Abstract
Recently there have been increasing research interests in synthesizing gene networks and several studies have been done. The motivations of those studies are two ways. One is that the construction of gene networks having desired properties is a complementary approach to investigating mechanism and functions of gene networks. The other is that the construction of gene networks could be the first step to tackling the control problem of living cells. In this paper, we propose a synthesis method of gene networks by network learning. The problem is to determine parameters of a gene network such that it possesses given gene expression pattern sequences as desired properties. The proposed method is applicable to the synthesis problem of a wider class of gene networks. The method is general enough to be able to apply to several types of synthesis problems by slight extensions. In fact, the method is extended so as to solve several types of synthesis problems: the synthesis problem of gene networks which possess several desired expression pattern sequences such as multiple expression pattern sequences, a cyclic expression pattern sequence and a stable cyclic expression pattern sequence, and the synthesis problem of controller gene networks. Several numerical experiments illustrate the performance of the proposed method.
