Iterated Reaction Graphs:  Simulating Complex Maillard Reaction Pathways

This study investigates a new method of simulating a complex chemical system including feedback loops and parallel reactions. The practical purpose of this approach is to model the actual reactions that take place in the Maillard process, a set of food browning reactions, in sufficient detail to be able to predict the volatile composition of the Maillard products. The developed framework, called iterated reaction graphs, consists of two main elements:  a soup of molecules and a reaction base of Maillard reactions. An iterative process loops through the reaction base, taking reactants from and feeding products back to the soup. This produces a reaction graph, with molecules as nodes and reactions as arcs. The iterated reaction graph is updated and validated by comparing output with the main products found by classical gas-chromatographic/mass spectrometric analysis. To ensure a realistic output and convergence to desired volatiles only, the approach contains a number of novel elements:  rate kinetics are treated as reaction probabilities; only a subset of the true chemistry is modeled; and the reactions are blocked into groups.