MSPattern: Efficient mining maximal subspace differential co-expression patterns in microarray datasets

Abstract

Traditional methods for microarray datasets analysis often find the co-expression genes. However, these methods may miss the genes which are differential co-expression patters under different datasets. Mining these differential co-expression patterns is more valuable for inferring regulator. In this paper, we develop an algorithm, MSPattern, to mine maximal subspace differential co-expression patterns. MSPattern constructs a weighted undirected gene-gene relational graph firstly. Then all the maximal subspace co-expression patterns would be mined by using gene-growth method in above graph. MSPattern also utilizes several techniques for generate maximal patterns without candidate SDC patterns maintenance. Evaluated by the gene expression datasets, the experimental results show our algorithm is more efficiently than traditional ones.