Biotechnology Research Institute for Drug Discovery
breadcrumb navigationOrganization > Biotechnology Research Institute for Drug Discovery > Research Groups > Computational Functional Genomics Research Group


* Computational Functional Genomics Research Group joined the Intelligent Bioinformatics Research Team, Artificial Intelligence Research Center as of April 1, 2016.
Computational Functional Genomics Research Group

Computational Functional Genomics Research Group: In order to provide a proper grounding in drug discovery and efficient production of useful compounds, we are developing computational methods for privacy-preserving data mining (PPDM) focusing on genomic data, identification of disease factors using omics data, and analyzing protein function on a genomic scale.

Group's Research Theme

A statistically significant combination found in breast cancer data
A statistically significant combination
found in breast cancer data



Development of machine learning and statistical methods to analyze biological big data

Advancement of biological technology allows us to observe various kinds of genome-wide information such as genome sequences, methylations and gene expression profiles. It is expected that the combining these heterogeneous datasets and analyzing them would identify causal factors of diseases. However, current machine learning technique and statistical methods are not enough to solve the problems. To tackle this problem, our group has developed machine learning and statistical methods such as identification of statistically significant combinatorial effects [PNAS 2013], allele-specific expression measurement and statistical framework to find its changes [NAR 2014], and causal inference from gene expression profile. Our group also applies these developed techniques to actual biological problems through collaboration with medical and biological researchers to identify causal factor of diseases.

Active-site of catechol O-methyltransferase (EC 2.1.1.6)
Active-site of catechol O-
methyltransferase (EC 2.1.1.6)

Development and application of enzyme reaction database; EzCatDB

Based on the enzyme active sites and their catalytic mechanisms, an enzyme reaction database, EzCatDB, was developed, and additional systems related to EzCatDB have been developed as well. In addition to the enzymes, biomolecules such as cofactors, substrates, products and intermediates, which are involved in enzyme reaction, are also analyzed and classified in the database.
Moreover, the knowledge on the enzyme reactions and techniques of database analyses are also applied to the research on the genome and secondary metabolism from filamentous fungi. More recently, a biosynthetic gene cluster for a ribosomal peptide was discovered from Aspergillus flavus [Fungal Genet. Biol. 2014].

Development of privacy-preserving data mining technologies for personal genomes
 

Development of privacy-preserving data mining technologies for personal genomes

Privacy protection is a challenging problem for handling personal genomes. In our institute, we aim to develop new technologies that enable to utilize such sensitive information while keeping donors’ privacy. For example, one of our goal is to develop efficient methods to search on a private genomic database without leaking a query DNA sequence to the server.

Development of computational methods for proteome analysis
Mitochondrial dysfunction leads to
various diseases
(PDF: 345KB)

Development of computational methods for proteome analysis

For improving efficiency in searching candidates of drug targets and disease-related genes, we develop computational methods for proteome analysis, such as prediction of protein subcellular localization and post-translational proteolytic processing sites. Especially, we focus on mitochondria as a promising drug target because dysfunction of the organelle leads to various diseases. As a first step towards development of mitochondria proteome analysis methods, we developed MitoFates [Mol. Cell Proteomics. 2015], which predicts mitochondrial targeting signal sequences and their cleavage sites. MitoFates can identify promising candidate mitochondrial proteins with less false positives for experimental validation compared to the previous methods.

Page top

Research Achievements

ページの先頭へ

Staff Members

Position & Name Email Address & Web SiteWeb Site
Group Leader
Kentaro TOMII
Email AddressEmail Address
Web Sitehttp://cbrc3.cbrc.jp/~tomii/
Senior Researcher
Nozomi NAGANO
Email AddressEmail Address
Web Sitehttp://cbrc3.cbrc.jp/~nagano/index.html
Senior Researcher
Jun SESE
Email AddressEmail Address
Web Sitehttp://seselab.org
Senior Researcher
Kana SHIMIZU
Email AddressEmail Address
Web Sitehttp://cbrc3.cbrc.jp/~shimizu/
Researcher
Kenichiro IMAI
Email AddressEmail Address
Web Sitehttp://cbrc3.cbrc.jp/~imai/
AIST Postdoctoral Researcher
Mariko MORITA
Email AddressEmail Address
Web Sitehttp://cbrc3.cbrc.jp/~morita/
AIST Postdoctoral Researcher
Kyungtaek LIM
Email AddressEmail Address
AIST Postdoctoral Researcher
Yoshinori FUKASAWA
Email AddressEmail Address
AIST Postdoctoral Researcher
Raissa RELATOR
Email AddressEmail Address

ページの先頭へ