Poster
Presentation 20:
Intelligent Multi-Agent System for Information Management
and Interpretation in Biotechnology Applications
Dennis J. Michaud1, Vikas
Agrawal2, Prasad S. Dhurjati1, and Keith S. Decker2
1Department of Chemical Engineering
2Dept. of Computer and Information Sciences
University of Delaware
Newark, DE 19716
michaud@che.udel.edu
(302) 831-6851
We are currently building an integrated software system
for heterogeneous database retrieval, data analysis, and visualization
that will help a biologist or engineer interpret and exploit the exponential
explosion of biological information available in on-line databases.
The system's framework consists of intelligent agents that coordinate the
gathering, filtering, and processing of information from many different
on-line databases, convert this information into useful knowledge, and
present the results to the user in an organized manner. Through the
use of query interpretation agents, the system shall be able to intelligently
select the necessary on-line or local data sources relevant to the query.
Database "wrapper" agents designed to transparently access this data have
been developed for a number of websites and others can easily be generated
by the user. The analysis tools necessary in converting this raw
biological data to valuable hypothesis are to be integrated within the
agent system. These tools shall eventually include comparative genome
mapping, analysis of gene expression and proteomic experiments, and dynamic
modeling of metabolic pathways. Visualization agents are also being
developed to present the query results in a concise and informative manner.
|