Search Results (1,430)

To solve the problem of low segmentation model accuracy due to the complex shape of carbon slag in the aluminum electrolysis fire-eye image and the blurring of the boundary between the slag and the surrounding electrolyte, this paper proposes a segmentation model of the fire-eye image based on an improved U-Net. The model reduces the depth of the traditional U-Net to four layers and uses the multiscale dilated convolution module (MDCM) in the down-sampling stage. Second, the Convolutional Block Attention Module (CBAM) is embedded in the skip connection part of the network to improve the ability of the model to extract contextual features from images of multiple scales, enhance the guidance of high-level features to low-level features, and make the model pay more attention to the critical regions. To alleviate the negative impact of the imbalance of positive and negative examples in the dataset, the weighted binary cross-entropy loss and the Dice loss are used to replace the traditional cross-entropy loss. The experimental results show that the segmentation accuracy of the improved model on the fire-eye dataset reaches 88.03%, which is 5.61 percentage points higher than U-Net. Full article

Circular RNAs (circRNAs) are a class of unique transcripts characterized by a covalently closed loop structure, which differentiates them from conventional linear RNAs. The formation of circRNAs occurs co-transcriptionally and post-transcriptionally through a distinct type of splicing known as back-splicing, which involves the formation of a head-to-tail splice junction between a 5′ splice donor and an upstream 3′ splice acceptor. This process, along with exon skipping, intron retention, cryptic splice site utilization, and lariat-driven intron processing, results in the generation of three main types of circRNAs (exonic, intronic, and exonic–intronic) and their isoforms. The intricate biogenesis of circRNAs is regulated by the interplay of cis-regulatory elements and trans-acting factors, with intronic Alu repeats and RNA-binding proteins playing pivotal roles, at least in the formation of exonic circRNAs. Various hypotheses regarding pathways of circRNA turnover are forwarded, including endonucleolytic cleavage and exonuclease-mediated degradation; however, similarly to the inconclusive nature of circRNA biogenesis, the process of their degradation and the factors involved remain largely unclear. There is a knowledge gap regarding whether these processes are guided by universal pathways or whether each category of circRNAs requires special tools and particular mechanisms for their life cycles. Understanding these factors is pivotal for fully comprehending the biological significance of circRNAs. This review provides an overview of the various pathways involved in the biogenesis and degradation of different types of circRNAs and explores key factors that have beneficial or adverse effects on the formation and stability of these unique transcripts in higher eukaryotes. Full article

Elucidating the gene regulatory mechanisms underlying the gut–brain axis is critical for uncovering novel gut–brain interaction pathways and developing therapeutic strategies for gut bacteria-associated neurological disorders. Most studies have primarily investigated how gut bacteria modulate host epigenetics and gene expression; their impact on host alternative splicing, particularly in the brain, remains largely unexplored. Here, we investigated the effects of the gut-associated probiotic Lacidofil^® on alternative splicing across 10 regions of the rat brain using published RNA-sequencing data. The Lacidofil^® altogether altered 2941 differential splicing events, predominantly, skipped exon (SE) and mutually exclusive exon (MXE) events. Protein–protein interactions and a KEGG analysis of differentially spliced genes (DSGs) revealed consistent enrichment in the spliceosome and vesicle transport complexes, as well as in pathways related to neurodegenerative diseases, synaptic function and plasticity, and substance addiction across brain regions. Using the PsyGeNET platform, we found that DSGs from the locus coeruleus (LConly), medial preoptic area (mPOA), and ventral dentate gyrus (venDG) were enriched in depression-associated or schizophrenia-associated genes. Notably, we highlight the App gene, where Lacidofil^® precisely regulated the splicing of two exons causally involved in amyloid β protein-based neurodegenerative diseases. Although the splicing factors exhibited both splicing plasticity and expression plasticity in response to Lacidofil^®, the overlap between DSGs and differentially expressed genes (DEGs) in most brain regions was rather low. Our study provides novel mechanistic insight into how gut probiotics might influence brain function through the modulation of RNA splicing. Full article

Background: Liver ultrasound segmentation is challenging due to low image quality and variability. While deep learning (DL) models have been widely applied for medical segmentation, generic pre-configured models may not meet the specific requirements for targeted areas in liver ultrasound. Quantitative ultrasound (QUS) is emerging as a promising tool for liver fat measurement; however, accurately segmenting regions of interest within liver ultrasound images remains a challenge. Methods: We introduce a generalizable framework using an adaptive evolutionary genetic algorithm to optimize deep learning models, specifically U-Net, for focused liver segmentation. The algorithm simultaneously adjusts the depth (number of layers) and width (neurons per layer) of the network, dropout, and skip connections. Various architecture configurations are evaluated based on segmentation performance to find the optimal model for liver ultrasound images. Results: The model with a depth of 4 and filter sizes of [16, 64, 128, 256] achieved the highest mean adjusted Dice score of 0.921, outperforming the other configurations, using three-fold cross-validation with early stoppage. Conclusions: Adaptive evolutionary optimization enhances the deep learning architecture for liver ultrasound segmentation. Future work may extend this optimization to other imaging modalities and deep learning architectures. Full article

Duchenne muscular dystrophy (DMD) is a severe X-linked genetic disorder caused by an array of mutations in the dystrophin gene, with the most commonly mutated regions being exons 48–55. One of the several existing approaches to treat DMD is gene therapy, based on alternative splicing and mutant exon skipping. Testing of such therapy requires animal models that carry mutations homologous to those found in human patients. Here, we report the generation of two genetically modified mouse lines, named “insT” and “insG”, with distinct mutations at the same position in exon 51 that lead to a frameshift, presumably causing protein truncation. Hemizygous males of both lines exhibit classical signs of muscular dystrophy in all muscle tissues except for the cardiac tissue. However, pathological changes are more pronounced in one of the lines. Membrane localization of the protein is reduced to the point of absence in one of the lines. Moreover, an increase in full-length isoform mRNA was detected in diaphragms of insG line mice. Although further work is needed to qualify these mutations as sole origins of dissimilarity, both genetically modified mouse lines are suitable models of DMD and can be used to test gene therapy based on alternative splicing. Full article

Background: An estimated 10–15% of all genetic diseases are attributable to variants in noncanonical splice sites, auxiliary splice sites and deep-intronic variants. Most of these unstudied variants are classified as variants of uncertain significance (VUS), which are not clinically actionable. This study investigated two novel splice-altering variants, CHM NM_000390.4:c.941-11T>G and CACNA1F NM_005183.4:c.2576+4_2576+5del implicated in choroideremia and cone dystrophy (COD), respectively, resulting in significant visual loss. Methods: Next-generation sequencing was employed to identify the candidate variants in CHM and CACNA1F, which were confirmed using Sanger sequencing. Cascade analysis was undertaken when additional family members were available. Functional analysis was conducted by cloning genomic regions of interest into gateway expression vectors, creating variant and wildtype midigenes, which were subsequently transfected into HEK293 cells. RNA was harvested and amplified by RT-PCR to investigate the splicing profile for each variant compared to the wildtype. Novel variants were reclassified according to ACMG/AMP and ClinGen SVI guidelines. Results: Midigene functional analysis confirmed that both variants disrupted splicing. The CHM NM_000390.4:c.941-11T>G variant caused exon 8 skipping, leading to a frameshift and the CACNA1F NM_005183.4:c.2576+4_2576+5del variant caused a multimodal splice defect leading to an in-frame insertion of seven amino acids and a frameshift. With this evidence, the former was upgraded to likely pathogenic and the latter to a hot VUS. Conclusions: This study adds to the mutational spectrum of splicing defects implicated in retinal degenerations by identifying and characterising two novel variants in CHM and CACNA1F. Our results highlight the importance of conducting functional analysis to investigate the consequences of intronic splice-altering variants and the significance of reclassifying VUS to confirm a genetic diagnosis. Full article

Alternative splicing is essential for the generation of various protein isoforms that are involved in cell differentiation and tissue development. In addition to internal coding exons, alternative splicing affects the exons with translation initiation codons; however, little is known about these exons. Here, we performed a systematic classification of human alternative exons using coding information. The analysis showed that more than 5% of cassette exons contain translation initiation codons (alternatively skipped exons harboring a 5′ untranslated region and coding region, 5UC-ASEs) although their skipping causes the deletion of translation initiation sites essential for protein synthesis. The splicing of 5UC-ASEs is under the repressive control of MATR3, a DNA/RNA-binding protein associated with neurodegeneration, and is distinctly regulated particularly in the human brain, muscle, and testis. Interestingly, MATR3 represses its own translation by skipping a 5UC-ASE in MATR3 to autoregulate its expression level. 5UC-ASEs are larger than other types of alternative exons. Furthermore, evolutionary analysis revealed that 5UC-ASEs have already appeared in cartilaginous fishes, have increased in amphibians, and are concentrated in the genes involved in transcription in mammals. Taken together, our analysis identified a unique set of alternative exons, 5UC-ASEs, that have evolutionarily acquired a repression mechanism for gene expression through association with MATR3. Full article

In forest ecosystems, the application of hyperspectral (HS) imagery offers unprecedented opportunities for refined identification and classification. The diversity and complexity of forest cover make it challenging for traditional remote-sensing techniques to capture subtle spectral differences. Hyperspectral imagery, however, can reveal the nuanced changes in different tree species, vegetation health status, and soil composition through its nearly continuous spectral information. This detailed spectral information is crucial for the monitoring, management, and conservation of forest resources. While Convolutional Neural Networks (CNNs) have demonstrated excellent local context modeling capabilities in HS image classification, their inherent network architecture limits the exploration and representation of spectral feature sequence properties. To address this issue, we have rethought HS image classification from a sequential perspective and proposed a hybrid model, the Spectral Weaver, which combines CNNs and Transformers. The Spectral Weaver replaces the traditional Multi-Head Attention mechanism with a Channel Attention mechanism (MCA) and introduces Centre-Differential Convolutional Layers (Conv2d-cd) to enhance spatial feature extraction capabilities. Additionally, we designed a cross-layer skip connection that adaptively learns to fuse “soft” residuals, transferring memory-like components from shallow to deep layers. Notably, the proposed model is a highly flexible backbone network, adaptable to both hyperspectral and multispectral image inputs. In comparison to traditional Visual Transformers (ViT), the Spectral Weaver innovates in several ways: (1) It introduces the MCA mechanism to enhance the mining of spectral feature sequence properties; (2) It employs Centre-Differential Convolutional Layers to strengthen spatial feature extraction; (3) It designs cross-layer skip connections to reduce information loss; (4) It supports both multispectral and hyperspectral inputs, increasing the model’s flexibility and applicability. By integrating global and local features, our model significantly improves the performance of HS image classification. We have conducted extensive experiments on the Gaofen dataset, multispectral data, and multiple hyperspectral datasets, validating the superiority of the Spectral Weaver model in forest hyperspectral image classification. The experimental results show that our model achieves 98.59% accuracy on multispectral data, surpassing ViT’s 96.30%. On the Jilin-1 dataset, our proposed algorithm achieved an accuracy of 98.95%, which is 2.17% higher than ViT. The model significantly outperforms classic ViT and other state-of-the-art backbone networks in classification performance. Not only does it effectively capture the spectral features of forest vegetation, but it also significantly improves the accuracy and robustness of classification, providing strong technical support for the refined management and conservation of forest resources. Full article

Ground-based cloud image features high-spatiotemporal resolution, presenting detailed local cloud structures and valuable weather information, which are crucial for meteorological forecasting. However, the inherent fuzziness and dynamism of ground-based clouds have hindered the development of effective prediction algorithms, resulting in low accuracy. This paper presents CloudPredRNN++, a novel method for predicting ground-based cloud dynamics, leveraging a deep spatiotemporal sequence prediction network enhanced with a self-attention mechanism. Initially, a Cascaded Causal LSTM (CCLSTM) with a dual-memory group decoupling structure is designed to enhance the representation of short-term cloud changes. Next, self-attention memory units are incorporated to capture the long-term dependencies and emphasize the non-stationary characteristics of cloud movements. These components are integrated into cloud dynamic feature mining units, which concurrently extract spatiotemporal features to strengthen unified spatiotemporal modeling. Finally, by embedding gradient highway units and adding skip connection, CloudPredRNN++ is constructed into a hierarchical recursive structure, mitigating the gradient vanishing and enhancing the uniform modeling of temporal–spatial features. Experiments on the sequence ground-based cloud dataset demonstrate that CloudPredRNN++ can predict the future cloud state more accurately and quickly. Compared with other spatiotemporal sequence prediction models, CloudPredRNN++ shows significant improvements in evaluation metrics, improving the accuracy of cloud dynamics forecasting and alleviating long-term dependency decay, thus confirming the effectiveness in ground-based cloud prediction tasks. Full article

Background: Integrating nonlinear behavior into the architecture of artificial neural networks is regarded as essential requirement to constitute their effectual learning capacity for solving complex tasks. This claim seems to be true for moderate-sized networks, i.e., with a lower double-digit number of layers. However, going deeper with neural networks regularly turns into destructive tendencies of gradual performance degeneration during training. To circumvent this degradation problem, the prominent neural architectures Residual Network and Highway Network establish skip connections with additive identity mappings between layers. Methods: In this work, we unify the mechanics of both architectures into Capsule Networks (CapsNet)s by showing their inherent ability to learn skip connections. As a necessary precondition, we introduce the concept of Adaptive Nonlinearity Gates (ANG)s which dynamically steer and limit the usage of nonlinear processing. We propose practical methods for the realization of ANGs including biased batch normalization, the Doubly-Parametric ReLU (D-PReLU) activation function, and Gated Routing (GR) dedicated to extremely deep CapsNets. Results: Our comprehensive empirical study using MNIST substantiates the effectiveness of our developed methods and delivers valuable insights for the training of very deep nets of any kind. The final experiments on Fashion-MNIST and SVHN demonstrate the potential of pure capsule-driven networks with GR. Full article

Targeted therapy indications for actionable variants in non-small-cell lung cancer (NSCLC) have primarily been studied in Caucasian populations, with limited data on Latin American patients. This study utilized a 52-genes next-generation sequencing (NGS) panel to analyze 1560 tumor biopsies from NSCLC patients in Chile, Brazil, and Peru. The RNA sequencing reads and DNA coverage were correlated to improve the detection of the actionable MET exon 14 skipping variant (METex14). The pathogenicity of MET variants of uncertain significance (VUSs) was assessed using bioinformatic methods, based on their predicted driver potential. The effects of the predicted drivers VUS T992I and H1094Y on c-MET signaling activation, proliferation, and migration were evaluated in HEK293T, BEAS-2B, and H1993 cell lines. Subsequently, c-Met inhibitors were tested in 2D and 3D cell cultures, and drug affinity was determined using 3D structure simulations. The prevalence of MET variants in the South American cohort was 8%, and RNA-based diagnosis detected 27% more cases of METex14 than DNA-based methods. Notably, 20% of METex14 cases with RNA reads below the detection threshold were confirmed using DNA analysis. The novel actionable T992I and H1094Y variants induced proliferation and migration through c-Met/Akt signaling. Both variants showed sensitivity to crizotinib and savolitinib, but the H1094Y variant exhibited reduced sensitivity to capmatinib. These findings highlight the importance of RNA-based METex14 diagnosis and reveal the drug sensitivity profiles of novel actionable MET variants from an understudied patient population. Full article

Network Intrusion Detection Systems (NIDSs) are vital for safeguarding Internet of Things (IoT) networks from malicious attacks. Modern NIDSs utilize Machine Learning (ML) techniques to combat evolving threats. This study systematically examined adversarial attacks originating from the image domain against ML-based NIDSs, while incorporating a diverse selection of ML models. Specifically, we evaluated both white-box and black-box attacks on nine commonly used ML-based NIDS models. We analyzed the Projected Gradient Descent (PGD) attack, which uses gradient descent on input features, transfer attacks, the score-based Zeroth-Order Optimization (ZOO) attack, and two decision-based attacks: Boundary and HopSkipJump. Using the NSL-KDD dataset, we assessed the accuracy of the ML models under attack and the success rate of the adversarial attacks. Our findings revealed that the black-box decision-based attacks were highly effective against most of the ML models, achieving an attack success rate exceeding 86% across eight models. Additionally, while the Logistic Regression and Multilayer Perceptron models were highly susceptible to all the attacks studied, the instance-based ML models, such as KNN and Label Spreading, exhibited resistance to these attacks. These insights will contribute to the development of more robust NIDSs against adversarial attacks in IoT environments. Full article

We focus on developing a lightweight model for resource-constrained devices, building on MobileViT, a hybrid model that combines the strengths of Transformers and CNNs to balance high accuracy and computational efficiency for image classification. Transformers, while effective at capturing global information, often have higher computational costs than CNNs due to the complexity of their self-attention mechanism. To address this, we introduce the Token Merging (ToMe) technique into MobileViT to reduce computational costs. However, because the number of tokens changes during merging, ToMe cannot be directly applied to MobileViT without adjustments. We propose simple methods, specifically reshaping features and removing skip connections, to resolve this issue. Additionally, we make adjustments to MobileViT’s structure to better support the application of ToMe. Our approach improves inference efficiency while retaining a competitive level of accuracy. The resulting models achieve a balance between performance and computational speed, offering a practical solution for hybrid architectures. This work shows the potential of ToMe-based techniques to broaden the range of lightweight model options, catering to diverse application requirements. Full article

Object tracking is a challenging task in computer vision. While simple tracking methods offer fast speeds, they often fail to track targets. To address this issue, traditional methods typically rely on complex algorithms. This study presents a novel approach to enhance object tracking speed via confidence-guided frame skipping. The proposed method is strategically designed to complement existing methods. Initially, lightweight tracking is used to track a target. Only in scenarios where it fails to track is an existing, robust but complex algorithm used. The contribution of this study lies in the proposed confidence assessment of the lightweight tracking’s results. The proposed method determines the need for intervention by the robust algorithm based on the predicted confidence level. This two-tiered approach significantly enhances tracking speed by leveraging the lightweight method for straightforward situations and the robust algorithm for challenging scenarios. Experimental results demonstrate the effectiveness of the proposed approach in enhancing tracking speed. Full article

Pan-sharpening aims to generate high-resolution multispectral (HRMS) images by combining high-resolution panchromatic (PAN) images with low-resolution multispectral (LRMS) data, while maintaining the symmetry of spatial and spectral characteristics. Traditional convolutional neural networks (CNNs) struggle with global dependency modeling due to local receptive fields, and Transformer-based models are computationally expensive. Recent Mamba models offer linear complexity and effective global modeling. However, existing Mamba-based methods lack sensitivity to local feature variations, leading to suboptimal fine-detail preservation. To address this, we propose a Conditional Skipping Mamba Network (CSMN), which enhances global-local feature fusion symmetrically through two modules: (1) the Adaptive Mamba Module (AMM), which improves global perception using adaptive spatial-frequency integration; and (2) the Cross-domain Mamba Module (CDMM), optimizing cross-domain spectral-spatial representation. Experimental results on the IKONOS and WorldView-2 datasets demonstrate that CSMN surpasses existing state-of-the-art methods in achieving superior spectral consistency and preserving spatial details, with performance that is more symmetric in fine-detail preservation. Full article