Inconsistency between COPASI UI parameter estimation and basico parameter estimation

[Theo-BRN]

Heya,

I don't have information and good examples at the moment as I'm still investigating it. But when I run save_model_and_data, open the newly created .cps file in CopasiUI, run a parameter estimation task, it doesn't work.

Specifically, I get simulated values much higher than expected (higher than I would see when doing the same in basico). But I can't recreate this when running a Time Course simulation. This could be due to events not working, or the experimental txt files not being interpreted correctly or just some calculation problem with Global Quantities.

I will try find time to better evidence the problem, I just wanted to flag it now.

Best,
Theo

[Theo-BRN]

May all be fine, I think CopasiUI's parameter estimation has different default tolerances for ODE solving, compared to basico. But... it still provided a curve for a global quantity calculation involving a divide by zero.

Just two questions then.
What's the easiest way to access the r_tol and a_tol of basico's parameter estimation task?
Is it possible to use conditionals in my global quantity definition so I can do some 'error handling' within the calculation?

Any help with these two would be really appreciated!

Best,
Theo

[fbergmann]

The tolerances are saved in the COPASI file, so basico and CopasiUI would use the same ones. To get a hold of the tolerances, there would be different ones:

• the one at the time course task (which you'll get as relative and absolute from the method settings)
• the one for the optimisation method (which is named slightly different for the different optimisation methods)

I'm happy to look at the script or the copasi file to see where the differences come into play.

[fbergmann]

I cant reproduce this here, and would need more information