21-06-2012, 05:50 PM
The Design of Agent-Based Tools for Bio-Informatics
The Design of Agent.pdf (Size: 231.91 KB / Downloads: 32)
Introduction
Information technology has always played an important
role in the development of genetics and molecular
biology. Information technologies helped to develop
many useful algorithms and programs, such as
GCG [3] and BLAST [l], and some very large sequence
databases, such as GenBank [6], TIGR [8],
and SwissProt [2], and tools for data management and
database search. As the interpretation of sequence
data to obtain biological signals and sequence characteristics
has become cenpal issue .of bio-informatics
research [7][9][10], and as it was estimated that a single
cell type has 15000 to 20000 expressed genes and
there are about 100,000 genes in human beings, it is
evident that information technologies will play an essential
role in future biological and genetic research.
An Integrated Web-Based Tool
We have designed an integrated and web-based tool
for processing database serach results. The detailed
workflow of database search tool is shown in Figure 1.
There are also two main pages in this tool. The first
page submits the query sequence to a database to
search for matches. The user may choose to query
a local’database or a public database. This page may
be invoked directly from a browser session. The second
page handles the search results. It lets user view
and annotate the search results. It also generates a report
and stores the search resluts and annotation into
the local database.
Agent-Based Approach
Currently, most workflow models or process models
are based on graph concepts, such as Petri nets, state
machines, or flow graphs [4][5]. Such graph-based
models are static in nature, even though that tasks
or actions can be enacted and created dynamically.
Thus, we propose to extend a workflow model with
software agents to better deal with the dynamic and
evolutionary part of real-world environment. Since
software agents are more intelligent and can be customized,
they are more flexible than tasks. Also, as
Internet agents can be mobile and dynamically downloaded,
can coordinate various activities, and can communicate
with other agents, they increase the expressiveness
and modeling power of a workflow model.
Conclusion
We have presented the design and implementation
of web-based tools for bio-informatics. The design and
implementation of the tools was first based on workflow
approach. As biological research and discovery is
a continuous and evolutionary process, pure workflow
approach is not sufficient in such environment. We
proposed to extend a workflow model with software
agents, where some tasks or actions can be software
agents. Agents can increase the expressiveness and
modeling power of a pure workflow model. However,
since we do not change the topology of a workflow
model, the processing steps remain unchanged. As
agent technology becomes more mature, it is expected
that agent technology will play a more important in
various software system design.