Java Adaptive Mathematical Modeling Engine (JAMME)

1
Java Adaptive Mathematical Modeling Engine(JAMME)

• Leeland Artra, Cell Systems Initiative (CSI)
• Zheng Li, Department of Bioengineering
• University of Washington, Seattle

2
Introduction of JAMME

• Proposed collaborative project between NSR
  (National Simulation Resource for Circulatory
  Mass Transport and Exchange NIH/NCRR) and CSI to
  provide a complete software database
  environment for real time biological modeling.
• Integration of NSRs JSIM model simulation
  environment with CSIs Moulage System framework,
  Metagraph data representation model, and a new
  object database management system.

3
Background (available software at NSR and CSI)

• JSIM is an integrative model simulation
  environment being developed at NSR.
• Moulage is a real time data sharing analysis
  system framework.
• Metagraph is a code name for a data and process
  representation language used to describe
  structures and methods that capture a
  computerized abstraction of the scientific
  process.
• The new object database management system being
  developed at CSI. Its primary goal is to provide
  a dynamic data engine that can handle complex
  hyper-graph / XML / Metagraph queries directly.

4
JSIM Design Goals

• Easy-to-use end-user interface. Low barriers to
  entry of math modelling for biologically-inclined
  users.
• Facilitate collaborative model development
  (integration)
• Integration at mathematical level.
• Integration at component level.
  (Plug-and-play).
• Connectivity to heterogeneous data sources.
• Facilitate collaborative software development.
• Facilitate distributed-computing.
• Flexibility and Extensibility at all levels
• math model classes and representations
• hierarchical component libraries
• numeric libraries

5
JSIM Architecture
6
Items Needed By JSIM

• Persistence Engine
• Intelligent Distributed Analysis Engine
• Visual Programming Ability
• Ability to Import Various Modeling Data
• Ability to Directly Interface With Experiment Data

7
Moulage Design Goals

• Create a development framework that automatically
  can provide
• Transparent access
• Ability to accept multiple data sources
• Easy sharing of data
• Data in form usable by object-oriented databases
• Multiple tools for analysis
• Dynamic version distribution

8
Moulage Base Components

• Inspector analysis, application
• DataObject data
• IOController IO operations
• DataObjectManager database
• Translator file to/from object

Translator
textual data
DataObjectManager
binary data
0111010101
0101011...
DataObject
remote data
4
3
2
1
Analysis / Application (Inspector)
9
JAMME Approach

• 100 Core Java
• Any needed classes built directly from on the 1.2
  core
• Carefully engineered to work within the Java
  Virtual Machine constraints
• Allow operation as
• Stand alone application
• Command line utility (batch processing)
• Applet (Web and Email integration)

10
Integration Overview

• Start with JSIM

JSIM
11
Integration Overview - Moulage

• Enhance by adapting to Moulage

Analysis
JSIM
MML Data
IO
12
Integration Overview - Metagraph

• Enhance by creating Metagraph Data Structures

Analysis
JSIM
MML Data
IO
13
Integration Overview - ODB

• Enhance by adding new object database

Analysis
JSIM
IO
IO
14
Integration Overview - ODBMS

• Enhance by allowing distributed database
  processing of analysis

Analysis
JSIM
IO
IO
ODBMS
Analysis
Analysis
Analysis
15
Integration Overview - Result

• Resulting in a set of reusable components

JAMME
Analysis
T-Cell Modeling
IO
IO
ODBMS
Analysis
Analysis
Analysis
Query Engine X