Introducing Quantitative Assessment of Accuracy for the Equilibrium Constant of Affinity Complexes

The equilibrium constant of affinity complexes (Kd) is among the most frequently determined physicochemical parameters, with thousands of papers reporting Kd values monthly. Kd is classically computed via nonlinear regression of a binding isotherm and, while it can be precise, it is often inaccurate due to error amplification. Currently, no method exists for quantitatively assessing the accuracy of Kd — here, we fill this knowledge gap. We introduce the accuracy confidence interval (ACI): a range within which the accurate value of a parameter lies with a defined probability. We also present the “ACI-Concept”: a general approach for determining the ACI of parameters computed with correct nonlinear regression models. The ACI-Concept combines regression-stability and error-propagation analyses. We apply the ACI-Concept to develop a workflow for determining the ACI of Kd from a single binding isotherm. We verify this workflow with computer-simulated and experimental binding isotherms. Finally, we implement this workflow in a user-friendly web application (https://aci.sci.yorku.ca) to facilitate its fast adoption by the broad research community. Knowing the ACI of Kd and other parameters computed through nonlinear regression will help researchers avoid misconceptions that can arise when relying solely on precision.