Multivalent bindings are a combination of several weaker bonds between two molecules such as a receptor and a ligand. Because the combined bonds are stronger than the individual ones, the molecules may interact longer. Thus, this concept is very interesting for targeted drug design. Theoretically, these bindings and unbindings are rare stochastic events. Literature on multivalency is vast, especially on specific experimental setups. However, there exists little theoretical studies for general n-valent settings. This thesis aims to fill that gap by modelling the binding process as kinetic rate matrices and applying the clustering algorithm PCCA+. While the binding and unbinding of the single ligand-receptor pairs happens on a fast time scale i.e. in the micro-perspective, the association or dissociation of a complex is acting on a slow timescale, thus in the macro-perspective. The existing kinITC method does not capture the switch between these time scales. Thus the method proposed in this thesis describes an alteration to kinITC, called kinITC+. The key findings of the thesis are: • It is possible to gain kinetic information from thermodynamic data. • The macroscopic binding rate kon is not constant, but ligand concentration dependent. • There is at least one counterexample to the assumption that the thermodynamic contribution of rebinding increases with valency.
