The fifth lecture examined in more formal detail the syntax of the Kappa language. Numerous rule examples are given to illustrate its expressiveness. Two aspects of Kappa's minimalist design are explained.
An agent interface cannot have two sites with the same name
This reflects the idea that two domains of the same type (example : two SH2 domains in a RasGAP) are generally not located in the same structural context and could, therefore, differ in their interaction conditions, justifying distinct names. This restriction is crucial, as it makes the complexity of identifying a Kappa pattern match in a host graph at most quadratic with the size of the pattern.
A Kappa rule is designed to be local
The intention is that it should never be necessary to check outside the drawing to determine whether a drawing is contained in a host graph. However, if the left-hand drawing of a rule consists of two connected components, the rule may induce an intra- or intermolecular reaction event. The distinction is important, as the rate constant of the two cases differs in their volume of dependence. Such rules oblige the simulator to monitor global information, causing a substantial increase in computation time. Simple Kappa models are used to illustrate the architecture of an input file. A more advanced model introduces basic mechanistic details just to copy a polymer. This model implements replay based on kinetic and energetic discrimination as explained in the lecture 3 and illustrates how error (entropy) can lead to energetically unfavorable polymerization. Although the model is analytically soluble in the steady state, its dynamics would be impractical to express with the usual equations of chemical kinetics. Demonstration and execution of the model introduces the graphical user interface of the Kappa platform.
