Biclustering algorithms for gene expression in bioinformatics.

摘要

Several clustering methods had been proposed to analyze the gene expression data obtained from micro array experiments. However, the results which were obtained from several standard clustering algorithms were limited. The clustering algorithms were applied either to a row or column of data matrix which represent the gene and conditions of the gene respectively. For this reason algorithms which can be applied simultaneously both to row and column dimensions of data matrix have been proposed. The main goal is to find out submatrices, that is, subgroup of genes and subgroups of gene conditions where genes shows highly correlated activities for every condition. Such algorithms are referred to as biclustering algorithms. Biclustering algorithms are also called coclustering and direct clustering and have also been used in fields such as information retrieval and data mining. In this dissertation, first I am going to give a brief introduction to bioinformatics and data mining and then I am going to analyze several existing approaches for biclustering and classify them in accordance with the type of biclusters they can find, patterns of biclusters that are discovered, method used to perform the search, the approaches used to evaluate the solutions and the target applications. I am also going to discuss one of the data mining approach for biclustering the gene expression.