LIGAND: A Versatile Computerized Approach for Characterization of Ligand-Binding Systems

We have developed a general strategy and a versatile computer program for analysis of data from ligand-binding experiments (e.g., radioreceptor assay systems for hormones, neurotransmitters, drugs). This method provides optimal (weighted least squares) estimates of “binding parameters” (affinity constants, binding capacities, nonspecific binding) for any number of ligands reacting simultaneously with any number of receptors. This approach provides two major advantages compared with other available methods: (i): It uses an exact mathematical model of the ligand-binding system, thereby avoiding the possible biases introduced by several commonly used approximations. (ii) It uses a statistically valid, appropriately weighted least-squares curve-fitting algorithm with objective measurement of goodness of fit, thereby avoiding the subjective graphical or simplified statistical methods which may introduce bias. Additional important features include the following. (i) The level of nonspecific binding is regarded as an unknown parameter, subject to uncertainty, which must be estimated simultaneously with other parameters of the system by appropriate statistical methods. This approach provides a more accurate and precise estimate of the parameters and their standard errors. (ii) Selected parameters can be forced to share a common value, or be fixed at any desired constant value. This feature facilitates hypothesis testing by appropriate statistical methods e.g., testing whether a particular experimental manipulation results in a change in affinity (K), binding capacity (R), or both parameters. (iii) One can combine results from multiple experiments by introduction of explicit scaling or “correction” factors which compensate for the commonly observed large degree of between-experiment variation of the overall binding capacity (Bmax) while other properties of the system (e.g., K values, relative binding capacities for high- and low-affinity sites) are highly reproducible. (iv) One can characterize complex cross-reacting systems involving any number of ligands reacting simultaneously with any number of binding sites. This enables one to pool results from several curves obtained using several different ligands.