Construction of Genetic Regulatory Network with Computational Life

为构建具有计算寿命的遗传调控网络，开展了以下系统和算法的开发。一种探索基因序列之间共同保守区域的系统。给定一组功能相同或相似的基因序列，我们开发了一种系统来识别序列之间保守的序列模式。我们的系统通过以下步骤来检测这种模式：构造一个图，其顶点是序列的每个可能的固定长度区域，并且如果任何两个区域的相似度超过给定的截止值，则绘制该图的边；并在图中提取最大密度子图。通过检测已知的保守区域，如大肠杆菌双组分系统、转录调控基因家族和头部休克蛋白，证明了本系统的有效性。一种使用基因表达谱的基因聚类系统。DNA芯片技术已被广泛应用于基因表达分析。然而，实验噪声阻碍了表达数据的有效利用。针对这一问题，我们通过傅立叶变换和小波变换将时间序列表达数据变换到频域，并去除信号中的高频成分，从而开发了一种抑制此类噪声的系统。代谢途径的多重比对算法。代谢途径被认为是最广为人知的分子网络之一。我们开发了一种算法，通过比较EC数表示的反应序列，找到几条路径之间共同保守的反应模式。通过发现糖代谢途径和氨基酸生物合成途径之间的保守反应，证明了算法的有效性。

Development of the following systems and algorithms have been carried out for the construction of genetic regulatory network with computational life.1. A system for exploring commonly conserved regions between gene sequences. Given a family of gene sequences of which function is the same as or similar to each other, we have developed a system for identifying sequence patterns that are conserved between the sequences. Our system detects such patterns by : construct a graph whose vertices are every possible fixed-length regions of the sequences, and whose edges are drawn if any two of regions have similarity more than a given cutoff ; and extract a maximum density subgraph in the graph. The effectiveness of our system is demonstrated by examining known conserved regions, such as Escherichia coli two-component systems, transcription regulation gene family, and head shock proteins.2. A clustering system of genes using gene expression profile. DNA chip technology has been widely used for gene expression analyses. The experimental noises, however, hindered the effective use of the expression data. To cope with this issue, we have developed a system for reducing such noises by transforming time-series expression data into the frequency domain by the Fourier and the wavelet transformations and removing the high-frequency components of the signals.3. A multiple alignment algorithm of metabolic pathways. Metabolic pathways have been recognized as one of the most well-known molecular networks. We have developed an algorithm for finding commonly-conserved reaction patterns between several pathways by comparing sequences of reactions represented by the EC numbers. The effectiveness of our algorithms is demonstrated by finding conserved reactions between sugar metabolism pathways and those between amino acid biosynthesis pathways.