Reasearch On The Algorithm Of Recognizing Cancer Driver Pathway Based On Multi-omics Data

Because of the importance of driving pathways for cancer precision medicine and personalized medicine,the problem of driving pathway identification has become a research hotspot in bioinformatics.The ensemble data generated by high-throughput sequencing technology is noisy and incomplete,and a single set of data contains limited information.So,the integration of multi-omics data not only improves the integrity and accuracy of the data,but also makes full use of the potential information in different omics data.This paper studies the problem of identifying driver pathway based on multi-omics data.The main work is as follows:An improved maximum weight submatrix problem model is proposed by integrating such three kinds of omics data as somatic mutations,copy number variations,and gene expressions.The model adjusts coverage and mutual exclusivity with the average weight of genes in a pathway,and simultaneously considers the correlation among genes,so that the pathway having high coverage but moderate mutual exclusivity can be identified.By introducing a kind of short chromosome code and a greedy based recombination operator,a parthenogenetic algorithm PA-IMWS is presented to solve the IMWS model.Experimental comparisons among algorithms GA,MOGA,iMCMC and PA-IMWS were performed on biological and simulated data sets.The experimental results show that,compared with above three algorithms,the PA-IMWS one based on the improved model can identify the gene sets with high coverage but moderate mutual exclusivity and scales well.An improved model for identify co-occurring mutated driver pathways is proposed by integrating omics data as somatic mutations,copy number variations,and gene expressions.The model has two mutational characteristics:(1)each pathway has high coverage and moderate mutual exclusion;(2)the mutated genes between the cooperatively driver pathways show significant co-occurrence in cancer samples.In addition,the relevance of all genes in the collaboratively driver pathway needs to be considered.By introducing a kind of short chromosome code and a greedy based recombination operator,a parthenogenetic algorithm PA-ICMDP is presented to solve the ICMDP model.Experimental comparisons among algorithms CoMDP and PA-ICMDP were performed on biological and simulated data sets.The experimental results show that,the PA-ICMDP based on the improved model can identify multiple important co-occurring mutation-driver pathways involved in key biological processes,such as cell survival and protein synthesis.PA-ICMDP method is suitable for mining genes related to cancer development.In addition,the EICMDP model and PA-EICMDP algorithm are proposed by extending the ICMDP and PA-ICMDP.Experimental results show that the extended model can effectively identify other important pathways that interact with known pathways.In summary,in this paper,the problem of cancer-driven pathway identification is studied.Based on the cancer multi-omics data,models and algorithms for pathway identification problems was proposed,which may be useful supplemental tools for detecting cancer pathways.