Modeling and Analysis of Gene Regulatory Networks - Modeling in Computational Biology and Biomedicine - page 74

2.4 Outlook

A large amount of data on gene regulation is currently available. This created an
upsurge of interest in modeling regulatory networks. Here, three usual approaches
to modeling were presented: continuous differential systems (based on sigmoidal
functions), piecewise affine systems (based on step functions), and discrete systems.
On the one hand, continuous systems present a higher level of details and offer the
ability to model dynamics. But on the other hand, discrete (or logical) systems are
easier to analyze, need a smaller amount of data, can be deduced from qualitative
observations, and thus allow modeling of larger systems. The choice then depends
on the nature of input data and of the biological question under consideration.
Furthermore, the well understood relationships between the continuous and discrete
approaches presented here, allow one to follow a classical and simple strategy to
model a new biological system: the discrete approach can be taken as a useful first
step as long as the input data are qualitative, then more accurate descriptions can be
achieved through continuous models based on the discrete ones when more precise
input data are available.
Research on gene regulatory networks is rather active, and many research
directions are relevant. From a computational point of view, it becomes crucial to
develop techniques that allow the modeling of large systems, for instance using
sensible model reductions or modular decompositions. Another direction consists
in using experimental design approaches to select sets of experiments that are
efficient to validate or to refute a model. From a biological point of view, a
number of outstanding questions are open. The stochasticity and robustness of
regulatory networks are not well understood. The dynamical influence of network
architectures, and the evolutionary processes that produce them, are also far from
being understood. Furthermore, models for the interplay between gene network and
other processes, such as metabolism and cell signaling, have to be developed.

2.5 Online Resources

Several computer tools are available to help modeling and analyze genetic regula-
tory networks. A few examples are:
GNA (Genetic Network Analyzer)
http://www.genostar.com/en/genostar-software/gnasim.html
Modeling and simulation of GRN, using piecewise linear models.The user specifies
the equations, the parameters (synthesis and degradation rates, thresholds), and
inequality constraints between them.
GINsim (Gene Interaction Network simulation)
http://gin.univ-mrs.fr
Modeling and simulation of GRN, based on a discrete, logical formalism. The
user may specify a model of a GRN in terms of asynchronous, multivalued logical
functions.