ホーム

MEMBERS | Program Members

Osamu Maruyama

MEMBERS

Message Download Consortium

MEMBERS

Program Members

Computational Biology

Osamu Maruyama
Degree: Doctor of Science (Kyushu University)
Research Interests: Algorithm and Bioinformatics
Unit:Number & Equation

Report

Figure1

Today, it seems that the connection between industry and science fields related to biology and life systems is getting stronger. Applications of life science research can be found in a variety of industrial and business fields, including drug discovery, medical services, functional foods, microorganism-derived material generation, biomass fuel, authentication technologies such as DNA identification, and protein engineering.

We are interested in developing efficient methods to discover biological knowledge from various existing knowledge and databases, and applying them to real biological data. One of the problems we have addressed extensively is the problem of finding sequence motifs, in which we are given a set of biological (DNA or amino acid) sequences and asked to find a short common pattern matching each of the given sequences. DNA sequences can be considered to be an encrypted blueprint of life and a protein is a molecular machine with a particular function, which is generated from the blueprint. Therefore, it is important to analyze sequences and understand their functions in gaining an insight into life systems.

Proteins usually function by interacting with each other (See Fig. 1). Thus, it is quite helpful to understand the mechanism of protein-protein interactions in gaining an insight into life systems and developing applications, including drug design. For this issue, we have presented a method to infer interacting sites and regions closely related to interactions from the sequences of proteins known to interact with a particular protein by finding sequence motifs from the sequences. As a result, we have succeeded in identifying reliable candidates of interacting sites from various proteins.

Figure 2

Gene expression is the process in which proteins are synthesized form genetic information. The first step of the process is “transcription”, in which particular region of DNA sequences are copied as messenger RNA (mRNA) (See Fig. 2). Transcription is “turned on” when transcription factors specific to a gene, which are complexes of proteins, bind to specific sites, called transcription factor binding sites, in the upstream regions of the gene. Thus, it is important to elucidate all the transcription factor binding sites of a gene for each of the transcription factors regulating the gene in understanding the structure of gene regulatory networks. For the problem of finding transcription factor binding sites, conventional computational methods tries to find sequence motifs common to the sequences of genes that are inferred to be co-regulated by common transcription factors from results of microarray experiments. Sequence motifs are often evaluated from statistical point of view. In recent years, with the advent of sequenced genomes of various species, “phylogenetic footprinting” has received much attention. It is a technique for finding more conserved regions than the surrounding sequences by comparing orthologous sequences of closely related species. We have proposed a novel phylogenetic footprinting method. One of the features of the method is that it uses two heterogeneous evaluating functions for sequence motifs. One is a maximum parsimony score and the other is a log-likelihood ratio based on an evolutionary probabilistic model on biding sites. In a performance comparison with existing other tools, our proposed method shows better qualities of found sequence motifs.

RETURN LIST