Gene Co-expression Network Analysis And Drug Prediction For Breast Cancer

Breast cancer is a malignant tumor that occurs in the breast epithelium tissue.In recent years,the morbidity of breast cancer has gradually increased and showed a younger trend,which is one of the most important diseases that threaten the health of women.The development of breast cancer is a complex biological process.At present,one of the key factors limiting the effectiveness of breast cancer treatment is that we are not completely sure about its molecular mechanism.Therefore,the in-depth study of molecular mechanism becomes an important method to improve the effectiveness of diagnosis and treatment.Traditional biology research mainly focuses on the characteristics and functions of single small molecules in cells.Although the method can reveal some life mechanism,it is undeniable that it ignores the biological characteristics of interaction between small molecules,and it can not explore the complex physiological activities of life.With the development of graph theory,network provides an important method for systemically studying the disease molecular mechanism.Weighted gene co-expression network analysis(WGCNA),based on the expression similarity among genes,is a common algorithm for network construction.It will be applied to gene expression data of breast cancer in this thesis.In this thesis,the gene expression data of breast cancer in TCGA database was used as the data source.First,in order to make differential analysis at the single gene level,we screened 9,376 differential genes associated with breast cancer development by using edgeR algorithm.Then we used WGCNA on normal samples and breast cancer samples to construct gene co-expression network,respectively.The algorithm only clusters the similarity of genes to one module.While the differential significance,biological processes and gene co-expression in each module are still unknown.Therefore,we focus on the differential analysis of modules.For each module,we conduct the differential analysis,including module difference connectivity analysis,module function analysis and differential gene correlation analysis,and find the hub genes and biological processes in each module.Finally,by analyzing the relationships between the drug signals and the hub genes,we obtain two candidate drugs astressin and protirelin.