Open dweindl opened 4 years ago
As with #924, I think transiently adding anything to a model that could also be done in SBML is a bad idea.
As defining observables is not officially supported in SBML I don't see any point in defining parameters there that are not used anywhere inside the SBML model.
I don't see any issues with defining observables with parameter assignments.
It's possible, but it's not SBML's circus and not SBML's monkeys... (#924)
I think it's great to use SBML for what it is made, and that is rather for representing the biology. Not the observation model or anything leading to an objective function for parameter estimation.
@dweindl is this fixed by PEtab?
@dweindl is this fixed by PEtab?
Wouldn't call it fixed, but worked around by adding those parameters to the model.
Currently there seems to be no elegant way of using output parameters (e.g. scaling parameters) that have not been defined in the SBML model.
Would be great if they would be automatically recognized from observable and noise model formulas and created in the model by sbml2amici. In order to be more explicit, an optional to pass additional parameters would also do the job.