Design of an extensible microvascular network information system
King-Ip (David) Lin* and Mohammad F. Kiani**
* Department of Mathematical Sciences, The University of Memphis, Campus Box 526429, Memphis, TN 38152-6429. ()
** School of Biomedical Engineering and Department of Radiation Oncology, University of Tennessee, Memphis, TN 38163.
Abstract
We propose an extensible Web-based microvascular network information system as part of the Microcirculation Physiome Project. We describe the design of the major components of the system, outline various technical challenges and propose solutions to them.
Introduction
The goal of the Physiome Project is to develop databases for physiological data, making them available via the Internet, and to build comprehensive models describing those physiological data sets [1]. Towards this goal, we propose the design of a microvascular network information system. We envision that the system will support the following:
· Efficient storage and management of microvascular network data. Users can retrieve information from the system via a user-friendly interface. Moreover, researchers can also deposit their data (e.g. results from their experiments) into the system.
· Capabilities of conducting pre-defined and user-defined mathematical modeling and simulation of network behavior, and comparing the findings to stored data.
Figure 1 outlines the system architecture. The WWW interface allows users at remote sites to access the system; the Geographical Information System (GIS) provides tools for network data collection and analysis [2].
Fig.1 Architecture of the Microvascular Network Information System
Database System
The database system provides tools for the storage and management of the microvascular network data. It interacts directly with, as well as providing data to, other parts of the system.
A challenge for the database system is to be able to handle the various types of data including alphanumeric data (e.g. vessel diameter), graphical data (e.g. topology) and complex data (e.g. adhesion molecules).
Traditional relational database systems (RDMBS) are inadequate for handling a wide variety of data types. Instead, we propose the use of new extensible object-relational database systems [3] to manage the data. These systems allow users to create their own data types (containing both data structures and operations on the structures) and incorporate them in the database directly. Users only need to define types like graphs, and provide the system with basic operators like “finding a loop” or “count the number of nodes”. The database system will then be able to treat these complex types as basic types. This allows users to write simple queries to retrieve data from the database directly instead of writing complex programs.
Mathematical Modeling
The simulation engine/mathematical modeling component of the database enables users to propose new models or use stored models to simulate various aspects of microvascular networks behavior. The microvascular network data can be supplied by the user or retrieved from the database. Simulation results can be stored in the database so that future users can query them without having to redo all the calculations.
The key challenge here is to design an engine that can simulate a wide variety of models. We are developing a component based and extensible simulation engine. The engine contains various components, each one conducting a specific task in the simulation. As new modeling techniques are introduced, the engine will be kept current by adding new components corresponding to the techniques.
Conclusion
We outlined the design of a microvascular network information system. We believe that the proposed system will be able to deliver the goals of the Physiome Project.
References
[1] J. B. Bassingthwaighte, “Design and Strategy for the Cardionome Project”, Analytical and Quantitative Cardiology, 430:325-339, 1997.
[2] N.M. Roth and M.F. Kiani, MF, “A ‘Geographic Information Systems’ based technique for the study of microvascular networks.” Annals of Biomedical Engineering, 27:42-47, 1999.
[3] M. Stonebraker and G. Kemnitz, “ The POSTGRES Next-Generation Database Management System”, Comm. of the ACM, 34(10), p. 78-92, October 1991.