search

PEtab-dev / PEtab

PEtab - an SBML and TSV based data format for parameter estimation problems in systems biology
https://petab.readthedocs.io
MIT License
57 stars 12 forks source link

Handling of local parameters #337

Open LeonardSchmiester opened 4 years ago

LeonardSchmiester commented 4 years ago

At the moment, local parameters cannot be directly estimated within PEtab, as this requires unique parameter Ids. One would need to make them global parameters. Maybe instead we could use a unique naming scheme like in AMICI (${reactionid}${local_par_id}) to also allow for local parameters

FFroehlich commented 4 years ago

https://github.com/PEtab-dev/PEtab/blob/8b03c8f20c8042c5375adabfe23bab19c3a649b0/petab/sbml.py#L115 ?

dweindl commented 4 years ago

https://github.com/PEtab-dev/PEtab/blob/8b03c8f20c8042c5375adabfe23bab19c3a649b0/petab/sbml.py#L115 ?

Good catch. So far unused here. Legacy... And from before I knew that libsbml already implements that same functionality.

Question is mostly if we want to handle local parameters that way.

JanHasenauer commented 4 years ago

Is there a disadvantage with this and would there be another good option?

dweindl commented 4 years ago

Is there a disadvantage with this and would there be another good option?

Disadvantage of allowing for local parameters: They don't necessarily have a unique ID. It's a per-reaction namespace. In PEtab we'd need globally identifiable parameter IDs. The common way is using ${reactionid}${local_par_id}, but this is still not necessarily unique.

JanHasenauer commented 4 years ago

Okay, but this is a problem of local parameters in general but not the suggested implementation. As we want to support available SBML models, we should in my opinion also support local parameters.

The common was sounds good and we should then be check for uniqueness.

dweindl commented 2 years ago

The changes proposed in #538 also provide a means to handle local parameters.