I'm currently working on building a simulator for the early earth to examine the spontaneous occurrence of food generated autocatalytic sets. No one in my group is a chemist, so the bottle neck for the project is implementing some sort of realistic chemistry in our model.
We are considering using the software outlined in this paper: http://arxiv.org/abs/1304.1356
This will allow us to create a reaction network from an initial set, based on a set of general rules. The paper demonstrates well how the formose reactions are executed, which we intend to include. What other reaction groups would it be important to include? Our initial set of molecules would be comprised of the food molecules for each set of reaction rules included, we don't intend to start from H,O,N,C for example, so fundamental chemical rules should not be too crucial, as they should be generalised into the major reaction groups we will use.