACT

BACKGROUND: α1-antitrypsin deficiency is caused by rare pathogenic variants in SERPINA1, the strongest genetic risk factor for COPD. Few studies have evaluated the effects of SERPINA1 variation on asthma severity accounting for critical gene-by-environment interactions with smoking.

OBJECTIVE: To characterize the influence of SERPINA1 variation on asthma severity.

METHODS: DNA samples from 847 non-Hispanic whites and 446 African Americans from the Severe Asthma Research Program underwent SERPINA1 resequencing to identify rare variants. An independent population of 1,955 individuals with asthma and α1-antitrypsin concentrations from a Cleveland Clinic Health System (CCHS) database were evaluated for severity measures.

MEASUREMENTS AND MAIN RESULTS: In whites, a history of minimum smoking significantly interacted with SERPINA1 low-to-rare frequency variation to determine risk for asthma-related healthcare utilization. This was attributed to PI type Z heterozygotes (MZ, N=11) who had a higher frequency of ED visits (6 [54.5%] MZ heterozygotes, OR=7.60, 95%CI=1.71-39.7, p=0.010), hospitalization (5 [45.5%], OR=16.1, 95%CI=2.64-150.4, p=0.0050) in the past year, and lifetime ICU admissions (6 [54.5%], OR=12.5, 95%CI=2.44-75.6, p=0.0032) compared to 146 individuals without SERPINA1 variants (30 [20.5%] reporting ED visits, 17 [11.6%] hospitalization, 15 [10.3%] ICU admission). SERPINA1 variant-ever smoking interactions in African Americans for ED visits (p=0.069) related to four of six compound heterozygotes reporting an ED visit. In CCHS, α1-antitrypsin concentrations were inversely associated with moderate-to-severe asthma risk (OR=0.97 per 10 mg/dL increase in α1-antitrypsin, 95%CI=0.94-0.99, p=0.010) and exacerbations (OR=0.84 per 10 mg/dL, 95%CI=0.76-0.94, p=0.002).

CONCLUSIONS: SERPINA1 variation and α1-antitrypsin concentrations impact asthma severity through gene-environment interactions with minimum smoking.

PMID:39954727 | DOI:10.1016/j.jaip.2025.01.039

Cell Stem Cell. 2025 Feb 12:S1934-5909(25)00010-4. doi: 10.1016/j.stem.2025.01.010. Online ahead of print.

ABSTRACT

Acute myeloid leukemia (AML) is an aggressive hematopoietic malignancy characterized by the blockage of myeloid cell differentiation and uncontrolled proliferation of immature myeloid cells. Here, we show that paraspeckle component 1 (PSPC1) is aberrantly overexpressed and associated with poor survival in AML patients. Using human AML cells and mouse models, we demonstrate that PSPC1 is not required for normal hematopoiesis, but it is critical and essential for AML cells to maintain their leukemic characteristics. PSPC1 loss induces robust differentiation, suppresses proliferation, and abolishes leukemogenesis in diverse AML cells. Mechanistically, PSPC1 exerts a pro-leukemia effect by regulating a unique leukemic transcription program via cooperative chromatin binding with PU.1 and activation of tumor-promoting genes, including NDC1, which is not previously implicated in AML. Our findings uncover a unique and crucial role of PSPC1 dependency in AML and highlight its potential as a promising therapeutic target for AML.

PMID:39954676 | DOI:10.1016/j.stem.2025.01.010

Cell Rep Methods. 2025 Feb 10:100984. doi: 10.1016/j.crmeth.2025.100984. Online ahead of print.

ABSTRACT

Probabilistic graphical models are powerful tools to quantify, visualize, and interpret network dependencies in complex biological systems such as high-throughput -omics. However, many graphical models assume sample homogeneity, limiting their effectiveness. We propose a flexible Bayesian approach called graphical regression (GraphR), which (1) incorporates sample heterogeneity at different scales through a regression-based formulation, (2) enables sparse sample-specific network estimation, (3) identifies and quantifies potential effects of heterogeneity on network structures, and (4) achieves computational efficiency via variational Bayes algorithms. We illustrate the comparative efficiency of GraphR against existing state-of-the-art methods in terms of network structure recovery and computational cost across multiple settings. We use GraphR to analyze three multi-omic and spatial transcriptomic datasets to investigate inter- and intra-sample molecular networks and delineate biological discoveries that otherwise cannot be revealed by existing approaches. We have developed a GraphR R package along with an accompanying Shiny App that provides comprehensive analysis and dynamic visualization functions.

PMID:39954675 | DOI:10.1016/j.crmeth.2025.100984

Cell Rep Methods. 2025 Feb 10:100985. doi: 10.1016/j.crmeth.2025.100985. Online ahead of print.

ABSTRACT

Cell-cell interaction (CCI) networks are key to understanding disease progression and treatment response. However, existing methods for inferring these networks often aggregate data across patients or focus on cell-type level interactions, providing a generalized overview but overlooking patient heterogeneity and local network structures. To address this, we introduce "random cell-cell interaction generator" (RaCInG), a model based on random graphs to derive personalized networks leveraging prior knowledge on ligand-receptor interactions and bulk RNA sequencing data. We applied RaCInG to 8,683 cancer patients to extract 643 network features related to the tumor microenvironment and unveiled associations with immune response and subtypes, enabling prediction and explanation of immunotherapy responses. RaCInG demonstrated robustness and showed consistencies with state-of-the-art methods. Our findings highlight RaCInG's potential to elucidate patient-specific network dynamics, offering insights into cancer biology and treatment responses. RaCInG is poised to advance our understanding of complex CCI s in cancer and other biomedical domains.

PMID:39954673 | DOI:10.1016/j.crmeth.2025.100985

Cell Rep Methods. 2025 Feb 10:100990. doi: 10.1016/j.crmeth.2025.100990. Online ahead of print.

ABSTRACT

The need for a deeper understanding of adverse drug reaction (ADR) mechanisms is vital for improving drug safety and repurposing. This study introduces Drug Adverse Reaction Mechanism Explainer (DREAMER), a network-based framework that uses a comprehensive knowledge graph to uncover molecular mechanisms underlying ADRs and disease phenotypes. By examining shared phenotypes of drugs and diseases and their effects on protein-protein interaction networks, DREAMER identifies proteins linked to ADR mechanisms. Applied to 649 ADRs, DREAMER identified molecular mechanisms for 67 ADRs, including ventricular arrhythmia and metabolic acidosis, and emphasized pathways like GABAergic signaling and coagulation proteins in personality disorders and intracranial hemorrhage. We further demonstrate the application of DREAMER in drug repurposing and propose sotalol, ranolazine, and diltiazem as candidate drugs to be repurposed for cardiac arrest. In summary, DREAMER effectively detects molecular mechanisms underlying phenotypes, emphasizing the importance of network-based analyses with integrative data for enhancing drug safety and accelerating the discovery of novel therapeutic strategies.

PMID:39954672 | DOI:10.1016/j.crmeth.2025.100990

Biosens Bioelectron. 2025 Feb 6;276:117228. doi: 10.1016/j.bios.2025.117228. Online ahead of print.

ABSTRACT

Hepatocellular carcinoma (HCC) poses a significant global health burden, with escalating incidence rates and substantial mortality. The predominant etiological factors include liver cirrhosis (LC) and chronic hepatitis B infections (CHB). Surveillance primarily relies on ultrasound and Alpha-fetoprotein (AFP), yet their efficacy, particularly in early HCC detection, is limited. Hence, there is a critical need for accurate non-invasive biomarkers to enhance surveillance and early diagnosis. Extracellular vesicles (EVs) hold promises as stable carriers of signaling molecules, offering potential in tumor diagnosis. Our study developed a novel tidal microfluidic chip for label-free EV isolation, enabling rapid and efficient enrichment from small plasma volumes. Through transcriptome sequencing and single-cell analysis, we identified HMMR and B4GALT2 as promising HCC-associated biomarkers in EVs. In a comprehensive clinical evaluation, bi-mRNAs in EVs exhibited superior diagnostic performance over AFP, particularly in distinguishing early-stage HCC or AFP-negative cases from high-risk individuals (CHB/LC). Notably, our study demonstrated the potential of bi-mRNAs to complement imaging examinations, enabling early detection of HCC lesions. In conclusion, the tidal microfluidic chip offers a practical solution for EV isolation, with the integration of EV-based biomarkers presenting opportunities for improved early detection and management of HCC in clinical practice.

PMID:39954520 | DOI:10.1016/j.bios.2025.117228

Europace. 2025 Feb 15:euaf034. doi: 10.1093/europace/euaf034. Online ahead of print.

NO ABSTRACT

PMID:39953950 | DOI:10.1093/europace/euaf034

J Inherit Metab Dis. 2025 Mar;48(2):e70013. doi: 10.1002/jimd.70013.

ABSTRACT

The Leloir pathway was elucidated decades ago, unraveling how galactose is metabolized in the body. Different inborn errors of metabolism in this pathway are known, the most frequent and well-studied being Classic Galactosemia (CG) (OMIM 230400) due to pathogenic variants in the GALT gene. Substrate reduction using dietary restriction of galactose is currently the only available treatment option. Although this burdensome diet resolves the life-threatening clinical picture in neonates, patients still face long-term complications, including cognitive and neurological deficits as well as primary ovarian insufficiency. Emerging therapies aim to address these challenges on multiple fronts: (1) restoration of GALT activity with nucleic acid therapies, pharmacological chaperones, or enzyme replacement; (2) influencing the pathological cascade of events to prevent accumulation of metabolites (Galactokinase 1 (GALK1) inhibitors, aldose reductase inhibitors), address myo-inositol deficiency, or alleviate cellular stress responses; (3) substrate reduction with synthetic biotics or galactose uptake inhibitors to eliminate the need for lifelong diet; and (4) novel approaches to mitigate existing symptoms, such as non-invasive brain stimulation and reproductive innovations. Early, personalized intervention remains critical for optimizing patient outcomes. We review the advances in the development of different treatment modalities for CG and reflect on the factors that need to be considered and addressed to reshape the landscape of treatment.

PMID:39953772 | DOI:10.1002/jimd.70013

Clin Transl Sci. 2025 Feb;18(2):e70080. doi: 10.1111/cts.70080.

ABSTRACT

Selection of rational antagonists of P2Y12 receptor for CAD patients who inherit CYP2C19 LoF alleles remains still conflicting. This study compared the clinical outcomes in CAD patients inheriting CYP2C19 LoF alleles undergoing PCI and treated with clopidogrel against alternative antagonists of P2Y12 receptor. A thorough literature search was performed across multiple scientific databases following the PRISMA guidelines and PICO model. Setting the statistical significance at p < 0.05 and RevMan software was used to calculate the risk ratios (RRs). Estimation of the pooled analysis revealed a significant 62% increased risk of major adverse cardiovascular events (MACE) in CAD patients inheriting CYP2C19 LoF alleles and treated with clopidogrel against those treated with alternative P2Y12 receptor antagonists such as prasugrel or ticagrelor (RR 1.62; 95% CI 1.42-1.86; p < 0.00001). In addition, Asian CAD patients were found at a significantly higher risk of MACE (RR 1.93; 95% CI: 1.49-2.49; p < 0.00001) juxtaposed to CAD patients of other ethnicities (RR 1.51; 95% CI: 1.29-1.78; p < 0.00001). Conversely, between these two treatment groups, taking clopidogrel against prasugrel/ticagrelor, who possess CYP2C19 LoF alleles, no significant differences in bleeding events were observed (RR 0.94; 95% CI 0.79-1.11; p = 0.47). CAD patients undergoing PCI who inherited CYP2C19 LoF alleles and treated with clopidogrel were associated with significantly higher risk of MACE against those treated with alternative antagonists of P2Y12 receptor, that is, prasugrel or ticagrelor.

PMID:39953666 | DOI:10.1111/cts.70080

Cancer Med. 2025 Feb;14(4):e70688. doi: 10.1002/cam4.70688.

ABSTRACT

BACKGROUND: Individualized dose prediction is critical for optimizing radiation treatment planning. This study introduces DESIRE, an enhanced UNet-based dose prediction model with progrEssive feature fuSion and dIfficult Region lEarning, tailored for nasopharyngeal carcinoma (NPC) patients receiving volumetric modulated arc therapy. We aimed to assess the impact of integrating DESIRE into the treatment planning process to improve plan quality.

METHODS: This retrospective study included 131 NPC patients diagnosed at Jiangxi Cancer Hospital between 2017 and 2020. Twenty patients were randomly allocated to a testing cohort, while the remaining 111 comprised a training cohort. Target delineation included three planning target volumes (PTVs): PTV70, PTV60, and PTV55, along with several organs at risk (OARs). The DESIRE model predicted dose distributions, and discrepancies between DESIRE's predictions and the ground truth (GT) were quantified using dosimetric metrics and gamma pass rates. Two junior physicians used DESIRE's predictions for treatment planning, and their plans were compared to the GT.

RESULTS: Most of DESIRE's predicted dosimetric metrics closely aligned with GT (mean difference < 1 Gy), with no significant differences (p > 0.05) in Dmean and D1 values across OARs. While significant differences were observed in PTV metrics, the mean differences in D98, D95, D50, and Dmean between DESIRE and GT did not exceed 1 Gy. Assisted by DESIRE, the junior physicians' plans were comparable to the GT in nearly all OARs, with no significant differences in dosimetric metrics. The conformity index (CI) and homogeneity index (HI) for PTV70 surpassed the GT (0.847 ± 0.036 vs. 0.827 ± 0.037 for CI, and 0.057 ± 0.009 vs. 0.052 ± 0.008 for HI). The average three-dimensional gamma passing rates were 0.85 for PTV70 and 0.87 for the 35-Gy isodose line.

CONCLUSIONS: The DESIRE model shows promise for patient-specific dose prediction, enhancing junior physicians' treatment planning capabilities and improving plan quality.

PMID:39953816 | DOI:10.1002/cam4.70688

New Phytol. 2025 Feb 15. doi: 10.1111/nph.70009. Online ahead of print.

ABSTRACT

Stomata regulate plant gas exchange via repeated turgor-driven changes of guard cell shape, thereby adjusting pore apertures. Grasses, which are among the most widespread plant families on the planet, are distinguished by their unique stomatal structure, which is proposed to have significantly contributed to their evolutionary and agricultural success. One component of their structure, which has received little attention, is the presence of a discontinuous adjoining cell wall of the guard cell pair. Here, we demonstrate the presence of these symplastic connections in a range of grasses and use finite element method simulations to assess hypotheses for their functional significance. Our results show that opening of the stomatal pore is maximal when the turgor pressure in dumbbell-shaped grass guard cells is equal, especially under the low pressure conditions that occur during the early phase of stomatal opening. By contrast, we demonstrate that turgor pressure differences have less effect on the opening of kidney-shaped guard cells, characteristic of the majority of land plants, where guard cell connections are rarely or not observed. Our data describe a functional mechanism based on cellular mechanics, which plausibly facilitated a major transition in plant evolution and crop development.

PMID:39953834 | DOI:10.1111/nph.70009

Sci Rep. 2025 Feb 14;15(1):5473. doi: 10.1038/s41598-025-89202-x.

ABSTRACT

Speech emotion recognition has seen a surge in transformer models, which excel at understanding the overall message by analyzing long-term patterns in speech. However, these models come at a computational cost. In contrast, convolutional neural networks are faster but struggle with capturing these long-range relationships. Our proposed system, MemoCMT, tackles this challenge using a novel "cross-modal transformer" (CMT). This CMT can effectively analyze local and global speech features and their corresponding text. To boost efficiency, MemoCMT leverages recent advancements in pre-trained models: HuBERT extracts meaningful features from the audio, while BERT analyzes the text. The core innovation lies in how the CMT component utilizes and integrates these audio and text features. After this integration, different fusion techniques are applied before final emotion classification. Experiments show that MemoCMT achieves impressive performance, with the CMT using min aggregation achieving the highest unweighted accuracy (UW-Acc) of 81.33% and 91.93%, and weighted accuracy (W-Acc) of 81.85% and 91.84% respectively on benchmark IEMOCAP and ESD corpora. The results of our system demonstrate the generalization capacity and robustness for real-world industrial applications. Moreover, the implementation details of MemoCMT are publicly available at https://github.com/tpnam0901/MemoCMT/ for reproducibility purposes.

PMID:39953105 | DOI:10.1038/s41598-025-89202-x

Sci Rep. 2025 Feb 14;15(1):5561. doi: 10.1038/s41598-024-82027-0.

ABSTRACT

Parkinson's disease is a neurological disorder, and early diagnosis is crucial for the treatment and quality of life of patients. Gait movement disorder is a significant manifestation of PD, and automated gait assessment is key to achieving automated detection of PD patients. With the development of deep learning, in order to improve the accuracy of early Parkinson's disease detection and enhance the robustness of motion recognition models, this study introduces an innovative deep learning approach, namely Multi-area Attention Spatiotemporal Directed Graph Convolutional Network (Ma-ST-DGN). The model effectively captures temporal and spatial information from the movement data of subjects to better understand subtle movement abnormalities in patients. Simultaneously, by reconstructing human skeleton features using directed graphs and introducing a multi-area self-attention mechanism, the model can adaptively focus on key information in different areas and apply more effective fusion strategies on features from different areas, thereby increasing sensitivity to potential signs of Parkinson's disease. By more effectively integrating global and local area information, the model captures subtle manifestations of PD. We use the first Parkinson's disease gait dataset, PD-Walk, consisting of walking videos of 95 PD patients and 96 healthy individuals. Extensive experiments on this clinical video dataset demonstrate that the model achieves the best performance to date, with an accuracy of 88.7%, far superior to existing sensor and vision-based Parkinson's gait assessment methods. Therefore, the method proposed in this study may be effective for early diagnosis of PD in clinical practice.

PMID:39952991 | DOI:10.1038/s41598-024-82027-0

Nat Commun. 2025 Feb 14;16(1):1651. doi: 10.1038/s41467-025-56832-8.

ABSTRACT

For the intricate and infrequent safety issues of batteries, online safety fault diagnosis over stochastic working conditions is indispensable. In this work, we employ deep learning methods to develop an online fault diagnosis network for lithium-ion batteries operating under unpredictable conditions. The network integrates battery model constraints and employs a framework designed to manage the evolution of stochastic systems, thereby enabling fault real-time determination. We evaluate the performance using a dataset of 18.2 million valid entries from 515 vehicles. The results demonstrate our proposed algorithm outperforms other relevant approaches, enhancing the true positive rate by over 46.5% within a false positive rate range of 0 to 0.2. Meanwhile, we identify the trigger probability for four safety fault samples, namely, electrolyte leakage, thermal runaway, internal short circuit, and excessive aging. The proposed network is adaptable to packs of varying structures, thereby reducing the cost of implementation. Our work explores the application of deep learning for real-state prediction and diagnosis of batteries, demonstrating potential improvements in battery safety and economic benefits.

PMID:39952987 | DOI:10.1038/s41467-025-56832-8

Sci Rep. 2025 Feb 14;15(1):5521. doi: 10.1038/s41598-025-86110-y.

ABSTRACT

While multi-modal deep learning approaches trained using magnetic resonance imaging (MRI) and fluorodeoxyglucose positron emission tomography (FDG PET) data have shown promise in the accurate identification of Alzheimer's disease, their clinical applicability is hindered by the assumption that both modalities are always available during model inference. In practice, clinicians adjust diagnostic tests based on available information and specific clinical contexts. We propose a novel MRI- and FDG PET-based multi-modal deep learning approach that mimics clinical decision-making by incorporating uncertainty estimates of an MRI-based model (generated using Monte Carlo dropout and evidential deep learning) to determine the necessity of an FDG PET scan, and only inputting the FDG PET to a multi-modal model when required. This approach significantly reduces the reliance on FDG PET scans, which are costly and expose patients to radiation. Our approach reduces the need for FDG PET by up to 92% without compromising model performance, thus optimizing resource use and patient safety. Furthermore, using a global model explanation technique, we provide insights into how anatomical changes in brain regions-such as the entorhinal cortex, amygdala, and ventricles-can positively or negatively influence the need for FDG PET scans in alignment with clinical understanding of Alzheimer's disease.

PMID:39952976 | DOI:10.1038/s41598-025-86110-y