Current Research

Structural Bioinformatics & Systems Biology

OUTLINE
Nowadays a number of genome sequences of several organisms are available. More than a thousand of proteins are usually coded in a genome and many structural-genomics projects are making a lot of effort to determine the structures of all proteins. On the other hand, there is significant progress in computational technology that enables us to challenge large-scale computer simulations. Therefore, the time is right to conduct bioinformatic research : dry biology relying on computers and theories. We address structural bioinformatics using biological databases and computer simulations. In this area, we are trying to elucidate protein functions based on protein structures and their motions. Proteins are very important functional molecules that comprise and maintain the body of a living organism. So far, our interest has been on the sequence-structure-function relationship of a protein. Currently, we are interested in higher-level functions emerging from the interactions of proteins.
TOPICS
(1) Protein structure prediction, de novo design, protein folding
A protein sequence folds into a unique three-dimensional structure and this structure in turn expresses a unique protein function. This means the protein structure can be deduced when the protein sequence was given. However, protein structure prediction is still very difficult. We are challenging this subject using the protein sequence-structure compatibility search and a protein structure database (fold recognition). To create a new protein with a desirable function, we have to design the functional shape of a protein. Using knowledge-based scoring functions, we are trying to establish a method to design protein sequences that fold into target templates with higher reliabilities. It is widely known that protein folding is accomplished very rapidly despite an enormously large protein conformational space. To investigate the mechanism of protein folding, we perform large-scale folding simulations on the super-computer and develop new methods to analyse a number of folding trajectories.
(2) Protein complex
A protein expresses its function via the interaction with the other molecules. In other words, proteins usually have the specific interaction surfaces. We investigate the appearances and disappearance of interaction surfaces by employing a special dataset : protein pairs whose sequences and structures are similar, but have different oligomeric states. We are also developing a new method to compare and classify the protein complexes and apply it to all protein complexes coded in the human genome. The results are deposited in the Structural Atlas of the Human Genome (SAHG) database.
(3) Interaction network, systems biology
Checking the protein complexes, we know protein A interacts with the protein B as well as protein C. Representing a protein as a node and an interaction as an edge, we can illustrate a small graph, B-A-C. By analysing the entire data of the protein interactions, we can depict a huge graph representing the protein-protein interaction network. We are investigating features of this graph in terms of protein domain architectures and functions. We are also exploring the sufficient and necessary conditions satisfied by an ideal minimum organism using genomic sequence data and a pathway database.
FUTURE WORK
A flow diagram of biological information processing exists : Protein sequences determine protein structures. Protein structures determine molecular functions. A set of molecular functions expresses cellular functions. A set of cellular functions specifies the tissue. Currently we are focusing on the level of molecular function. I hope we can enhance our research toward cellular functions in the future.