Identifying ovarian cancer micro RNA bio-Markers using a sequential wrapper method

A microRNA (miRNA) is a small non-coding RNA molecule. The main task of microRNA is the posttranscriptional regulation of gene expression. miRNAs can act as either oncogenes or tumor suppressors by targeting the expression of cancer-related genes. So, miRNAs can be used as biomarkers for the diagnosis, prognosis, and treatment of cancer. Microarray-based expression analysis is a common approach for detecting candidate miRNAs which are differentially expressed in normal and malignant tissue samples. Biomarkers finding is equivalent to a feature selection problem. The selection of a subset of features increases the accuracy of classification and reduces the cost of computation, clinical costs and the possibility of over-fitting, which is likely to be increased by increasing the number of miRNAs relative to the number of samples. In this study, a sequential wrapperbased approach was used to select biomarkers from miRNAs involved in ovarian cancer. This method provides the best prediction for the classification of cancerous and normal samples by selecting a subset of miRNAs sequentially and uses the LDA classifier. The proposed method identified 8 out of 2565 miRNAs as biomarkers that they can separate healthy and cancerous samples using 10-fold cross-validation and achieved an accuracy of 100%. These eight miRNAs include: hsa-miR-760, hsamiR-320b, hsa-miR-1290, hsa-miR-3197, hsa-miR-4258, hsa-miR-6131, hsa-miR-6800-5p .We evaluated the selected miRNAs by using their target genes and analyzed Gene-miRNA pathway by using Cytoscape Software. The analysis confirms the significant relationship between selected biomarkers and ovarian cancer

کلیدواژه ها:

Micro RNA ، Biomarker ، Feature Selection Method ، Sequential Wrapper Method