An agent-based software simulation suite for modeling gene regulatory networks

Tuesday, October 09, 2012 — Poster Session I
1:00 p.m. – 3:00 p.m	Natcher Conference Center, Building 45	NIAID	BIOINFO-7

Authors

• R.M. Henderson
• D.E. Hurt
• V. Nagarajan
• M. Quiñones
• J. Lawson
• B. Peck
• R. Singh
• J. Hunsberger
• F. Chibane
• A. Elkahloun
• F. McMahon
• P. Munson
• Y. Demirkale
• M. Alda
• D. Chuang
• M. Hultner
• Y. Huyen

Abstract

​Existing genetic regulatory pathways analysis tools allow researchers to map their experimental data onto gene networks, but do not allow researchers to actively simulate dynamic perturbations to specific nodes. For example, existing tools allow users to see that repression of A activates B, which then activates C and D, and then activation of D represses E. However, there is no way to easily see the effect on E by the activation of A without manually walking through the network node by node. In this example, only five nodes are represented; the problem becomes increasingly complex, since the solution space increases exponentially as the number of network components increases linearly. We present a modeling and simulation suite, GRANITE, (Genetic Regulatory Analysis of Networks Investigational Tools Environment), an agent-based approach to modeling large, complex, and dynamic systems. We show that GRANITE is expressive enough to capture any kind of interaction network, can modularly use any kinetic model, is computationally tractable and scalable, and allows researchers to interact and dynamically perturb the system at different hierarchical levels to learn its rules for emergent behavior.

back to top