Study On Cancer-driven Gene Recognition Based On Cancer Genome Data

Cancer is a major life-threatening and health-threatening disease,and research in cancer-related fields remains a hotspot for bioinformatics research.Numerous studies have shown that very few important genes play a decisive role in the process of cancer.Therefore,an important task in cancer genome research is to promote mutations or genes.Function-related driver mutations in the genome,also called driver pathways or modules,activate the mechanism of cancer development,induce cancer,and give cancer cells selective benefits.Therefore,accurate identification of cancer driving genes and more accurate diagnosis of cancer are very important for clinicians to effectively treat cancer.In this article,we will study the identification of cancer driver genes based on cancer genomic data.Research methods for identifying cancer driver genes can be divided into two categories,depending on the type of input data used.1)The "de novo" method relies solely on genetic data as input data for discovering new gene interactions.And cancerrelated driver genes 2)A combination of knowledge-based methods in the form of pathways,networks,and functional phenotypes,in addition to genomic data,and prior knowledge to identify driver genes.The main research method in this article is the latter.The main work of this article is introduced below.1)I propose a method to identify the cancer driver module based on community detection(CMBCD).This method integrates connection information in the form of protein-protein interaction(PPI),mutual exclusion,and coverage,and recursively uses the Louvain algorithm to identify cancer driver modules.With respect to the recovery of known oncogenes,CMBCD is superior to some state-of-the-art calculation methods for TCGA pancancer data,classifying normal and tumor samples,and a module rich in mutations of a particular cancer type.To provide.CMBCD identifies gene modules that contain known oncogenes and genes that rarely mutate in pancancer data.2)In this paper,a cancer-driven gene recognition method based on Multi-task learning(CDGBML)is proposed.It uses somatic genomic data to drive gene priorities,combine genomic information with previous biological knowledge,and accurately prioritize cancer driver genes through multitask learning.model.Through testing on various datasets,experimental results show that CDGBM performance is good enough and better than existing methods.This is an accurate and easy-to-use cancer genome driver gene sequencing method that helps improve the detection of cancer driver genes..And the accuracy of priorities facilitates accurate diagnosis,personalized treatment,and better clinical decisions.In summary,this article studies the identification of driver genes in the cancer genome and proposes two effective identification models and algorithms.Experimental results show that based on the proposed models and algorithms,some biologically important driving pathways and genes can actually be identified.