MOTIVATION : Boolean network models are suitable to simulate gene
regulatory networks (GRNs) in the absence of detailed kinetic information.
However, reducing the biological reality implies making
assumptions on how genes interact (interaction rules) and how their
state is updated during the simulation (update scheme). The exact
choice of the assumptions largely determines the outcome of the
simulations. In most cases, however, the biologically correct assumptions
are unknown. An ideal simulation thus implies testing different
rules and schemes to determine those that best capture an observed
biological phenomenon. This is not trivial, since most current methods
to simulate Boolean network models of GRNs and to compute their
attractors impose specific assumptions that cannot be easily altered,
as they are built into the system.
Results : To allow for a more flexible simulation framework, we
developed ASP-G. We show the correctness of ASP-G in simulating
Boolean network models and obtaining attractors under different
assumptions by successfully recapitulating the detection of attractors
of previously published studies. We also provide an example of how
performing simulation of network models under different settings help
determine the assumptions under which a certain conclusion holds.
The main added value of ASP-G is in its modularity and declarativity,
making it more flexible and less error-prone than traditional approaches.
The declarative nature of ASP-G comes at the expense of
being slower than the more dedicated systems but still achieves a
good efficiency w.r.t. computational time.