Chongfan Technology
News
20
2024
-
08
The latest Nature sub-journal of Harbin Institute of Technology Jiang Qinghua and other teams
Author:
Inference of intercellular communication (CCC) is essential for a better understanding of complex cellular dynamics and regulatory mechanisms in biological systems. However, accurate extrapolation of space at single-cell resolutionCCCsremains a major challenge.
August 18, 2024,Harbin Institute of TechnologyJiang Qinghua, Ji Yong, Harbin Medical University,Xu Zhaochun Common NewsletterInNature Communications Published online entitled“Deciphering cell–cell communication at single-cell resolution for spatial transcriptomics with subgraph-based graph attention network”The research paper,The studyA general method called DeepTalk is proposed to infer spatial CCC at single-cell resolution by integrating single-cell RNA sequencing (scRNA-seq) data and spatial transcriptomics (ST) data.
DeepTalk utilization diagramConcernNetwork (GAT) Integrate scRNA-seq and ST data to enable accurate cell type identification of single-cell ST data and deconvolution of point-based ST data. DeepTalk can then leverage subgraph-based GAT to capture connections between multiple levels of cells, further enabling spatially resolved CCC inference at single-cell resolution.DeepTalk has achieved excellent results in discovering meaningful spatial CCCs on multiple cross-platform datasets, demonstrating its superior ability to dissect cellular behavior in complex biological processes.
Intercellular communication is a basic biological process, which plays a vital role in immune cooperation, organ development, stem cell niche and other biological phenomena.Advances in single-cell transcriptomics, particularly single-cell RNA sequencing (scRNA-seq), have revolutionized the study of single cells, providing unprecedented insights into their composition, function, and dynamics. Calculation tools suchCellPhoneDBCellChat, NicheNet, CytoTalk, scTensor, iTALK, ICELLNET, SingleCellSignalR, and Scriabin have been developed to infer and decipher CCC networks. These tools employ innovative methods and statistical tests to quantify the probability of each interaction, helping to evaluate the null hypothesis and inferences about specific cell types involved in cell-to-cell communication. In addition, NicheNet and CytoTalk revealed complex gene-gene interactions within cells in response to external stimuli, while scTensor utilized hypergraphs to reveal the complexity of higher-order intercellular communication and visualize complex interaction networks. Nevertheless, a significant drawback of current single-cell analysis methods is that they do not fully capture spatial information. The lack of cellular spatial information increases the probability of false positives, resulting in physically separated cells being misclassified as those involved in cell-to-cell communication.Therefore, spatial information should be incorporated into the CCC analysis to promote the development of the field and enhance the understanding of this complex process.
Spatial transcriptomics (ST) provides valuable spatial information about cells or points that are composed of multiple or partial cells.These techniques facilitate the measurement of spatial gene expression in two-dimensional (2D) or three-dimensional (3D) tissue samples with varying degrees of cellular resolution. The introduction of ST significantly improves the accuracy and reliability of spatial CCC reasoning in biology and biomedical fields. Recently, several approaches have emerged to decipher the underlying mechanisms of climate change in the spatial context. For example, CellPhoneDBv3(CPDB3) limits the interaction between cell types located in the same microenvironment, which is determined by spatial information. SVCA and MISTy employ probabilistic and machine learning models, respectively, to identify spatially correlated gene interactions between cells. NCEM uses a function to fit relationships between cell type, spatial environment, and gene expression.In addition, SpaOTsc perform a structured optimal transport mapping between scRNA-seq and ST data, assign the spatial location of cells, and use cell-cell distance as a transport cost to infer the ligand-receptor signaling network that mediates spatial constraints.
Although this method quantifies the possibility of cell-cell interactions, its specificity to image-based spatial expression data and its reliance on predefined paths limit its versatility and applicability to a certain extent.COMMOT is an extension of the SpaOTsc optimal transport framework. Considering the complex ligand-receptor interaction and the limitation of effective cell-cell communication distance, the optimal transport analysis tool is applied to ST data to infer the communication direction. These methods have great promise for direct analysis of CCC in the spatial context. The limitations encountered, however, stem from gene throughput and spatial resolution of ST data. In addition, most of these methods mainly focus on the identification of CCC between paired cell types, while ignoring CCC analysis between paired individual cells.There are still few methods that can simultaneously overcome the limitations of ST data and infer CCC at single-cell resolution, thereby limiting the understanding of the coordinated activities exhibited by various cells in biological processes.
The study developed DeepTalk, an innovative method that combines cell-specific gene expression data and cell spatial affinity, to predict CCC at single-cell resolution.DeepTalk employ graph attention networks (GATs) and subgraph-based GATs to uncover the complex mechanisms of CCCs in the spatial context of healthy and diseased tissues. By integrating ST data with scRNA-seq data from the same region, the limitations of limited gene flux and insufficient spatial resolution of ST data are effectively overcome. Extensive evaluation using various publicly available datasets verified the superior performance and robustness of the DeepTalk in identifying spatial CCCs.The findings show that DeepTalk have great potential in discovering meaningful CCC patterns under different conditions and provide valuable insights into spatial intercellular dynamics within tissues.
LATEST NEWS
Thank you for visiting the official website of Chongfan Technology. If you have cooperation intentions or suggestions, please contact us through the following methods, and we will reply as soon as possible, thank you!
Address: Room 403, Building 6, Phase III of R&D, No. 36 Xiyong Avenue, High tech Zone, Chongqing, China.
Telephone: +86-13658337211
E-mail: Sales@cfkeji.net
Website: www.cfkeji.net
Mobile Version
