J Transl Med. 2025 Jan 14;23(1):65. doi: 10.1186/s12967-024-06044-3.

ABSTRACT

BACKGROUND: Parkinson's Disease (PD) is a neurodegenerative disorder, and eye movement abnormalities are a significant symptom of its diagnosis. In this paper, we developed a multi-task driven by eye movement in a virtual reality (VR) environment to elicit PD-specific eye movement abnormalities. The abnormal features were subsequently modeled by using the proposed deep learning algorithm to achieve an auxiliary diagnosis of PD.

METHODS: We recruited 114 PD patients and 125 healthy controls and collected their eye-tracking data in a VR environment. Participants completed a series of specific VR tasks, including gaze stability, pro-saccades, anti-saccades, and smooth pursuit. After the tasks, eye movement features were extracted from the behaviors of fixations, saccades, and smooth pursuit to establish a PD diagnostic model.

RESULTS: The performance of the models was evaluated through cross-validation, revealing a recall of 97.65%, an accuracy of 92.73%, and a receiver operator characteristic area under the curve (ROC-AUC) of 97.08% for the proposed model.

CONCLUSION: We extracted PD-specific eye movement features from the behaviors of fixations, saccades, and smooth pursuit in a VR environment to create a model with high accuracy and recall for PD diagnosis. Our method provides physicians with a new auxiliary tool to improve the prognosis and quality of life of PD patients.

PMID:39810187 | DOI:10.1186/s12967-024-06044-3

BMC Med Educ. 2025 Jan 14;25(1):66. doi: 10.1186/s12909-025-06648-3.

ABSTRACT

BACKGROUND: Reflection fosters self-regulated learning by enabling learners to critically evaluate their performance, identify gaps, and make plans to improve. Feedback, in turn, provides external insights that complement reflection, helping learners recognize their strengths and weaknesses, adjust their learning strategies, and enhance clinical reasoning and decision-making skills. However, reflection alone may not produce the desirable effects unless coupled with feedback. This study aimed to investigate the impact of feedback integrated reflection on meaningful learning and higher order MCQ score among under-grade medical students.

OBJECTIVE: To evaluate the impact of feedback-integrated reflection versus reflection alone on higher-order MCQ scores among undergraduate medical students in a gynecology clinical setting.

METHODS: A randomized controlled trial was conducted with 68 final-year medical students randomly assigned to a study group (feedback-integrated reflection) and a control group (reflection alone). Both groups completed a pre-test, followed by six daily teaching sessions on gynecology topics. Participants engaged in written reflections after each session, and the study group additionally received individualized feedback. Independent sample t-tests were used to compare pre and post-test scores between the groups, while paired t-tests assessed within-group improvements.

RESULTS: Pre-test scores were comparable between the study group (11.68 ± 2.60, 38.93%) and the control group (11.29 ± 2.38, 37.15%; P = 0.52). Post-test scores showed a significant improvement in the study group (20.88 ± 2.98, 69.32%) compared to the control group (15.29 ± 2.66, 51.00%; P = 0.0001). The percentage gain in learning was 35.43% for the control group (reflection alone) and 78.77% for the study group (feedback-integrated reflection). The normalized learning gain (NLG) was calculated to compare the effectiveness of the intervention (feedback-integrated reflection) with that of the control (reflection alone). The study group demonstrated a mean normalized learning gain of 69.07%, compared to 29.18% in the control group. The net learning gain, calculated as the difference in normalized learning gains between the study and control groups, was found to be 39.89%.

CONCLUSION: The findings highlight the effectiveness of feedback-integrated reflection versus reflection alone in fostering deeper learning by improving higher-order MCQ scores in a gynecologic setting in the undergraduate medical education.

TRIAL REGISTRATION: This trial was registered retrospectively on 27th July 2024. Trial registration no is CTU/07/2024/010/RMU.

PMID:39810114 | DOI:10.1186/s12909-025-06648-3

BMC Genom Data. 2025 Jan 14;26(1):4. doi: 10.1186/s12863-024-01293-z.

ABSTRACT

BACKGROUND: miRNAs (microRNAs) are endogenous RNAs with lengths of 18 to 24 nucleotides and play critical roles in gene regulation and disease progression. Although traditional wet-lab experiments provide direct evidence for miRNA-disease associations, they are often time-consuming and complicated to analyze by current bioinformatics tools. In recent years, machine learning (ML) and deep learning (DL) techniques are powerful tools to analyze large-scale biological data. Hence, developing a model to predict, identify, and rank connections in miRNAs and diseases can significantly enhance the precision and efficiency in investigating the relationships between miRNAs and diseases.

RESULTS: In this study, we utilized miRNA-disease association data obtained by biotechnological experiments to develop a DL model for miRNA-disease associations. To improve the accuracy of prediction in this model, we introduced two labeling strategies, weight-based and majority-based definitions, to classify miRNA-disease associations. After preprocessing, data was trained with a novel model combining gated recurrent units (GRU) and graph convolutional network (GCN) to predict the level of miRNA-disease associations. The miRNA-disease association datasets were from HMDD (the Human miRNA Disease Database) and categorized by two distinct labeling approaches, weight-based definitions and majority-based definitions. We classified the miRNA-disease associations into three groups, "upregulated", "downregulated" and "nonspecific", by regression analysis and multiclass classification. This GRU-GCN coordinated model achieved a robust area under the curve (AUC) score of 0.8 in all datasets, demonstrating the efficacy in predicting potential miRNA-disease relationships.

CONCLUSIONS: By introducing innovative label-preprocessing methods, this study addressed the relationships between miRNAs and diseases, and improved the ambiguity of the results in different experiments. Based on these refined label definitions, we developed a DL-based model to refine and predict the results of associations between miRNAs and diseases. This model offers a valuable tool for complementing traditional experimental methods and enhancing our understanding of miRNA-related disease mechanisms.

PMID:39810100 | DOI:10.1186/s12863-024-01293-z

J Imaging Inform Med. 2025 Jan 14. doi: 10.1007/s10278-024-01375-5. Online ahead of print.

ABSTRACT

Optical coherence tomography angiography (OCTA) is an emerging, non-invasive technique increasingly utilized for retinal vasculature imaging. Analysis of OCTA images can effectively diagnose retinal diseases, unfortunately, complex vascular structures within OCTA images possess significant challenges for automated segmentation. A novel, fully convolutional dense connected residual network is proposed to effectively segment the vascular regions within OCTA images. Firstly, a dual-branch structure Recurrent Residual Convolutional Neural Network (RRCNN) block is constructed utilizing RecurrentBlock and convolutional operations. Subsequently, the ResConvNeXt V2 Block is built as the backbone structure of the network. The output from the ResConvNeXt V2 Block is then fed into the side branch and the next ResConvNeXt V2 Block. Within the side branch, the Group Receptive Field Block (GRFB) processes the results from the previous and current layers. Ultimately, the side branch results are added to the backbone network outputs to produce the final segmentation. The model achieves superior performance. Experiments were conducted on the ROSSA and OCTA-500 datasets, yielding Dice scores of 91.88%, 91.72%, and 89.18% for the respective datasets, and accuracies of 98.31%, 99.02%, and 98.02%.

PMID:39810043 | DOI:10.1007/s10278-024-01375-5

Psychon Bull Rev. 2025 Jan 14. doi: 10.3758/s13423-024-02616-y. Online ahead of print.

ABSTRACT

We examined the intricate mechanisms underlying visual processing of complex motion stimuli by measuring the detection sensitivity to contraction and expansion patterns and the discrimination sensitivity to the location of the center of motion (CoM) in various real and unreal optic flow stimuli. We conducted two experiments (N = 20 each) and compared responses to both "real" optic flow stimuli containing information about self-movement in a three-dimensional scene and "unreal" optic flow stimuli lacking such information. We found that detection sensitivity to contraction surpassed that to expansion patterns for unreal optic flow stimuli, whereas this trend was reversed for real optic flow stimuli. Furthermore, while discrimination sensitivity to the CoM location was not affected by stimulus duration for unreal optic flow stimuli, it showed a significant improvement when stimulus duration increased from 100 to 400 ms for real optic flow stimuli. These findings provide compelling evidence that the visual system employs distinct processing approaches for real versus unreal optic flow even when they are perfectly matched for two-dimensional global features and local motion signals. These differences reveal influences of self-movement in natural environments, enabling the visual system to uniquely process stimuli with significant survival implications.

PMID:39810018 | DOI:10.3758/s13423-024-02616-y

Sci Rep. 2025 Jan 14;15(1):1962. doi: 10.1038/s41598-025-86455-4.

ABSTRACT

The NLRP3 inflammasome, regulated by TLR4, plays a pivotal role in periodontitis by mediating inflammatory cytokine release and bone loss induced by Porphyromonas gingivalis. Periodontal disease creates a hypoxic environment, favoring anaerobic bacteria survival and exacerbating inflammation. The NLRP3 inflammasome triggers pyroptosis, a programmed cell death that amplifies inflammation and tissue damage. This study evaluates the efficacy of Variational Graph Autoencoders (VGAEs) in reconstructing gene data related to NLRP3-mediated pyroptosis in periodontitis. The NCBI GEO dataset GSE262663, containing three samples with and without hypoxia exposure, was analyzed using unsupervised K-means clustering. This method identifies natural groupings within biological data without prior labels. VGAE, a deep learning model, captures complex graph relationships for tasks like link prediction and edge detection. The VGAE model demonstrated exceptional performance with an accuracy of 99.42% and perfect precision. While it identified 5,820 false negatives, indicating a conservative approach, it accurately predicted 4,080 out of 9,900 positive samples. The model's latent space distribution differed significantly from the original data, suggesting a tightly clustered representation of the gene expression patterns. K-means clustering and VGAE show promise in gene expression analysis and graph structure reconstruction for periodontitis research.

PMID:39809940 | DOI:10.1038/s41598-025-86455-4

Nat Biotechnol. 2025 Jan 14. doi: 10.1038/s41587-024-02528-1. Online ahead of print.

ABSTRACT

Efficient and accurate nanocarrier development for targeted drug delivery is hindered by a lack of methods to analyze its cell-level biodistribution across whole organisms. Here we present Single Cell Precision Nanocarrier Identification (SCP-Nano), an integrated experimental and deep learning pipeline to comprehensively quantify the targeting of nanocarriers throughout the whole mouse body at single-cell resolution. SCP-Nano reveals the tissue distribution patterns of lipid nanoparticles (LNPs) after different injection routes at doses as low as 0.0005 mg kg-1-far below the detection limits of conventional whole body imaging techniques. We demonstrate that intramuscularly injected LNPs carrying SARS-CoV-2 spike mRNA reach heart tissue, leading to proteome changes, suggesting immune activation and blood vessel damage. SCP-Nano generalizes to various types of nanocarriers, including liposomes, polyplexes, DNA origami and adeno-associated viruses (AAVs), revealing that an AAV2 variant transduces adipocytes throughout the body. SCP-Nano enables comprehensive three-dimensional mapping of nanocarrier distribution throughout mouse bodies with high sensitivity and should accelerate the development of precise and safe nanocarrier-based therapeutics.

PMID:39809933 | DOI:10.1038/s41587-024-02528-1

Sci Rep. 2025 Jan 14;15(1):1924. doi: 10.1038/s41598-024-84869-0.

ABSTRACT

To address the challenges of unbalanced class labels with varying maturity levels of tomato fruits and low recognition accuracy for both fruits and stems in intelligent harvesting, we propose the YOLOX-SE-GIoU model for identifying tomato fruit maturity and stems. The SE focus module was incorporated into YOLOX to improve the identification accuracy, addressing the imbalance in the number of tomato fruits and stems. Additionally, we optimized the loss function to GIoU loss to minimize discrepancies across different scales of fruits and stems. The mean average precision (mAP) of the improved YOLOX-SE-GIoU model reaches 92.17%. Compared to YOLOv4, YOLOv5, YOLOv7, and YOLOX models, the improved model shows an improvement of 1.17-22.21%. The average precision (AP) for unbalanced semi-ripe tomatoes increased by 1.68-26.66%, while the AP for stems increased by 3.78-45.03%. Experimental results demonstrate that the YOLOX-SE-GIoU model exhibits superior overall recognition performance for unbalanced and scale-variant samples compared to the original model and other models in the same series. It effectively reduces false and missed detections during tomato harvesting, improving the identification accuracy of tomato fruits and stems. The findings of this work provide a technical foundation for developing advanced fruit harvesting techniques.

PMID:39809915 | DOI:10.1038/s41598-024-84869-0

Sci Rep. 2025 Jan 14;15(1):1926. doi: 10.1038/s41598-024-81244-x.

ABSTRACT

The rapid expansion of the coal mining industry has introduced significant safety risks, particularly within the harsh environments of open-pit coal mines. The safe and stable operation of belt conveyor idlers is crucial not only for ensuring efficient coal production but also for safeguarding the lives of coal mine workers. Therefore, this paper proposes a method based on deep learning for real-time detection of conveyor idler faults. The selected YOLOv5 network is analyzed and improved based on the training results. First, the coordinate attention mechanism is integrated into the model to reassign the weights across different channels. Subsequently, the α-CIoU localization loss function replaces the traditional CIoU to enhance the model's regression accuracy. Experimental results demonstrate that the enhanced YOLOv5 algorithm achieves a 95.3% mAP on the self-constructed infrared image dataset, surpassing the original algorithm by 2.7%. Moreover, with a processing speed of 285 FPS, it accurately performs the defect detection of conveyor idlers while satisfying real-time operational requirements.

PMID:39809903 | DOI:10.1038/s41598-024-81244-x

Commun Biol. 2025 Jan 14;8(1):48. doi: 10.1038/s42003-024-07334-8.

ABSTRACT

Idiopathic pulmonary fibrosis (IPF) is an irreversible lung condition that progresses over time, which ultimately results in respiratory failure and mortality. In this study, we found that PLAC8 was downregulated in the lungs of IPF patients based on GEO data, in bleomycin (BLM)-induced lungs of mice, and in primary murine alveolar epithelial type II (pmATII) cells and human lung epithelial cell A549 cells. Overexpression of PLAC8 facilitated autophagy and inhibited apoptosis of pmATII cells and A549 cells in vitro. Moreover, inhibition of autophagy or overexpression of p53 partially abolished the effects of PLAC8 on cell apoptosis. ATII cell-specific overexpression of PLAC8 alleviated BLM-induced pulmonary fibrosis in mice. Mechanistically, PLAC8 interacts with VCP-UFD1-NPLOC4 complex to promote p53 degradation and facilitate autophagy, resulting in inhibiting apoptosis of alveolar epithelial cells and attenuating pulmonary fibrosis. In summary, these findings indicate that PLAC8 may be a key target for therapeutic interventions in pulmonary fibrosis.

PMID:39810019 | DOI:10.1038/s42003-024-07334-8

Genome Biol. 2025 Jan 14;26(1):9. doi: 10.1186/s13059-024-03471-9.

ABSTRACT

BACKGROUND: Streptomyces is a highly diverse genus known for the production of secondary or specialized metabolites with a wide range of applications in the medical and agricultural industries. Several thousand complete or nearly complete Streptomyces genome sequences are now available, affording the opportunity to deeply investigate the biosynthetic potential within these organisms and to advance natural product discovery initiatives.

RESULTS: We perform pangenome analysis on 2371 Streptomyces genomes, including approximately 1200 complete assemblies. Employing a data-driven approach based on genome similarities, the Streptomyces genus was classified into 7 primary and 42 secondary Mash-clusters, forming the basis for comprehensive pangenome mining. A refined workflow for grouping biosynthetic gene clusters (BGCs) redefines their diversity across different Mash-clusters. This workflow also reassigns 2729 known BGC families to only 440 families, a reduction caused by inaccuracies in BGC boundary detections. When the genomic location of BGCs is included in the analysis, a conserved genomic structure, or synteny, among BGCs becomes apparent within species and Mash-clusters. This synteny suggests that vertical inheritance is a major factor in the diversification of BGCs.

CONCLUSIONS: Our analysis of a genomic dataset at a scale of thousands of genomes refines predictions of BGC diversity using Mash-clusters as a basis for pangenome analysis. The observed conservation in the order of BGCs' genomic locations shows that the BGCs are vertically inherited. The presented workflow and the in-depth analysis pave the way for large-scale pangenome investigations and enhance our understanding of the biosynthetic potential of the Streptomyces genus.

PMID:39810189 | DOI:10.1186/s13059-024-03471-9

NPJ Syst Biol Appl. 2025 Jan 14;11(1):8. doi: 10.1038/s41540-024-00480-z.

ABSTRACT

Genome-scale metabolic models (GSMM) are commonly used to identify gene deletion sets that result in growth coupling and pairing product formation with substrate utilization and can improve strain performance beyond levels typically accessible using traditional strain engineering approaches. However, sustainable feedstocks pose a challenge due to incomplete high-resolution metabolic data for non-canonical carbon sources required to curate GSMM and identify implementable designs. Here we address a four-gene deletion design in the Pseudomonas putida KT2440 strain for the lignin-derived non-sugar carbon source, p-coumarate (p-CA), that proved challenging to implement. We examine the performance of the fully implemented design for p-coumarate to glutamine, a useful biomanufacturing intermediate. In this study glutamine is then converted to indigoidine, an alternative sustainable pigment and a model heterologous product that is commonly used to colorimetrically quantify glutamine concentration. Through proteomics, promoter-variation, and growth characterization of a fully implemented gene deletion design, we provide evidence that aromatic catabolism in the completed design is rate-limited by fumarase hydratase (FUM) enzyme activity in the citrate cycle and requires careful optimization of another fumarate hydratase protein (PP_0897) expression to achieve growth and production. A double sensitivity analysis also confirmed a strict requirement for fumarate hydratase activity in the strain where all genes in the growth coupling design have been implemented. Metabolic cross-feeding experiments were used to examine the impact of complete removal of the fumarase hydratase reaction and revealed an unanticipated nutrient requirement, suggesting additional functions for this enzyme. While a complete implementation of the design was achieved, this study highlights the challenge of completely inactivating metabolic reactions encoded by under-characterized proteins, especially in the context of multi-gene edits.

PMID:39809795 | DOI:10.1038/s41540-024-00480-z

Genome Res. 2025 Jan 14. doi: 10.1101/gr.279372.124. Online ahead of print.

ABSTRACT

Following amputation, zebrafish regenerate their injured caudal fin through lineage-restricted reprogramming. Although previous studies have charted various genetic and epigenetic dimensions of this process, the intricate gene regulatory programs shared by, or unique to, different regenerating cell types remain underinvestigated. Here, we mapped the regulatory landscape of fin regeneration by applying paired snRNA-seq and snATAC-seq on uninjured and regenerating fins. This map delineates the regulatory dynamics of predominant cell populations at multiple stages of regeneration. We observe a marked increase in the accessibility of chromatin regions associated with regenerative and developmental processes at 1 dpa, followed by a gradual closure across major cell types at later stages. This pattern is distinct from that of transcriptomic dynamics, which is characterized by several waves of gene upregulation and downregulation. We identified and in vivo validated cell-type-specific and position-specific regeneration-responsive enhancers and constructed regulatory networks by cell type and stage. Our single-cell resolution transcriptomic and chromatin accessibility map across regenerative stages provides new insights into regeneration regulatory mechanisms and serves as a valuable resource for the community.

PMID:39809530 | DOI:10.1101/gr.279372.124

Bioresour Technol. 2025 Jan 12:132064. doi: 10.1016/j.biortech.2025.132064. Online ahead of print.

ABSTRACT

Pinene is a plant volatile monoterpenoid which is used in the fragrance, pesticide, and biofuel industries. Although α-pinene has been synthesized in microbial cell factories, the low synthesis efficiency has thus far limited its production. In this study, the cell growth and α-pinene production of the engineered yeast were decoupled by a dynamic regulation strategy, resulting in a 101.1-fold increase in α-pinene production compared to the control. By enhancing the mevalonate pathway and expanding the cytosolic acetyl-CoA pool, α-pinene production was further increased. Overexpression of the transporter Sge1 resulted in a redistribution of global gene transcription, leading to an increased flux of α-pinene synthesis. By optimizing the aeration flow rate in 3-L bioreactors, the α-pinene production reached 1.8 g/L, which is the highest reported α-pinene production in cell factories. Our research provides insights and fundamentals for the efficient synthesis of monoterpenoids in microbial cell factories.

PMID:39809385 | DOI:10.1016/j.biortech.2025.132064

Dev Cell. 2025 Jan 9:S1534-5807(24)00771-8. doi: 10.1016/j.devcel.2024.12.032. Online ahead of print.

ABSTRACT

Reactivation of cell division is crucial for the regeneration of damaged tissues, which is a fundamental process across all multicellular organisms. However, the mechanisms underlying the activation of cell division in plants during regeneration remain poorly understood. Here, we show that single-cell endodermal ablation generates a transient change in the local mechanical pressure on neighboring pericycle cells to activate patterned cell division that is crucial for tissue regeneration in Arabidopsis roots. Moreover, we provide strong evidence that this process relies on the phytohormone ethylene. Thus, our results highlight a previously unrecognized role of mechano-chemical control in patterned cell division during regeneration in plants.

PMID:39809282 | DOI:10.1016/j.devcel.2024.12.032

Biomed Pharmacother. 2025 Jan 13;183:117829. doi: 10.1016/j.biopha.2025.117829. Online ahead of print.

ABSTRACT

Soluble guanylate cyclase (sGC) stands as a pivotal regulatory element in intracellular signalling pathways, mediating the formation of cyclic guanosine monophosphate (cGMP) and impacting diverse physiological processes across tissues. Increased formation of reactive oxygen species (ROS) is widely recognized to modulate cGMP signalling. Indeed, oxidatively damaged, and therefore inactive sGC, contributes to poor vascular reactivity and more severe neurological damage upon stroke. However, the specific involvement of cGMP in redox signalling remains elusive. Here, we demonstrate a significant cGMP-dependent reduction of reactive oxygen and nitrogen species upon sGC activation under hypoxic conditions, independent of any potential scavenger effects. Importantly, this reduction is directly mediated by downregulating NADPH oxidase (NOX) 4 and 5 during reperfusion. Using an in silico simulation approach, we propose a mechanistic link between increased cGMP signalling and reduced ROS formation, pinpointing NF-κB1 and RELA/p65 as key transcription factors regulating NOX4/5 expression. In vitro studies revealed that p65 translocation to the nucleus was reduced in hypoxic human microvascular endothelial cells following sGC activation. Altogether, these findings unveil the intricate regulation and functional implications of sGC, providing valuable insights into its biological significance and ultimately therapeutic potential.

PMID:39809128 | DOI:10.1016/j.biopha.2025.117829

Comput Biol Med. 2025 Jan 13;186:109665. doi: 10.1016/j.compbiomed.2025.109665. Online ahead of print.

ABSTRACT

BACKGROUND: Effective drugs for tendinopathy are lacking, resulting in significant morbidity and re-tearing rate after operation. Applying systems biology to identify new applications for current pharmaceuticals can decrease the duration, expenses, and likelihood of failure associated with the development of new drugs.

METHODS: We identify tendinopathy signature genes employing a transcriptomics database encompassing 154 clinical tendon samples. We then proposed a systems biology based drug prediction strategy that encompassed multiplex transcriptional drug prediction, systematic review assessment, deep learning based efficacy prediction and Mendelian randomization (MR). Finally, we evaluated the effects of drug target using gene knockout mice.

RESULTS: We demonstrate that sirolimus is a repurposable drug for tendinopathy, supported by: 1) Sirolimus achieves top ranking in drug-gene signature-based multiplex transcriptional drug efficacy prediction, 2) Consistent evidence from systematic review substantiates the efficacy of sirolimus in the management of tendinopathy, 3) Genetic prediction indicates that plasma proteins inhibited by mTOR (the target of sirolimus) are associated with increased tendinopathy risk. The effectiveness of sirolimus is further corroborated through in vivo testing utilizing tendon tissue-specific mTOR gene knockout mice. Integrative pathway enrichment analysis suggests that mTOR inhibition can regulate heterotopic ossification-related pathways to ameliorate clinical tendinopathy.

CONCLUSIONS: Our study assimilates knowledge of system-level responses to identify potential drugs for tendinopathy, and suggests sirolimus as a viable candidate. A systems biology approach could expedite the repurposing of drugs for human diseases that do not have well-defined targets.

PMID:39809087 | DOI:10.1016/j.compbiomed.2025.109665

PLoS One. 2025 Jan 14;20(1):e0316548. doi: 10.1371/journal.pone.0316548. eCollection 2025.

ABSTRACT

This study introduces a high-resolution wind nowcasting model designed for aviation applications at Madeira International Airport, a location known for its complex wind patterns. By using data from a network of six meteorological stations and deep learning techniques, the produced model is capable of predicting wind speed and direction up to 30-minute ahead with 1-minute temporal resolution. The optimized architecture demonstrated robust predictive performance across all forecast horizons. For the most challenging task, the 30-minute ahead forecasts, the model achieved a wind speed Mean Absolute Error (MAE) of 0.78 m/s and a wind direction MAE of 33.06°. Furthermore, the use of Gaussian noise concatenation to both input and label training data yielded the most consistent results. A case study further validated the model's efficacy, with MAE values below 0.43 m/s for wind speed and between 33.93° and 35.03° for wind direction across different forecast horizons. This approach shows that combining strategically deployed sensor networks with machine learning techniques offers improvements in wind nowcasting for airports in complex environments, possibly enhancing operational efficiency and safety.

PMID:39808682 | DOI:10.1371/journal.pone.0316548

PLoS One. 2025 Jan 14;20(1):e0317585. doi: 10.1371/journal.pone.0317585. eCollection 2025.

ABSTRACT

Urban waterfront areas, which are essential natural resources and highly perceived public areas in cities, play a crucial role in enhancing urban environment. This study integrates deep learning with human perception data sourced from street view images to study the relationship between visual landscape features and human perception of urban waterfront areas, employing linear regression and random forest models to predict human perception along urban coastal roads. Based on aesthetic and distinctiveness perception, urban coastal roads in Xiamen were classified into four types with different emphasis and priorities for improvement. The results showed that: 1) the degree of coastal openness had the greatest influence on human perception while the coastal landscape with a high green visual index decreases the distinctiveness perception; 2) the random forest model can effectively predict human perception on urban coastal roads with an accuracy rate of 87% and 77%; 3) The proportion of low perception road sections with potential for improvement is 60.6%, among which the proportion of low aesthetic perception and low distinctiveness perception road sections is 10.5%. These findings offer crucial evidence regarding human perception of urban coastal roads, and can provide targeted recommendations for enhancing the visual environment of urban coastal road landscapes.

PMID:39808675 | DOI:10.1371/journal.pone.0317585

J Thorac Imaging. 2024 Sep 20. doi: 10.1097/RTI.0000000000000808. Online ahead of print.

ABSTRACT

PURPOSE: Radiological follow-up of oncology patients requires the detection of metastatic lung lesions and the quantitative analysis of their changes in longitudinal imaging studies. Our aim was to evaluate SimU-Net, a novel deep learning method for the automatic analysis of metastatic lung lesions and their temporal changes in pairs of chest CT scans.

MATERIALS AND METHODS: SimU-Net is a simultaneous multichannel 3D U-Net model trained on pairs of registered prior and current scans of a patient. It is part of a fully automatic pipeline for the detection, segmentation, matching, and classification of metastatic lung lesions in longitudinal chest CT scans. A data set of 5040 metastatic lung lesions in 344 pairs of 208 prior and current chest CT scans from 79 patients was used for training/validation (173 scans, 65 patients) and testing (35 scans, 14 patients) of a standalone 3D U-Net models and 3 simultaneous SimU-Net models. Outcome measures were the lesion detection and segmentation precision, recall, Dice score, average symmetric surface distance (ASSD), lesion matching, and classification of lesion changes from computed versus manual ground-truth annotations by an expert radiologist.

RESULTS: SimU-Net achieved a mean lesion detection recall and precision of 0.93±0.13 and 0.79±0.24 and a mean lesion segmentation Dice and ASSD of 0.84±0.09 and 0.33±0.22 mm. These results outperformed the standalone 3D U-Net model by 9.4% in the recall, 2.4% in Dice, and 15.4% in ASSD, with a minor 3.6% decrease in precision. The SimU-Net pipeline achieved perfect precision and recall (1.0±0.0) for lesion matching and classification of lesion changes.

CONCLUSIONS: Simultaneous deep learning analysis of metastatic lung lesions in prior and current chest CT scans with SimU-Net yields superior accuracy compared with individual analysis of each scan. Implementation of SimU-Net in the radiological workflow may enhance efficiency by automatically computing key metrics used to evaluate metastatic lung lesions and their temporal changes.

PMID:39808543 | DOI:10.1097/RTI.0000000000000808