High-order Subgraph Expression Pattern Recognition And Application Based On Spatial Transcriptome

With the rapid development of high-throughput sequencing technology,gene expression data with tissue location information can be measured.This is essential for analyzing the function and phenotype of cells in multicellular organisms,completely changing the research of tissue function and disease pathology,and making the related research of spatial transcriptomics become the international frontier of current life science and medical research.However,its potential is limited by current data analysis strategies.Most of the current research studies the recognition of the spatial expression pattern of a single gene,ignoring the influence of multiple genes,but the high-order interaction between genes Interactions have been proven to be very helpful for understanding genomics data.Therefore,studying the high-order expression patterns of spatial transcriptomes of interactions between multiple genes is a scientific interpretation of expression patterns in biological tissues.The key issue is also a major challenge posed by the spatial heterogeneity of life science research to computational science.In order to solve the above-mentioned problems in life science research work,this article gives the definition of high-order expression patterns of the spatial transcriptome,and develops a calculation method for identifying and characterizing spatially specific regulatory edges or the high-order expression mode of the regulatory motif.The algorithm can identify and characterize spatially specific regulatory edges or regulatory motifs.In this thesis,the activities of the regulatory edges and regulatory motifs with spatial location information are quantified by geometric standard scores,random walks,single samples,and other methods,and then based on a model that recognizes spatially-specifically expressed genes,a recognition spatial transcriptome is constructed.The algorithm framework of high-order expression patterns,and the statistical significance of the algorithm is verified.With the help of the regulatory relationship data of the Kyoto Gene and Genome Encyclopedia database and the spatial transcriptome data of existing breast cancer samples,this thesis applies the algorithm to identify and analyze the spatially specific high-order expression of the spatial transcriptome in breast cancer samples.Through extensive verification and analysis on spatial transcriptome data,the high-order pattern recognition algorithm of spatial transcriptome designed in this thesis can effectively identify the spatially specific high-order expression patterns of spatial transcriptome data,and provide relevant research on spatial transcriptome new platforms and tools.At the same time,we used the identified high-order expression pattern of the spatial transcriptome as a new spatial feature for experimental analysis and application.The results show that the high-order expression pattern of the spatial transcriptome identified by the algorithm in this thesis is significant in topological analysis.,Has significant biological significance in enrichment analysis.After it is applied to automatic expression histology,it can also reflect the internal structure of the original biological tissue to a certain extent.Therefore,the recognition and application algorithms of high-order patterns in the spatial transcriptome provide a new perspective for studying the spatial heterogeneity of tissues,and also provide a platform for systematic analysis of cancer from the perspective of high-order patterns.