Individualized Differential Gene Expression Studies Based On Relative Expression Abundance Ranking

High-throughput gene chip technology has been widely used in basic medical research fields such as tumor pathogenesis and tumor prognostic marker identification.In these studies,the most basic analysis of gene expression profile data is to identify differential expression genes(DEG) between the two types of samples.However,the widely used DEG recognition method can only identify DEGs at group level,and cannot provide patient-specific information for individualized diagnosis and treatment.Researchers have proposed an algorithm based on relative expression orderings(REO) for individualized differential expression analysis.However,these methods treat all gene pairs with reversal REO patterns in disease samples equally in the process of differential expression recognition,and ignore the fact that other factors may affect the contribution of genes,and there are certain defects.The application of differentially expressed gene recognition algorithms in cancer has always been a research hotspot in bioinformatics.Triple-negative breast cancer(TNBC) has become a more serious type of breast cancer due to its high risk of death.At present,the research on differential expression of TNBC genes mostly analyzes differentially expressed genes at the population level;however,the heterogeneity of TNBC is high,and the use of differentially expressed genes at the population level may not accurately represent the differential expression information of TNBC.In order to solve the above problems,this paper has done the following work on the differentially expressed gene recognition algorithm and its application in breast cancer:(1)Considering the influence of the rank position and relative rank difference of genes on the contribution of genes in DEG recognition,this paper designs a new algorithm IndDEA based on REO.In order to better compare the performance of IndDEA algorithm,this paper applies RankComp V2,an algorithm for group-level DEG recognition,to the field of individualized DEG recognition,and compares it with IndDEA algorithm with RankComp and Pen DA.The IndDEA algorithm performed well in both simulated and real lung data,and RankComp V2 performed slightly worse than IndDEA but also performed stable.At the same time,this paper applies IndDEA and RankComp V2 to the ischemic cardiomyopathy data collection.In the enrichment experiment of DEG,new biological pathways related to ICM are discovered,which proves that the DEG results of RankComp V2 and IndDEA algorithms are biological Significance,can provide new information for the study of disease mechanism.(2)The designed IndDEA algorithm was applied to find individualized DEGs in TNBC data,and further studies were carried out with these individualized DEGs.Through the analysis of the frequency and clinical information of these DEG samples,it was found that the primary tumor stage,lymph node staging,and tumor grade were significantly related to the frequency of sample disorder.At the same time,this study found a series of highly up-regulated,down-regulated and highly stable genes from these DEGs.Through the functional enrichment analysis of these genes,the highly up-regulated and highly stable biological pathways in TNBC were obtained.In the clustering and subtype analysis of these DEGs,the two subtypes obtained through hierarchical clustering can distinguish disease samples with different lymph node stages.Survival analysis showed that there was a difference between the disease-free survival time and the overall survival time between the two groups.