Radioligand assays-methods and applications. IV. Uniform regression of hyperbolic and linear radioimmunoassay calibration curves.

In order to handle all types of radioimmunoassay (RIA) calibration curves obtained in our laboratory in the same way, we tried to find a non-linear expression for their regression which allows calibration curves with different degrees of curvature to be fitted. Considering the two boundary cases of the incubation protocol we derived a hyperbolic inverse regression function: x = a1y + a0 + a-1y-1, where x is the total concentration of antigen, ai are constants, and y is the specifically bound radioactivity. An RIA evaluation procedure based on this function is described providing a fitted inverse RIA calibration curve and some statistical quality parameters. The latter are of an order which is normal for RIA systems. There is an excellent agreement between fitted and experimentally obtained calibration curves having a different degree of curvature.