Analysis Of The Prognostic Value And Risk Model Construction Of EMT-related MiRNAs In Breast Cancer Based On TCGA Database

Background and purpose The incidence of breast cancer(BC)has surpassed that of lung cancer as the most common cancer worldwide and the leading cause of cancer death among women worldwide.Although the survival rate of breast cancer has been significantly improved with the development and progress of diagnostic techniques,surgical treatment and adjuvant therapy,recurrence and metastasis are still the root cause affecting the prognosis of breast cancer patients.One of the most common factors in tumor metastasis is epithelial-mesenchymal transition(EMT).During the EMT process,epithelial cells lose their original polarity and intercellular adhesion capabilities,and at the same time gain the ability of mesenchymal cells to migrate and invade surrounding tissues,thereby causing distant metastasis of tumor cells and promoting tumor development.,affecting the prognosis of patients.Studies have shown that the EMT process of tumor cells is regulated by miRNA.miRNA is one of many types of non-coding RNA,usually 20-25 nucleotides in length,which can bind to the 3’-untranslated region(UTR)of target m RNA through incomplete sequence matching to inhibit its translation and stability.It participates in various processes of cells,including cell proliferation,differentiation,apoptosis,stress resistance,fat metabolism,and tumor occurrence and metastasis.This study used the miRNA expression data in the TCGA database to find EMT-related miRNAs in breast cancer,and constructed a breast cancer molecular prognosis model based on EMT-related miRNAs,which provided a certain reference value for predicting the prognosis of breast cancer patients.Methods The breast cancer miRNA expression data and patient clinical information were downloaded from the TCGA database,and differentially expressed miRNAs were screened.Combining the patient’s survival information with the differentially expressed miRNA data,survival analysis was performed to obtain the differential miRNAs that affect the patient’s prognosis,and then the target genes of the differential miRNAs related to prognosis were predicted using the miRDB,miRTar Base and Target Scan databases,and the target genes were correlated with EMT The related genes were intersected,and the obtained genes were subjected to functional enrichment analysis and protein interaction network construction to obtain EMT-related miRNAs.Patients with complete clinical information were randomly divided into a modeling group and a verification group at a ratio of 7:3,and the survival information of the modeling group was included,and single-factor COX regression analysis and multi-factor COX regression analysis were performed to finally obtain miRNAs that could be used to build risk models Expression profile,use its regression coefficient to calculate risk score,build prognostic risk model and calculate ROC value,and evaluate the sensitivity and specificity of the model.The survival information of the verification group was incorporated into the risk score calculation equation,and the ROC value was calculated to verify the accuracy of the risk score model.Then,the calculated risk score of the modeling group was combined with the patient’s age and TNM stage for univariate COX regression analysis and multivariate COX regression analysis to obtain independent factors affecting the prognosis of breast cancer patients.Using the "survival" package of R language and The "rms" package draws the nomogram,draws the calibration curves of the nomogram 1,3,and 5 years,calculates the C index,and evaluates the accuracy of the nomogram.Results The miRNA expression data of 1090 breast cancer tissues and 103 normal tissues were downloaded from the TCGA database,and a total of 290 differentially expressed miRNAs were obtained through screening,including 200 up-regulated miRNAs and 90 downregulated miRNAs.Patients with incomplete survival information were excluded,and the survival time and survival status of 1064 patients were included,and combined with differentially expressed miRNAs for survival analysis,a total of 33 miRNAs related to breast cancer prognosis were obtained(P<0.05).The target genes of 33 miRNAs were predicted by miRDB,miRTar Base and Target Scan databases,and a total of 858 target genes were obtained.1184 EMT-associated genes(EAGs)were downloaded from the db EMT2.0 database.By intersecting 858 target genes and 1184 EAGs,a total of 114 genes were obtained,corresponding to 18 miRNAs.Gene enrichment GO results show that biological processes(BP)are mainly enriched in cell development regulation,cellular components(CC)are mainly enriched in transcription regulation and autophagy,and molecular functions(MF)are mainly enriched in protein serine/threonine Kinase activity and protein phosphatase binding.KEGG results showed that it was mainly enriched in cell senescence.The constructed protein interaction network showed that the top 10 core genes were SMAD3,MAPK1,STAT3,PIK3R1,RAC1,HIF1 A,MYC,TGFBR2,VEGFA,SIRT1.Including the survival information of the modeling group,after univariate and multivariate COX regression analysis,six miRNAs(hsa-miR-204-5p,hsamiR-195-3p,hsa-miR-877-5p,hsa-miR-605-5p,hsa-miR-106a-5p and hsa-miR-556-3p)were used to construct risk models.The modeling group was divided into high and low risk groups,and the obtained K-M survival curve showed that the overall survival prognosis of the high risk group was poor,and the difference between the two groups was statistically significant(P<0.001),and the 3-year and 5-year ROC The areas under the curves were 0.702 and 0.709,respectively.Similarly,the K-M survival curves of the two groups in the verification group showed that the overall survival prognosis of the highrisk group was poor,and the difference between the two groups was statistically significant(P=0.01),and the 3-year and 5-year ROC The areas under the curves were 0.612 and 0.644.It is proved that the model has strong feasibility.The risk model can be included in the nomogram with age and TNM stage to predict the 1-,3-,and 5-year survival rates of patients,and the C index is 0.776,which shows that the prediction result of the nomogram is more accurate.Conclusion The breast cancer molecular prognosis model based on EMT-related miRNA was successfully constructed,which can better predict the prognosis of breast cancer patients.