Attractor Equivalence: An Observational Semantics for Reaction Networks

ABSTRACT. We study observational semantics for networks of chemical reactions as used in systems biology. Reaction networks without kinetic information, as we consider, can be identified with Petri nets. We present a new observational semantics for reaction networks that we call the attractor equivalence. The main idea of the attractor equivalence is to observe reachable attractors and reachability of an attractor divergence in all possible contexts. The attractor equivalence can support powerful simplifications for reaction networks as we illustrate at the example of the Tet-On system. Alternative semantics based on bisimulations or traces, in contrast, do not support all needed simplifications.