Automatic Generation of Drug Metabolic Pathways from ADME Ontology on OWL-DL

1
Automatic Generation of Drug Metabolic Pathways
from ADME Ontology on OWL-DL

• Konagaya Akihiko
• RIKEN Genomic Sciences Center
• Project Director
• Advanced Genome Information Technology Research
  Group

2
Motivation

• Coming of Personalized Genome Era
• Polymorphism in Drug Response Genes
• Detection of Drug-Drug Interaction

In silico prediction of individual differences
in drug response and drug-drug interactions on
multiple dose
3
Issues

• Detection of Personal Genome Variation
• Inference of Drug-Drug Interactions on Multiple
  Dose
• Quantitative Analysis by Drug Metabolic Pathway
  Simulation

4
What is a pathway?
5
Pathway
KEGG
6
Metabolic Pathway
http//www.expasy.org/cgi-bin/show_thumbnails.pl
7
Typical Pathway Analogy
8
Approaches for Metabolic Pathway Models

• Dynamic Approach
• Generated from primitive reactions depending on
  Trigger

• Static Approach

A Priori Defined
Body/Cell
Primitives
KEGG
http//www.genome.jp/kegg/
9
Why Dynamic Approach?

• Combinatorial Explosion of Molecular Pathways
• Integration of Continuants and Processes on
  Primitive Molecular Interactions
• Representation of Pathways as Aggregation of
  Primitive Molecular Events

10
How many colors can you see in rainbow?
11
Real Rainbow Color
From 360 nm400 nm
to 760 nm830 nm

• All the colors you can see with your own eyes!

12
Explicit Knowledge of Colors
Purple
Indigo
Blue
Green
Yellow
Orange
Red
13
Ontology for Rainbow Colors
RGB Value
Purple
800080
Indigo
000080
Blue
0000FF
Green
008000
Yellow
FFFF00
Orange
FF8000
Red
FF0000
14
Which are Purple?
800050
700080
600080
800080
800060
800070
500080
15
Color Representation by Primitives
RGB Representation
R 700nm, G 546.1nm, B 435.8nm.
Purple
Red
?
???
360nm
830nm
16
Ontology Schema
17
??????????????????(continuants)??????(process)??
?
?????????????????????????
Trigger (SN-38_at_lever)
Situated (Carboxyl esterase_at_lever)
Process (Irinotecan-SN38 Metabolism_at_lever)
Resultant (SN-38_at_lever)
18
???????????????????????
??????????????(??????????2006)
19
Prototype System
20
Controlled Vocabulary
21
Generated Pathway
22
Detected Drug Interactions
23
Ontology-driven Hypothetical Assertion
24
Conclusion

• Drug Interaction Ontology can be represented by
  OWL-DL in terms of processes, continuants and
  events.
• Drug metabolic pathways can be dynamically
  generated by the aggregation of primitive
  molecular events with OWL-DL and Prolog.
• Drug interaction can be detected by logical
  inference and mapped onto drug interaction
  ontology.

25
Future Works

• Expansion of Drug Interaction Ontology
• Automatic Generation of ADME models
• Integration of Drug Interaction Ontology and ADME
  simulation

26
Acknowledgement
Sumi Yoshikawa RIKEN GSC Ryuzo Azuma
RIKEN GSC Takeo Arikuma Tokyo Institute
of Technology Kentaro Watanabe Tokyo Institute
of Technology (Hitachi Ltd., Japan. ) Kazumi
Matsumura RIKEN GSC (DAIICHI PURE CHEMICALS CO.,
LTD., Japan. )
27
???????????????
Thank You for Listening