Identification And Study Of Cancer Biomarkers In Combination With Multi Omics Data

In the field of cancer research and medicine, biomarkers can diagnose the condition of cancer patients early, provide timely treatment, and also can predict the condition of cancer, having a very high guiding value to cancer treatment. Many studies have reported that genes can be used as candidate biomarkers that are used in the diagnosis of disease or cancer, prognosis and efficacy.With the development of high-pass sequencing technology, the study of cancer biomarkers began to develop from single-omic data to multi-omics data, but the integration of multi-omics data remained in the simple integration stage. We can not find the internal relationship between multiple sets of data, we study from the gene expression and DNA methylation integration data for cancer biomarkers research and analysis.The research contents of this thesis are as follows:1?The traditional feature selection method often considers the high classification performance of the feature selection results in the high dimensional small sample data,while ignoring the stability of the feature selection results. In this paper,we propose to retain the putative cancer associated important genes in the characterization of gene expression data and to obtain a stable combination of gene features.2?Since the 450K methylation chip covers only 2% of the total methylation site,the use of a simple fusion approach may result in biased results. In this paper, we propose the method to fuse the extended 450K methylation microarray data and gene expression data for the first time, analyze cancer biomarkers from multiple levels, and use existing DNA methylation data as much as possible to fuse multiple sets of data to retain more information, to be stable and reliable with the promotion of potential cancer biomarkers, This method is more accurate and reliable than traditional method. Analysis of a variety of cancer-specific potential cancer biomarkers and a variety of cancer common potential biomarkers of cancer for medical research and clinical treatment to provide guidance and help.3?Finally, in this paper, we construct a classifier model based on fuzzy rules to verify the classification effect of the potential cancer biomarkers for cancer, and to compare the simple fusion of our methods with traditional methylation data and DNA methylation data by cross validation method, we found that our method is superior to the traditional method, and our method is better than the traditional method of predicting the independent sample. Basing on the found potential cancer biomarkers, we get a more robust and easy to understand classification rule system.