Analysis And Application Of Time-Course Gene Expression Data

Along with the research and extensive applications of DNA biochip technology, gene expression data analysis has become a hotspot in life science field. As we know, DNA microarray technology is a very powerful tool to explore the mode of genes expressed in a cell. The main challenge is how to analyze the resulting large amounts of gene expression data. Recently, a new kind of gene expression data called time course gene expression data has received a lot of attention. Time course gene expression data is got by recording expression levels at different time points during a temporal cellular process. In present, there have been many clustering algorithms to analyze time course gene expression data, such as k-means clustering, hierarchical clustering, statistical model based clustering and so on. But in these methods, time course gene expression data is seen as a simple vector in multi-dimension space and self-related information between different time points in the data is ignored completely and so cannot influence the final clustering result effectively.This paper aims to explore novel time course gene expression data clustering algorithms and two autoregressive model (AR) based dynamic clustering methods are proposed. In this paper, some current clustering techniques and validation methods are reviewed and time course gene expression data is briefly introduced. The emphasis of this paper is as follows: Firstly, an improved AR model and Bayesian posterior probability based clustering method is proposed and the principle and way of how to use this method for clustering time course gene expression data are elaborated. Secondly, an AR model based fuzzy dynamic clustering algorithm is proposed and also the principle and way of how to use this method for clustering are elaborated. Aiming at utilizing only the class-conditional probability density or called likelihood in the original dynamic clustering method, an improved Bayesian posterior probability based clustering algorithm is proposed. And by integrating fuzzy theory, a fuzzy dynamic clustering algorithm is proposed, which adjusts the forecast process of AR model by fuzzy membership degrees and consequently overcomes the localization of the order p =1 of AR model in the original dynamic algorithm. At last, some discussion on time course gene expression model is presented based on regression techniques. And MATLAB is used to implement all the algorithms and experiment results on some datasets demonstrate the effectiveness, feasibility and advantage of the proposed methods over some present clustering methods.