Analyzing gene behaviors with genetic computation

Computational methods have become very important in analyzing biosequence data. However, most of the current computer-assisted methods are limited to finding motifs only. In real life, genes can be regulated in many ways, involving various interactions among regulatory elements. In this paper we proposed a new framework for analyzing gene behaviors in terms of the correlation between motifs and gene expressions. This framework consists of motif-finding algorithms, constructive induction and standard inductive learners. It complements the current development of microarray and genechip technology by providing a deeper insight of gene behaviors. A genome-wide gene expression analysis demonstrated the value of the studies of this new framework.