Abstract

An analysis is made of the experimental errors in the determination of agonist dissociation constants (Ka) by the method of Furchgott (Advan. Drug Res. 3 : 21-55, 1966), which employs β-haloalkylamines to bring about partial receptor inactivation. Several methods of weighted and unweighted regression analysis are outlined. These are evaluated by trials in 500 computer-simulated experiments, with respect to the variance and bias of the estimates of Ka and q (proportion of receptors not inactivated). The relative values of the different methods of analysis depend on the form of the relation between receptor activation and response, the proportion of "spare receptors," the relation of error or variation in response to the magnitude of the response, the degree of receptor inactivation brought about and the selection of doses given after partial receptor inactivation. Of the simple methods of analysis, the most generally useful seems to be graphical or regreesion analysis based on the linear equation used by Furchgott, but for this method the points to be plotted should be taken from the upper part of the postunactivation dose-response curve. Precision can be improved by appropriate weighted regreesion analysis.