Chinese Journal of Computers   Full Text
Title	Mining Synchronous and Asynchronous Co-Regulated Gene Clusters from Time Series Microarray Data
Authors	YIN Ying ZHAO Yu-Hai ZHANG Bin WANG Guo-Ren
Address	(College of Information Science and Engineering, Northeastern University, Shengyang 110004)
Year	2007
Issue	No.8(1302—1314)
Abstract &   Background	Abstract Gene co-regulation falls into two major categories, i.e. synchronous and asynchoronous co-regulation. This paper proposes a new model Reg-Cluster, which groups synchronous and asynchoronous co-regulated genes together if they have the same code. Further, an effective clustering algorithm with several efficient pruning rules, namely FBLD, is designed to identify all maximal Reg-Clusters in a "First Breadth-first and Last Depth-first" manner. The resultant clusters contain the detailed and complete co-regulation information, which facilitates the study of genetic regulatory networks. Moreover, the method can be extended to the analysis of 3D gene-sample-time microarray data. The FBLD algorithm has been implemented on both real and synthetic datasets and the results from the real dataset has been submitted to Gene Ontology. Experimental results prove the effectiveness and efficiency of the proposed method. keywords synchronous/asynchronous co-regulation; activation/inhibition co-regulation; clustering; time series; gene ontology background The complexity of a biological system provides a great diversity of correlations among genes/gene clusters, including synchronous and asynchronous co-regulations, each of which can be further divided into two categories: Activation and inhibition. Most existing methods can only identify the synchronous activation patterns, such as shifting, scaling and shifting-and-scaling, however, few focus on capturing both synchronous and asynchronous co-regulations. This paper focuses on identifying synchronous and asynchronous co-regulation patterns simultaneously. Furthermore, the detailed and complete co-regulation information including synchronous/asynchronous activation co-regulation, inhibition co-regulation and the number of time-lag points in genes/gene clusters, which facilitates the study of genetic regulatory networks, can be easily derived from the resulting clusters analysis. This research is supported by National Natural Science Foundation of China (60573089) and the National Key Technologies Research and Development Programming (2004BA721A05). One mission of all these projects is to mine interesting patterns of significant meaning from bio/medical data. The research work of this paper is encouraged by this background and is considered as a significant part of pattern discovery. Bioinformatics is a research direction of data mining group. What they are interested in includes clustering, classification, and association analysis on bio/medical data. Several related papers have been published or accepted by some journals or international conferences.