AIHARA Kazuyuki The University of Tokyo, Institute of Industrial Science, Professor, 生産技術研究所, 教授 (40307218)
ICHINOSE Natsuhiro Kyoto University, Graduate School of Informatics, Reader, 情報学研究科, 助手 (70302750)
CHEN Lounan Osaka Sangyo University, Faculty of Engineering, Associate Professor, 工学部, 助教授 (20298813)
伊庭 斉志 東京大学, 大学院・新領域創成科学研究科, 助教授 (40302773)
|Budget Amount *help
¥72,900,000 (Direct Cost: ¥72,900,000)
Fiscal Year 2004: ¥16,000,000 (Direct Cost: ¥16,000,000)
Fiscal Year 2003: ¥16,000,000 (Direct Cost: ¥16,000,000)
Fiscal Year 2002: ¥15,400,000 (Direct Cost: ¥15,400,000)
Fiscal Year 2001: ¥25,500,000 (Direct Cost: ¥25,500,000)
In this research, we have presented an application of genetic algorithms to the gene network inference problem. It is one of the active topics in recent Bioinformatics. The objective is to predict a regulating network structure of the interacting genes from observed outcome, i.e., expression pattern. The task consists of modeling the rules of regulation and inferring the network structure from observed data. The GA is applied to train the model with observed data to predict the regulatory pathways, represented as influence matrix. We have implemented a reverse engineering method based on genetic algorithms in a quantitative and linear biological framework. The merit of this approach is that it can be applied with small amount of data, optimize large amount of parameters simultaneously and can be applied on nonlinear models. The GA implementation includes multiple stage evolution and matrix chromosomes. This method has been applied on simulated expression patterns and experimentally obs
erved expression patterns. In this research, we used the knowledge of designing electric circuit by GA.
As for another important topic, we have proposed a dynamic differential Bayesian networks (DDBNs) and nonparametric regression model. This model is an extended model of traditional dynamic Bayesian networks (DBNs), which can incorporate temporal information in a natural way and directly handle real-valued data obtained from microarrays without any transformation. In addition, it can cope with differential information between gene expression levels, without any loss to the traditional advantage, i.e., the capability of estimating non-linear relationships between genes. We have applied DDBNs to analyze simulated data and real data, i.e., Saccharomyces cerevisiae cell cycle gene expression data. We have confirmed the effectiveness of our approach in the sense that some edges have been successfully detected only by DDBNs, not by DBNs.
In recent years, base sequences have been increasingly unscrambled through attempts represented by the human genome project. Accordingly, the estimation of the genetic network has been accelerated. However, no definitive method has become available for drawing a large effective graph. To solve these difficulties, we have proposed a method which allows for coping with an increase in the number of nodes by laying out genes on planes of several layers and then overlapping these planes. This layout involves an optimization problem which requires maximizing the fitness function. To demonstrate the effectiveness of our approach, we show some graphs using actual data on 82 genes, 552 genes, and artificial data modeled from a scale-free network of 1,000 genes. We also described how to lay out nodes by means of stochastic searches, e.g., stochastic hill-climbing and simulating annealing methods. The experimental results have shown the superiority and usefulness of stochastic searches in comparison with the simple random search. Less