Skip Navigation

Results and Future Plans

Advanced Genome Information Technology Research Group
Promoting Health and Longevity through the Use of Grid-and Ontology-based Genomic Analysis Technologies

Akihiko Konagaya

Former Project Director of the Advanced Genome Information Technology Research Group

Professor Konagaya served as Project Director of the above-mentioned research group and team leader of the BioKnowledge Federation Research Team within the same research group.

How can we make use of genomics information to promote long life? This is a current theme central to the activities of the Advanced Genome Information Technology Research Group. To advance our research objectives, we employ a wide spectrum of methods ranging from high-density single nucleotide polymorphism (SNP) microarrays (for screening and analysis of human mutations), systematized models for ontology-based*1 studies of drug interactions, automatic generation of pharmacokinetics models based on ontological data on drug interactions, and high-performance pharmacokinetics simulations based on a Grid*2 system.
The Japan of today is one of the few countries in the world that can boast a long-living population, with more than 30,000 centenarians and women enjoying an average life expectancy exceeding 85. Despite this, however, it has been reported that more than 80% of Japanese aged over 100 have health problems in one form or another, which is an unflatteringly high percentage compared with healthy elderly people living in the United States or China. As one possible factor, the side effects of drugs have often been cited. Great variability of drug response exists among individuals, which may reflect genetic differences arising from multiple types of drug response genes (with CYP3A4 being a representative) as well as individual variations in liver and kidney functions. For elderly people who typically have diminished drug metabolic capacities, they tend to be more easily affected by drug interactions.

*1 Ontology
Ontology literally means the study of being or existence. Here, it refers to technologies associated with new terminology, which are used for specific concepts and contexts.

*2 Grid
A technology allowing the secure sharing of computers or experimental data among multiple research institutions.

Drug interactions between irinotecan (anticancer drug CPT11) and ketoconazole (antifungal agent) whose drug metabolic pathways were automatically inferred from the Drug Interaction Ontology

Ketoconazole and CYP3A4 have high binding affinity for each other. CYPO3A4 present in the liver can impede the metabolism of irinotecan. At the same time, both irinotecan and ketoconazole can bind with albumin within the blood stream. It was found that the two drugs may thus compete for binding with albumin.

Technologies for predicting drug metabolism

In the above contexts, we have undertaken research to find ways for making pre-treatment predictions on, for example, the effects of combined drug use or aging, to create models capable of integrating individuals' genetic information, physiological parameters and drug metabolic pathways. So far, as a step toward creating a prediction system of drug interactions for clinical use, we have developed a number of relevant technologies including a genetic variation analysis system (VARSearch), drug interaction ontology (DIO), and a pharmacokinetics simulator (PPDViewer).
The genetic variation analysis system (VARSearch) is being developed in collaboration with Toshio Kojima, Team Leader of the Computational and Experimental Systems Biology Group. Analyses for genotyping and gene copy number polymorphisms are conducted using this system, which involves SNP microarrays that can handle genetic data stored across 500,000 spots. The system is designed for rapid processing of terabytes of experimental data through its distributed parallel databases*3 and advanced user interface. SNP data in the order of 109 are visualized to allow efficient identification of sites of genetic variations.
Drug interaction ontology (DIO) has been developed as an ontology technology for systematized integration of information focused on intermolecular interactions to facilitate response studies related to drug response genes, drug actions and drug metabolites. Coupled with Semantic WEB*4 technology and rational inference algorithms, this system makes use of up-to-date drug information available on the Internet to create dynamic pharmacokinetics models, thereby identifying genes implicated in drug interactions during combined drug treatment.
The pharmacokinetics simulator (PPDViewer) can generate models for simulations of intermolecular interactions within cells, drug absorption and removal in organs, blood flow between organs, and drug excretion. Pharmacokinetics analyses are invariably dependent on many physiological parameters including polymorphisms of drug response genes, sex, height, body weight, and organ size. Using a defined set of such factors, we can produce simulations of pharmacokinetic parameters (such as plasma concentration of drug-time/area under curve [AUC]), providing quantitative information on the effects of drug interactions. As a vast amount of data is involved, the system exploits the use of Grid technologies to achieve the necessary computational power (figures above).

*3 Distributed parallel databases
A technology allowing rapid search functions via parallel access to replicated copies of databases in multiple computers.

*4 Semantic WEB technology
A technology allowing the sharing of Internet information rendered in a mode comprehensive (that is, ontology) to the computer.

Quantitative comparison between drug interactions predicted in pharmacokinetics simulations using Grid technology

The ADME simulation models obtained from the Drug Interaction Ontology enable quantitative evaluation of side effects caused by multiple drug administration and genetic variation of drug response genes. In the case of the irinotecan and ketoconazole co-administrated ADME model, the blood concentration of SN-38, the active form of irinotecan may increase up to 6% and 13% in maximum concentration and area under the blood concentration time curve (AUC), respectively. The model also indicates that severe side effects may be caused by the mutation of UGT1A1 which metabolizes SN-38 to its soluble form SN-38G by glucuronic acid conjugation; SN-38 is doubled when metabolic activity decreased by 30%.

Building a long-lived healthy society via sustained international cooperation

Through the research activities outlined above, our research group has played a pioneering role in advancing the applications of ontology and Grid technologies in the life sciences. Ontology research has been experiencing substantial growth in the context of active cooperation between Keio University, Osaka University and JCORE (a research base for ontology in Japan). It is hoped that through linkage with international players such as NCORE (US ontology research base) and ECORE (European ontology research base), JCORE will become an important source of innovation for the global development of ontology.
Since 2004, the Life Sciences Grid Workshop has been held annually as an international event. Conference proceedings are compiled and published in special issues of international journals or in book volumes. Large-scale data processing and large-scale computation are the fundamental elements of bioinformatics, which necessitates rapid sharing of the computational environment between research bases supported by life sciences-focused Grid research. In this regard, our group will continue to actively develop and promote technologies that advance the goals of research collaboration, which is dependent on secure sharing of experimental data and scientific knowledge.
To realize the dream of a happy healthy society blessed with longevity, it has become increasingly obvious that we must make intelligent use of information on individuals' genetic variations. Accordingly, new technologies for personalized genetics analyses need to be developed to facilitate the processes of integration/systematization and modeling of relevant information ranging from genomics data to physiological parameters. As a pioneering team for the development of ontology, and modeling and simulating technologies, we are strongly motivated by our mission to advance genomic sciences and promote the growth of the genomics industry.