A novel one-way clustering based gene expression data biclustering method

Gene expression data biclustering is very important for the research on gene regulatory mechanisms. Especially biclustering has also been proved to be very useful to analyze data matrix other than gene expression data. Compared with the traditional clustering methods, bicluster detection is very different since the elements of one bicluster may be greatly distributed among the original data matrix. In this paper a novel one-way bicluster detection method is proposed. It makes use of the existing traditional clustering algorithms such as K-means as an intermediate tool to do data clustering. Based on the clustering results and a characteristic of bicluster, the biclusters are detected one by one. Furthermore an efficient submatrices and tables creation method is proposed to save the memory storage and accelerate the processing speed. At the end of the paper an experiment with the simulated data are presented.