Screening Of Mirna For Poor Prognosis Of Breast Cancer Based On TCGA Database And Risk Evaluation

Objective:The cancer with the highest incidence among women worldwide is breast cancer,which greatly affects the lives of women around the world and is the main cause of death among women.In China,breast cancer has an increasing incidence year by year.At present,the mainstream treatment methods for breast cancer include surgery,endocrine therapy,chemical therapy and radiation therapy.Although the cure rate of breast cancer is high,there are still many patients with relapse due to drug resistance or metastasis leading to poor prognosis.Therefore,it is important to find biomarkers that can predict the poor prognosis of breast cancer.There is currently no joint indicator for accurate diagnosis of breast cancer prognosis.MicroRNA（miRNA）is an endogenous,non-coding small RNA that is typically about 22 nucleotides in size and binds to the3’-UTR of the targeted gene to further regulate transcription of the gene in the negative direction.Post expression level.A single miRNA can simultaneously control hundreds of genes and play important regulatory roles in various biological processes of cells.For example,differentiation,apoptosis and proliferation have shown that miRNA are involved in many tumors such as breast cancer.Abnormal expression is involved in the process of tumorigenesis and drug relapse.miRNA have great potential as biomarkers for cancer,which is helpful for diagnosis and prognosis.Although there have been reports on miRNA as tumor markers,there is still a lack of more sensitive and specific combinatorial markers for miRNA that judge the prognosis of breast cancer.Therefore,this study will screen ten miRNA of breast cancer samples in the TCGA database as a whole,and use the overall risk score to evaluate the prognosis of patients,which proves that it can be used as a biomarker with better sensitivity and better prognosis.Methods:Tissue sample data was obtained from the Human Tumor Genome Database,the TCGA database（full name in English,The Cancer Genome Atlas）.The database was developed by the National Cancer Center and the National Human Genome Research Institute,and 33 tumor types were included.The data of miRNA in breast cancer patients in the TCGA database were processed,and fold change>2 and FDR<0.05 were set as threshold values.miRNA with significant differences in expression of miRNA in cancer tissues and adjacent tissues were screened out.A normal tissue and 1103 breast cancer tissue samples were drawn using the R language package（Edger package）.The Vlookup function is used to integrate patient survival data with expression data and clinical pathology parameters.We use R（R version 3.3.3）,use the Survival package to perform Cox univariate regression analysis using the Coxph function model,output deviation coefficient（β）,hazard ratio（HR）,P value,and so on.Variables that may affect patient survival were obtained by Cox univariate regression analysis,further used as a single-line variable for the Cox regression model in the analysis,using R,using the Survival package to perform Cox multivariate regression analysis using the Step function.Ifβ<0,it indicates that the variable is a protection factor,such asβ>0,indicating that the variable is a risk factor;correspondingly,if HR<1,it indicates that the variable is a protection factor,and if HR>1,it indicates the variable As a risk factor;P<0.05 as the significance test level,P<0.05,indicating that the variable is an independent prognostic factor.Our team used R,using the Survival package to use the Predict function model,and calculated the risk score for each patient based on the formula risk score=h₀（t）(Exp _miRNA1*β_miRNA1+……+Exp_miRNAn*β_miRNAn),according to the risk score.Values divide patients into two groups.Statistical analysis was performed using R（R version 3.3.3）.Data are expressed as mean±standard deviation（SD）and statistically compared by paired t-test.A P value<0.05 was considered to be statistically significant.Results:The results of two-dimensional hierarchical clustering of 491 sample differentially expressed genes were obtained.A total of 370 miRNA were screened and significantly different from normal tissues,including 108 down-regulated 108 miRNA and significantly up-regulated 262 miRNA.2.Univariate Cox regression analysis was used to evaluate the relationship between the expression levels of 262 miRNA（DEMis）and the overall survival（OS）of the patients.It was found that 39 miRNA were significantly associated with OS（P<0.05）,and OS-related up-regulation There are 21 miRNA.3.Stepwise multivariate Cox regression analysis of the first 21 miRNA,of which 10miRNA were included in the prediction model.The predictive model was defined as a linear combination of the expression levels of ten-miRNA weighted by their relative coefficients in a multivariate Cox regression test,Survival Risk Score（SRS）=（0.3968×Exphsa-mir-148b）+（-0.1469×Exphsa-mir-449c）+（-0.1878×Exphsa-mir-106a）+（-0.1356×Exphsa-mir-181d）+（0.0757×Exphsa-mir-9-3）+（0.3366×Exphsa-mir-549a）+（-0.1712×Exphsa-mir-556）+（-0.1234×Exphsa-mir-618）+（0.3965×Exphsa-mir-466）+（-0.0837×Exphsa-mir-135a-1）.The kaplan-Meier total survival and ROC curves were analyzed separately.The 10 miRNA did not have an ideal evaluation effect.4.According to the ten-miRNA median risk score,1176 patients in the study were divided into high-risk group（n=538）or low-risk group（n=538）,and kaplan-Meier overall survival curve showed prognosis of high-risk score patients.Significantly poor（log-rank P=0<0.001）.The AUC value of the ROC curve analysis was 0.712,indicating that the ten-miRNA feature model has good sensitivity and specificity in predicting the risk of survival in breast cancer patients.5.TargetScan and miRDB online analysis tools were used to predict the target genes of ten-miRNA,and 565 target genes were obtained by overlapping the mRNA of breast cancer tissues in the TCGA database.GO function analysis showed that the target genes were involved in biological processes such as anatomical morphogenesis.The plasma membrane fraction is composed of cytological components,RNA polymerase II transcription factor activity and other molecular functions;KEGG pathway enrichment analysis shows that the target gene participates in signaling pathways such as Ras.Conclusion:1.The combination of 10 miRNA in the TCGA breast cancer data identified by the COX risk regression model,namely hsa-mir-148b,hsa-mir-449c,hsa-mir-106a,hsa-mir-181d,hsa-mir-9-3,hsa-mir-549a,hsa-mir-556,hsa-mir-618,hsa-mir-466,hsa-mir-135a-1 constitute a combined marker,and the prognosis of breast cancer patients with high-risk scores is significantly poor.It has good sensitivity and specificity and can be used as a reliable biomarker for predicting the prognosis of breast cancer patients.2.The target gene of ten-miRNA is involved in biological processes and signaling pathways related to the occurrence and development of breast cancer.