Explore the vast range of titles available.

Single-cell RNA sequencing provides high-throughput gene expression information to explore cellular heterogeneity at the individual cell level. A major challenge in characterizing high-throughput gene expression data arises from challenges related to dimensionality, and the prevalence of dropout events. To address these concerns, we develop a deep graph learning method, scMGCA, for single-cell data analysis. scMGCA is based on a graph-embedding autoencoder that simultaneously learns cell-cell topology representation and cluster assignments. We show that scMGCA is accurate and effective for cell segregation and batch effect correction, outperforming other state-of-the-art models across multiple platforms. In addition, we perform genomic interpretation on the key compressed transcriptomic space of the graph-embedding autoencoder to demonstrate the underlying gene regulation mechanism. We demonstrate that in a pancreatic ductal adenocarcinoma dataset, scMGCA successfully provides annotations on the specific cell types and reveals differential gene expression levels across multiple tumor-associated and cell signalling pathways. A major challenge in analyzing scRNA-seq data arises from challenges related to dimensionality and the prevalence of dropout events. Here the authors develop a deep graph learning method called scMGCA based on a graph-embedding autoencoder that simultaneously learns cell-cell topology representation and cluster assignments, outperforming other state-of-the-art models across multiple platforms.

Single-cell ATAC-seq technology advances our understanding of single-cell heterogeneity in gene regulation by enabling exploration of epigenetic landscapes and regulatory elements. However, low sequencing depth per cell leads to data sparsity and high dimensionality, limiting the characterization of gene regulatory elements. Here, we develop scAGDE, a single-cell chromatin accessibility model-based deep graph representation learning method that simultaneously learns representation and clustering through explicit modeling of data generation. Our evaluations demonstrated that scAGDE outperforms existing methods in cell segregation, key marker identification, and visualization across diverse datasets while mitigating dropout events and unveiling hidden chromatin-accessible regions. We find that scAGDE preferentially identifies enhancer-like regions and elucidates complex regulatory landscapes, pinpointing putative enhancers regulating the constitutive expression of CTLA4 and the transcriptional dynamics of CD8A in immune cells. When applied to human brain tissue, scAGDE successfully annotated cis-regulatory element-specified cell types and revealed functional diversity and regulatory mechanisms of glutamatergic neurons. Single-cell ATAC-seq reveals gene regulation at individual cell levels but struggles with data sparsity. Here, authors introduce scAGDE, a deep graph learning framework that improves cell embedding and clustering, outperforming existing methods and uncovering key regulatory mechanisms.

Recent advances in spatial transcriptomics have enabled simultaneous preservation of high‐throughput gene expression profiles and the spatial context, enabling high‐resolution exploration of distinct regional characterization in tissue. To effectively understand the underlying biological mechanisms within tissue microenvironments, there is a requisite for methods that can accurately capture external spatial heterogeneity and interpret internal gene regulation from spatial transcriptomics data. However, current methods for region identification often lack the simultaneous characterizing of spatial structure and gene regulation, thereby limiting the ability of spatial dissection and gene interpretation. Here, stDCL is developed, a dual graph contrastive learning method to identify spatial domains and interpret gene regulation in spatial transcriptomics data. stDCL adaptively incorporates gene expression data and spatial information via a graph embedding autoencoder, thereby preserving critical information within the latent embedding representations. In addition, dual graph contrastive learning is proposed to train the model, ensuring that the latent embedding representation closely resembles the actual spatial distribution and exhibits cluster similarity. Benchmarking stDCL against other state‐of‐the‐art clustering methods using complex cortex datasets demonstrates its superior accuracy and effectiveness in identifying spatial domains. Our analysis of the imputation matrices generated by stDCL reveals its capability to reconstruct spatial hierarchical structures and refine differential expression assessment. Furthermore, it is demonstrated that the versatility of stDCL in interpretability of gene regulation, spatial heterogeneity at high resolution, and embryonic developmental patterns. In addition, it is also showed that stDCL can successfully annotate disease‐associated astrocyte subtypes in Alzheimer's disease and unravel multiple relevant pathways and regulatory mechanisms. stDCL, a dual graph contrastive learning method, captures spatial heterogeneity and interprets gene regulation in spatial transcriptomics data. Integrating spatial and gene expression data through graph embeddings, stDCL provides robust spatial characterization and accurate region identification, reconstructs spatial hierarchies, and identifies disease‐associated cell subtypes, unveiling new insights into tissue microenvironments and disease mechanisms.

Single‐cell RNA sequencing (scRNA‐seq) is a robust method for studying gene expression at the single‐cell level, but accurately quantifying genetic material is often hindered by limited mRNA capture, resulting in many missing expression values. Existing imputation methods rely on strict data assumptions, limiting their broader application, and lack reliable supervision, leading to biased signal recovery. To address these challenges, authors developed Bis, a distribution‐agnostic deep learning model for accurately recovering missing sing‐cell gene expression from multiple platforms. Bis is an optimal transport‐based autoencoder model that can capture the intricate distribution of scRNA‐seq data while addressing the characteristic sparsity by regularizing the cellular embedding space. Additionally, they propose a module using bulk RNA‐seq data to guide reconstruction and ensure expression consistency. Experimental results show Bis outperforms other models across simulated and real datasets, showcasing superiority in various downstream analyses including batch effect removal, clustering, differential expression analysis, and trajectory inference. Moreover, Bis successfully restores gene expression levels in rare cell subsets in a tumor‐matched peripheral blood dataset, revealing developmental characteristics of cytokine‐induced natural killer cells within a head and neck squamous cell carcinoma microenvironment. The accurate measurement of genetic material encounters challenges due to limited intracellular mRNA capture, leading to many missing expression values. A distribution‐agnostic deep learning model, informed by external cues from bulk RNA‐seq data, is developed to address this issue. This model precisely reconstructs gene expression patterns, offering valuable insights into the developmental maturation mechanisms of cytokine‐induced NK cells.

It is well-known that social phenomena can be reflected in art. In the past research time, there have been some research results about the Dunhuang murals and gender equality, respectively. However, there is no such one as combining the two meaningful topics together and applying them in modern society. The aim of this study is to analyze the trend of murals painted in Dunhuang by comparing the characteristics in a different time, from the Beiliang Dynasty to the Yuan Dynasty, and then compare the trend of feminization of the mural images with the changes in women's status during this 1000 years. After comparing and analyzing, the different, even contrary, results lead to deeper thoughts: from what perspective the women's status should be analyzed and described? Definitely, the four dimensions, including economic participation and opportunity, educational attainment, health and survival, and political empowerment, should not be ignored and must be treated seriously. That is meaningful for the research of Dunhuang Grottoes, the images painted in Dunhuang, the studies of ancient China, today’s gender inequality, and many other areas. It is confident that this article will fill academic gaps and provide a new idea to other researchers.

Pulmonary hypertension (PH) is an uncommon yet severe condition characterized by sustained elevation of blood pressure in the pulmonary arteries. The delaying treatment can result in disease progression, right ventricular failure, increased risk of complications, and even death. Early recognition and timely treatment are crucial in halting PH progression, improving cardiac function, and reducing complications. Within this study, we present a highly promising hybrid model, known as bERIME_FKNN, which constitutes a feature selection approach integrating the enhanced rime algorithm (ERIME) and fuzzy K-nearest neighbor (FKNN) technique. The ERIME introduces the triangular game search strategy, which augments the algorithm's capacity for global exploration by judiciously electing distinct search agents across the exploratory domain. This approach fosters both competitive rivalry and collaborative synergy among these agents. Moreover, an random follower search strategy is incorporated to bestow a novel trajectory upon the principal search agent, thereby enriching the spectrum of search directions. Initially, ERIME is meticulously compared to 11 state-of-the-art algorithms using the IEEE CEC2017 benchmark functions across diverse dimensionalities such as 10, 30, 50, and 100, ultimately validating its exceptional optimization capability within the model. Subsequently, employing the color moment and grayscale co-occurrence matrix methodologies, a total of 118 features are extracted from 63 PH patients' and 60 healthy individuals' images, alongside an analysis of 14,514 recordings obtained from these patients utilizing the developed bERIME_FKNN model. The outcomes manifest that the bERIME_FKNN model exhibits a conspicuous prowess in the realm of PH classification, attaining an accuracy and specificity exceeding 99%. This implies that the model serves as a valuable computer-aided tool, delivering an advanced warning system for diagnosis and prognosis evaluation of PH. •The performance of ERIME algorithm is enhanced by Triangular gaming search and Random follower search.•Compared with other high-performance optimizers, ERIME obtains higher quality optimal solutions in IEEE CEC 2017 functions.•The bERIME for pulmonary hypertension is proposed using FKNN.•bERIME_FKNN has technical advantages in the analysis of pulmonary hypertension.•bERIME_FKNN can be used as a tool to assist in the diagnosis of pulmonary hypertension.

As carbon cycling and global environmental protection gain increasing attention, forest disturbance research has intensified worldwide. Constrained by limited data availability, existing frameworks often rely on extracting individual spectral bands for simple binary disturbance detection, lacking systematic approaches to visualize and classify causes of disturbance over large areas. Accurately identifying disturbance types is critical because different disturbances (e.g., fires, logging, pests) exhibit vastly different impacts on forest structure, successional pathways and, consequently, forest carbon sequestration and storage capacities. This study proposes an integrated remote sensing and deep learning (DL) method for forest disturbance type identification, enabling high-precision monitoring in Northeast China from 1992 to 2023. Leveraging the Google Earth Engine platform, we integrated Landsat time-series data (30 m resolution), Global Forest Change data, and other multi-source datasets. We extracted four key vegetation indices (NDVI, EVI, NBR, NDMI) to construct long-term forest disturbance feature series. A comparative analysis showed that the proposed convolutional neural network (CNN) model with six feature bands achieved 5.16% higher overall accuracy and a 6.92% higher Kappa coefficient than a random forest (RF) algorithm. Remarkably, even with only six features, the CNN model outperformed the RF model trained on fifteen features, achieving a 0.4% higher overall accuracy and a 0.58% higher Kappa coefficient, while utilizing 60% fewer parameters. The CNN model accurately classified forest disturbances—including fires, pests, logging, and geological disasters—achieving a 92.26% overall accuracy and an 89.04% Kappa coefficient. This surpasses the 81.4% accuracy of the Global Forest Change product. The method significantly improves the spatiotemporal accuracy of regional-scale forest monitoring, offering a robust framework for tracking ecosystem dynamics.

Point cloud registration is a multifaceted problem that involves a series of procedures. Many deep learning methods employ complex structured networks to achieve robust registration performance. However, these intricate structures can amplify the challenges of network learning and impede gradient propagation. To address this concern, the soft-hard correspondence (SHC) framework is introduced in the present paper to streamline the registration problem. The framework encompasses two modes: the hard correspondence mode, which transforms the registration problem into a correspondence pair search problem, and the soft correspondence mode, which addresses this new problem. The simplification of the problem provides two advantages. First, it eliminates the need for intermediate operations that lead to error fusion and counteraction, thereby improving gradient propagation. Second, a perfect solution is not necessary to solve the new problem, since accurate registration results can be achieved even in the presence of errors in the found pairs. The experimental results demonstrate that SHC successfully simplifies the registration problem. It achieves performance comparable to complex networks using a simple network and can achieve zero error on datasets with perfect correspondence pairs.

Ultra-thin asphalt overlay has become the mainstream measure of road preventive maintenance due to its good economic benefits and road performance. However, hot mix asphalt concrete technology is widely used at present, which is not the most ideal way to promote energy saving and emission reduction in the field of road maintenance. At the same time, the ultra-thin friction course based on cold mix technology, such as slurry seal layer, micro-surface, and other technologies, are still far behind the hot mix friction course in terms of crack resistance. In this research, by establishing an integrated design of materials and structures, a cold paving technology called “high-toughness cold-mixed ultra-thin pavement (HCUP)” is proposed. The high-viscosity emulsified bitumen prepared by using high-viscosity and high-elasticity modified bitumen is used as the binder and sticky layer of HCUP. The thickness of HCUP is 0.8–2.0 cm, the typical thickness is 1.2 cm, and the nominal maximum size of the coarse aggregate is 8 mm. Indoor tests show that HCUP-8 has water stability, anti-skid performance, high temperature performance, peeling resistance, and crack resistance that are not weaker than traditional hot-mixed ultra-thin wear layers such as AC-10, Novachip, and GT-8. At the same time, the test road paving further proved that HCUP-8 has excellent road performance with a view to providing new ideas for low-carbon and environmentally friendly road materials.

AbstractsQuantitatively defining the relationship between laser powder bed fusion (LPBF) process parameters and the resultant microstructures for LPBF fabricated alloys is one of main research challenges. To date, achieving the desired microstructures and mechanical properties for LPBF alloys is generally done by time-consuming and costly trial-and-error experiments that are guided by human experience. Here, we develop an approach whereby an image-driven conditional generative adversarial network (cGAN) machine learning model is used to reconstruct and quantitatively predict the key microstructural features (e.g., the morphology of martensite and the size of primary and secondary martensite) for LPBF fabricated Ti-6Al-4V. The results demonstrate that the developed image-driven machine learning model can effectively and efficiently reconstruct micrographs of the microstructures within the training dataset and predict the microstructural features beyond the training dataset fabricated by different LPBF parameters (i.e., laser power and laser scan speed). This study opens an opportunity to establish and quantify the relationship between processing parameters and microstructure in LPBF Ti-6Al-4V using a GAN machine learning-based model, which can be readily extended to other metal alloy systems, thus offering great potential in applications related to process optimisation, material design, and microstructure control in the additive manufacturing field.