Quantifying thermodynamic parameters of multi-step binding using a novel approach to Bayesian inference

Abstract

Multistep protein-protein or protein-ligand binding interactions are of critical biological importance in numerous cellular processes. However, probing the thermodynamics of complex binding by isothermal titration calorimetry (ITC) or other methods often relies on applications of simple models with strong assumptions not representative of the underlying interactions, which can lead to overly narrow uncertainty estimates. Focusing on ITC, standard practice in the field is to fit binding curves to simple models which return a single enthalpy and free energy term representing an overall picture of binding thermodynamics, but offering minimal information on intermediate steps.