Bioinformatic Analysis To Predict Small Non-Coding Rnas As Potential Biomarkers For Her2 Positive Breast Cancer

Background and aim: breast cancer(bc) is one of the most common cancers and main cause of cancer deathin women worldwide. based on cell surface receptors, different classifications are considered.(her2)/neu is a growth factor receptor gene that is amplified or overexpressed in her2 enrichedsubtype and it is associated with more aggressive disease. due to this aggressive behavior of her2positive tumors, early detection is vital which has been improved by using biomarkers.micrornas(mirnas) seem to be suitable biomarkers which have high sensitivity and stability as wellas they can be detectable in non-invasive methods. expression of mirnas in many common humancancers is changed, so they could be considered as potential biomarkers for early detection of breastcancer. to identify crucial candidate biomarkers for breast cancer, we analyzed mirna expressionprofiles of bc from the cancer genome atlas(tcga) breast cancer mirna/rnaseq dataset. weidentified some mirnas by combination bioinformatics analysis. methods: several bioinformatics tools based on feature selection were utilized to investigate the mostimportant biomarkers. data of the gse69085 study were obtained from tcga and some tools likelimma used for differential expression also xgboost and svm_1 were used. then we selected mirnaswhich demonstrated the most deferentially expressed between her2 positive samples and her2negative ones. results: in this study, all data was merged and calculated by some features like log2 fold change(logfc) and adjusted pvalue ratios. we showed mir-30, mir-149, mir635, mir-190b, mir-184,mir-425 could be considered as potential biomarkers among 2500 non-coding rnas. conclusion: big data and bioinformatics algorithms is utilized to predict signature mirnas as biological markers in breast cancer.