Far be it for me to tell you, o esteemed user, how to create a model of your system of interest. You probably already have several models you'd like to try out. All that's necessary is that the model be expressed in terms of a set of differential equations, and that you specify the initial conditions. There are nevertheless some points which should be borne in mind:
Firstly, if your model involves terms with more complicated kinetics, which replace several explicit terms, you might consider using the full set instead. For example, enzyme catalyzed reactions can under certain conditions be described by the Michaelis-Menten approximation, greatly simplifying the reaction set description. The conditions for applicability of this, however, are that the substrate population be significantly greater than the enzyme population, and that the intermediate enzyme-substrate complex be in a quasi-steady state for most of the reaction course. If the populations of the reactants in the system are low, then it is possible that these conditions do not hold. Thus, it may be necessary to include explicitly the reactions for the creation and decay of the enzyme-substrate complex.
Secondly, the terms in each differential equation must correspond to reactions in the model. Thus, if you require an oscillatory solution, and you impose this with an explicit oscillatory term, there is probably no easy way to translate this into a stochastic simulation. Remember that the algorithm works by extracting from the differential equations information pertaining to the actual reactions.