MacELLIPSE, A Graphical Aid to the Problem of the Joint Confidence Region: a Practical Example for Ligand Binding Experiments

Abstract

The analysis of receptor binding experiments involving saturation or displacement curves may represent a problem when data are obtained from a number of experimental groups characterized by different tissues or different treatments. The conceptual difficulty of comparing such binding parameters and the complexity in graphically representing many curves is often overcome using tables and statistical tests for uncorrelated parameters. These, however, might be misleading. In fact, computing parameters using a nonlinear least squares (NLLS) technique may yield a significant covariance or correlation in the parameter estimates. In the present paper we suggest a practical method for comparing curves from different experimental groups using the LIGAND program and present the program MacELLIPSE that permits the graphical representation of the binding results together with the correlation existing between the estimated parameters.