A common-subsequence-based approach for mining deep order preserving submatrix

Abstract

As an effective biclustering model, order-preserving submatrix (OPSM) has been widely applied to biological gene expression data mining. Recently, biologists hope to find deep OPSMs with long patterns and comparatively few support rows, which are not only useful on the interpretation of gene regulatory networks but also have essential biological significance. Unfortunately, the traditional exact mining algorithms based on Apriori principle could not deal with the deep OPSM problem, since they often take a large minimum support threshold for pattern pruning, and inevitably miss some significant deep OPSMs. Therefore, this paper proposes a new exact algorithm for mining deep OPSMs, which obtain all the deep OPSMs by finding the common subsequences shared by every two rows. Experiments have been done in both real and synthetic data sets, and the results show that our algorithm is suitable for the full mining of deep OPSMs with a small support, which could even find all the deep OPSMs with a minimum support threshold of 2. Compared with the traditional sequential pattern mining algorithms which depend on relatively large support threshold, our algorithm is an effective one to solve the deep OPSM problem.