The Study On Models Building And Structure Analysis Of Gene Regulatory Networks

Bioinformatics is a new subject on processing and analysis life information. In recent years, modeling gene regulatory networks become one of the effective means of systems biology research. In the highly connected cells environment, in order to get the gene regulatory networks, we make the genome as a whole organization structure, study function and behavior of gene from the systematic point of view. As the advances in information technology and computational science, gene regulatory network has been extensively studied in the last decade. On the basis of understanding the functions of cell metabolism in detail, it plays a significant role in exploring the mechanism of life activities and causes of disease.This paper is mainly aimed at several issues based on structure building and analysis of gene regultory networks. Including: proving the ordering of Boolean network in random states, sutructure analysis of probability Boolean gene regulatory network basing on the parameters, building fuzzy Boolean gene regulatory network in intelligent methods. Besides the study on construction and deconstruction on gene regulatory networks, a new hybrid clustering algorithm for analyzing time-course gene expression. To be concrete, the contributions of this dissertation are as follows:1. Combined with the small-world network theory,take a method for randomized Boolean networks, Based on the theory of complex networks,the classical random sampling random Boolean networks and simulation Prove the model as a Boolean gene regulatory network topology of the network, and presents a simple way to build a Boolean network.2. In the second part, a new method is presented for decomposing the structure of gene regulatory networks, by taking the probability Boolean network (PBN) as a typical model. Basic steps as follows: first, according to the strcuture of probability Booloean networks, search for key genes in the abstract network with nodes and connections; second, starting from these key nodes(genes), calculating mutual information between geneon the shortest path with maximum average mutual information passing can be found. Each ordinary node can be distinguished which sub-network it belongs to. Ultimately, the choice of function parameters and scale of sub-network depends on the biological consideration.3. A new gene regulatory network model via the fuzzy logic is proposed. Conmbined with the research on the structure of gene regulatory networks and obtained experience, judging genes expression level on the fuzzy rule, to contribute a new model of gene regulatory network.4. A novel fuzzy clustering algorithm is proposed for analysing time-course gene expression data. Compared with conventional hard partition clustering algorithms, fuzzy clustering algorithms are robust to the scaling transformation of a dataset. However, they cannot make full use of the important dynamic information in time-course gene expression data. Accordingly, autoregressive (AR) model can be introduced into fuzzy clustering algorithm; we explore the proposed clustering algorithm DFC, which can analyze a time-course gene expression data as a set of time series dynamically. In this way, the important dynamic information in time-course gene expression data is used adequately. And the forecast processes in AR model is adjusted using the corresponding membership functions, such that better clustering results for time-course gene expression data is obtained.