The Key Mediating Patterns Recognition Algorithm On Cancer Omics

Cancer is one of the main diseases threatening human life.By 2030,13 million people worldwide will die from cancer.With the rapid development of high-throughput technology,a large amount of cancer genome maps and other data have been generated,which provides the possibility to study the biological mechanism of cancer.At the same time,due to the complex,multi-level,and complementary characteristics of big data in cancer omics,its calculation,analysis and interpretation have brought great challenges.Existing studies have shown that somatic mutations and DNA methylation affect gene expression,which are closely related to the occurrence of cancer.Cancer omics data are both independent and interrelated.Research on the association relationship between omics can make better use of massive cancer omics big data,realize data-driven biomedical research,and promote people's understanding of the mechanism of cancer occurrence and development.And provide new strategies for early detection,prevention and adjuvant treatment of diseases.Based on genome,transcriptome and epigenetic data,this article systematically and multi-dimensionally analyzes big data of cancer genome map.Using graph theory to model and data mining multi-omics data,explore the association relationship of multi-omics data and propose mediation graph mode.From the perspective of network communication,a calculation method for identifying key mediation patterns between omics has been developed.The algorithm can identify the key mediator genes and key mediation pathways that characterize the relationship between omics.Based on the reverse local path propagation algorithm,this paper first constructed a mediation graph pattern recognition algorithm framework to solve the multi-source shortest path problem,and then measured the importance of mediation nodes and mediation paths through five improved centrality indicators,and then obtained omics The key mediates nodes and pathways.Through extensive verification and analysis on cancer omics data,the key mediation map pattern recognition algorithm designed in this paper can accurately identify the mediation genes and mediation pathways that describe the relationship between omics,and provide relevant research on omics relationship.New platforms and tools.We applied the algorithm to breast cancer somatic mutation and differential expression,copy number variation and differential expression,DNA methylation and differential expression data,identified key mediator genes and key mediation pathways between omics,and performed experiments on them Analysis and application verification.In the multi-source shortest path problem,the key mediation graph pattern recognition algorithm can identify mediation nodes and mediation paths more quickly than Freud algorithm and reverse local path propagation algorithm.It only uses 20% of the mediation nodes of the Freud algorithm,and less than 3% of the mediation paths can connect a set of source nodes and a set of target nodes with a similar number.The verification by biological saliency analysis shows that the key mediation diagram model has significant biological significance.Therefore,the key mediation graph pattern recognition and application algorithms provide a new perspective for the study of the relationship between the groups,and also provide a platform for the systematic analysis of cancer from the perspective of network propagation.