How to get Prism's nonlinear regression to fit different models to different data sets.

Prism's nonlinear regression analysis only lets you pick one equation to fit to all the data sets selected for the analysis. If you want to fit different models to different data sets, you have two choices:

• Repeat the analysis several times. When you choose nonlinear regression from the Analyze dialog, you choose which data sets are analyzed on the right side of that dialog. Choose one data set (or several) the first time you fit a curve. Then run nonlinear regression again, choosing different data sets and a different equation.
• Write your own equation, and use special syntax to make some lines apply to only one data set. Below is an example. It fits column A to a model that defines total binding and column B to a model that defines nonspecific binding only. The first two lines of the equation are evaluated for all data sets, the third line is only evaluated for data set A, while the last line is only evaluated for data set B. To fit this model, you would want to set the constraint that the parameter NS is shared between data sets.

Specific=X*Bmax/(X+Kd)
Nonspecific=NS*X
<A>Y=Specific + Nonspecific
<B>Y=Nonspecific

If your goal is to fit one model, but with different constant values for each data set, see this FAQ.