Research And Application Of Cancer Cooperative Driver Pathway Identification Method Based On Multi-omics Data

Identifying the key drivers of cancer occurrence from large-scale cancer data is crucial for a comprehensive understanding of the carcinogenic process,which helps researchers achieve accurate diagnosis of cancer,rapid development of targeted drugs,and personalized treatment for clinical cases.Early studies primarily screened genes with significant high-frequency mutations as driving genes.However,this method was greatly influenced by environmental factors and lacked universality.In fact,driver genes typically aggregate into a pathway and cooperatively drive the occurrence of cancer.Pathways also commonly drive the carcinogenic transformation process of normal cells through cooperation.Therefore,identifying the cooperative driver pathway has become a major focus of cancer research.Many methods have been proposed,but they still have problems such as excessive dependence on batch sequencing data leads to susceptibility to noise,insufficient utilization of other important cancer data except gene-related data.Furthermore,inconsistent format makes it difficult to compare and analyze the results.Therefore,it is still highly significant to develop more efficient methods and integrate multiple methods to form robust recognition systems.To address the problem that the existing methods are usually only use batch data and ignore the problem of cellular heterogeneity,we propose a novel method called CDMFinder based on single-cell data and prior knowledge.CDMFinder firstly utilizes gene co-expression specificity from various cancer subtypes and normal cell expression data to construct expression association networks,and fuses these networks with gene interaction network to obtain a gene functional association network,and thus effectively reduces network complexity while capturing in-depth functional associations between genes.It then adopts an overlapping Markov clustering on this functional network to mine functional clusters,along with a greedy strategy with hybrid weight function to identify driver pathways from these clusters.Finally,a distance function based on the co-occurrence of interaction and mutation is introduced to identify cooperative driver pathway sets.CDMFinder fully integrates a variety of genetic factors,including expression,mutation,and subtype specificity,and demonstrates excellent performance.To address the problem that the existing methods mainly focus on genomics data,neglect known pathway information and other related multi-omics data,thus cannot faithfully decipher the carcinogenic process,we propose CDPMiner to discover cooperative driver pathways by multiplex network embedding,which can jointly model relational and attribute information of multi-type molecules.CDPMiner first uses the pathway topology to quantify the weight of genes in different pathways,and optimizes the relations between genes and pathways.Then it constructs an attributed multiplex network consisting of miRNAs,lncRNAs,genes and pathways,embeds the network through deep joint matrix factorization to mine more essential information for pathway-level analysis and reconstructs the pathway interaction network.Finally,CDPMiner leverages the reconstructed network and mutation data to define the driver weight between pathways to discover cooperative driver pathways.CDPMiner can effectively fuse multi-omics data to discover more and better driver pathways,which indeed cooperatively trigger cancers and are valuable for carcinogenesis analysis.Aiming at the problem that the existing methods are usually developed in different languages and the data format and result presentation form are also different,which makes it challenging for researchers to compare and analyze driver pathways identified by different methods,we carry out a unified reconstruction and optimization of various existing methods,integrate and form a cooperative driver pathway online identification system called CDPTool.The system displays the identification results in a visual and unified manner,making it easier for researchers in relevant fields to compare and analyze driver pathways conveniently and quickly.