Dissertation > Medicine, health > Basic Medical > Human biochemistry, molecular biology

Research on Disease Related miRNA Mining Based on Protein-Protein Interaction Network

Author ZuoMingXiang
Tutor WangYaDong
School Harbin Institute of Technology
Course Computer Science and Technology
Keywords miRNA disease protein-protein interation network topological parameter
CLC R341
Type Master's thesis
Year 2008
Downloads 187
Quotes 0
Download Dissertation

miRNA plays an important role of regulator in embryo development and disease process in human body. With the miRNA research outspreading, the biological data on miRNA is increasing rapidly. As a result of that, based on finding links among biological data, bioinformatics researchers are beginning to investigate the relations between miRNA and disease from a perspective of full data, to fill the shortcomings of traditional experimental method. Investigation on disease related miRNA mining will show significant effect on processes of disease diagnostics and therapy, drug target detecting, and drug screening.The method of disease related miRNA mining based on protein-protein interaction network (PPIN) is the main research interest of this paper. Firstly, an analysis of the characteristics of the biological data on miRNA and disease is done. Thus, we chose a series of effective biological data as the basis of this paper. Then, based on chosen biological data, we define and construct disease-associated protein-protein interaction networks, and make a statistical analysis of topological parameters (including degree, cluster coefficient, shortest path number) on the networks as well. Thirdly, we establish the links between miRNA and PPIN through the characteristics of miRNA targeting data. So after the mathematical model of Bayesian posterior probability was introduced into the analysis of topological parameters on disease-associated PPIN, we finally achieve the goal of disease related miRNA prediction.Besides, this paper successfully makes a prediction on breast cancer related miRNA prediction, and does 4-fold cross-validation several times to validate the effectivity of our algorithms. As the result shows the forecast accuracy of our algorithms achieve about 32% on the basis of 0.32 of Youden’s index, the validity of the algorithms of disease related miRNA mining based on PPIN in this paper is strongly demonstrated.

Related Dissertations
More Dissertations