Database of Dynamic Signatures Generated
by Regulatory Networks (DSGRN)Bree Cummins^1 , Tomas Gedeon1(B), Shaun Harker^2 ,
and Konstantin Mischaikow^2(^1) Department of Mathematical Sciences,
Montana State University, Bozeman, MT 59715, USA
[email protected]
(^2) Department of Mathematics,
Rutgers, The State University of New Jersey, New Brunswick, USA
Abstract.We present a computational tool DSGRN for exploring net-
work dynamics across the global parameter space for switching model
representations of regulatory networks. This tool provides a finite par-
tition of parameter space such that for each region in this partition a
global description of the dynamical behavior of a network is given via a
directed acyclic graph called a Morse graph. Using this method, para-
meter regimes or entire networks may be rejected as viable models for
representing the underlying regulatory mechanisms.
1 Introduction
An important challenge in systems biology today is the lack of robust tools that
can translate static network information into actionable information about a
network’s dynamics. It is the dynamics of the network that ultimately correlates
with the cellular state and determines its phenotype. Lack of understanding of
all potential dynamics that a network structure supports is one of the reasons
for the apparent lack of correspondence between genotype and phenotype.
We have developed an efficient mathematically rigorous computational tool-
box, called Dynamic Signatures Generated by Regulatory Networks (DSGRN),
that computes the range of dynamic behaviors supported by a given network.
DSGRN is based on a new mathematical framework for nonlinear dynamics,
which moves away from consideration of individual solutions at particular para-
meter values. In networks with 5–10 nodes with 30–50 parameters any random
sampling of parameters and initial conditions will only cover a negligible portion
of possible dynamical behaviors. Furthermore, comparing individual solutions to
experimental data, which typically carries significant uncertainty, cannot be used
to reject potential models because many nearby parameters and initial condi-
tions will produce solutions that fit the data equally well. The main applications
of our tool have been to(1)describe coarse dynamics for the entire parameter
space for a given network, allowing exploration and quantification of different
dynamic signatures supported by the network architecture, and(2)compare
©cSpringer International Publishing AG 2017
J. Feret and H. Koeppl (Eds.): CMSB 2017, LNBI 10545, pp. 300–308, 2017.
DOI: 10.1007/978-3-319-67471-1 19