Laboratories

Laboratory of Bioinformatics and Systems Biology (Department of Biological Sciences, Graduate School of Science) doesn't recruit students for the academic year 2018

Professor Toshihisa TAKAGI
Department of Biological Sciences, Graduate School of Science(Hongo)
E-mail: tt{at}bs.s.u-tokyo.ac.jp
Lab HP

Introduction

【Keyword】database, knowledge processing, text mining

Reconstruction of biological systems through knowledge processing technologies

 Understanding life as a system requires integrated analysis of not only genomic sequences and protein structures, but also various forms of data related to expression, localization, interactions, pathways, networks, and phenotypes, as well as the relationships between these concepts and their biological constraints and contexts. This will not be possible without reconstructing these life systems using computers so that their properties and behaviors can be examined. To this end, we are pursuing the following research topics.

1.Knowledge extraction from the literature
 Most established biological knowledge such as pathways and networks, as well as the experiments from which they are established, is represented as (predominantly English) text and diagrams in research papers and textbooks. We are therefore developing methods of efficiently extracting knowledge from this massive amount of data using natural language processing and information retrieval technologies. Specifically, we are developing technologies for automatic extraction of medical and biological concepts related to proteins, genes, compounds, and diseases, as well as the relationships among them. We are also developing new systems for supporting knowledge discovery by using extracted knowledge to re-interpret experimental data. Other research areas include developing methods for integrating the content of multiple academic papers, image and illustration search systems, and natural language query systems for biological or medical questions.

2.Biological system analysis from interactions
 Biological systems realize complex functions through interactions between biomolecules. It is therefore vital to compare the interactions and networks of biomolecules in multiple species,and to identify local structures particular to a given species so as to associate these features with the biological processes that they realize. Our lab develops the information technologies and systems that are required in these efforts.

3.Ontology construction
 Genome sequences are simple, in that they are indicated using combinations of only four characters:ATGC. Nonetheless, it is difficult to represent complex knowledge concerning these biological functions and mechanisms in computers without losing their fundamental essence. Analytical comparisons of multiple biological systems require unified methods of representing biological functions and mechanisms. Notational systems for representing biological functions and mechanisms are called ontologies, and our lab is working towards construction of ontologies related predominantly to pathways and networks.

4.Database integration technologies
 Understanding biological systems requires the unification of various types of data and knowledge, as well as intrinsic knowledge discovery. We have therefore joined database centers such as the DBCLS and NBDC to develop technologies for unified representation of complex biological information by means of knowledge representation technologies such as RDF. We are also working toward the unification of a wide variety of databases.

5.Methods for representing large-scale data
 These days, researchers are required to handle previously unimaginable amounts of data, so we are developing methods for data representation to aid the efficient interpretation of large-scale biological datasets. For example, omics data is often presented as huge, complex networks that defy intuitive interpretation, so we are working on network visualization tools that can dynamically and automatically extract information from such networks and present it in a more comprehensible format, similar to that of Google Maps.

 

Laboratories

The University of Tokyo
Graduate School of Frontier Sciences, The University of Tokyo

Page Top