GraphPad Curve Fitting Guide

"Interrupted"

"Interrupted"

Previous topic Next topic No expanding text in this topic  

"Interrupted"

Previous topic Next topic JavaScript is required for expanding text JavaScript is required for the print function Mail us feedback on this topic!  

What does 'interrupted' mean?

Prism reports 'interrupted' in two situations:

Nonlinear regression is iterative and the maximum number of iterations is specified in the Diagnostics tab, with 1000 being the default. With data that doesn't really define all the parameters in the model, Prism can just keep iterating until it hits the limit. In that case it stops and reports "interrupted". You can lift the limit in the Diagnostics tab of the nonlinear regression dialog, but usually the problem is that the data simply doesn't enough information for the model to fit.

The fit was slow, and you clicked "Interrupt" on the progress dialog.

Checklist

If the maximum number of iterations was set to a low value, set it to a higher value and try again. If you have lots of data points and lots of parameters, nonlinear regression can sometimes require hundreds of iterations.

If the maximum number of iterations was already set to a high value, you can try a still higher value, but most likely Prism is still not going to be able to find a best-fit curve. Things to check:

Did you enter the right model?

Does your data provide enough information to define that model. For example if you fit a log(dose) response curve, does your data show a sigmoidal shape, enough to define the top or bottom plateaus? If your data doesn't define all the parameters, consider constraining one or more parameters to constant values.

Does the curve defined by your initial values come near your data? Check the option on the diagnostics tab to plot that curve.

If you entered constraints, were they entered correctly?

If you didn't enter any constraints, consider whether you can constrain one or more parameters to a constant value? For example, in a dose-response curve can you constrain the bottom plateau to be zero?

Can you share a parameter over all the data sets (global fitting)?