Discovery of Bidirectional Contiguous Column Coherent Bicluster in Time-Series Gene Expression Data.
Yun Xue,Zhihao Ma,Huixin Xu,Zhihao Lu,Xiaohui Hu,Chaoyi Pang
DOI: https://doi.org/10.1007/s13042-015-0464-0
2015-01-01
International Journal of Machine Learning and Cybernetics
Abstract:The application of high-throughput microarray has led to massive gene expression data, urging effective methodology for analysis. Biclustering comes out and serves as a useful tool, performing simultaneous clustering on rows and columns to find subsets of coherently expressed genes and conditions. Specially, in analysis of time–series gene expression data, it is meaningful to restrict biclusters to contiguous time points concerning coherent evolutions. In this paper, BCCC-Bicluster is proposed as an extension of CCC-Bicluster. An exact algorithm based on frequent sequential mining is proposed to find all maximal BCCC-Biclusters. The newly defined Frequent-Infrequent Tree-Array (FITA) is constructed to speed up the traversal process, with useful strategies originating from Apriori property to avoid redundant work. To make it more efficient, the bitwise operation XOR is applied to capture identical or opposite contiguous patterns between two rows. The algorithm is tested in simulated data, yeast microarray data and human microarray data. The experimental results show the proposed algorithm had better performance on the ability to recover the planted biclusters in the synthetic data than CCC-Biclusters and outperformed the one without FITA in speed and scalability. In the enrichment analysis, BCCC-Biclusters are proven to find more significant GO terms involved in biological processes than other three kinds of up-to-date biclusters.