search

LeonardSchmiester / Benchmark-Models

This repository is not maintained anymore. Please use: https://github.com/Benchmarking-Initiative/Benchmark-Models-PEtab
BSD 3-Clause "New" or "Revised" License
8 stars 5 forks source link

Noise parameter documentation #36

Closed yannikschaelte closed 5 years ago

yannikschaelte commented 5 years ago

Document how the noise parameter construct is meant (this lead to some problems in the import and seems to be handled not fully coherently in the converted models ...).

LoosC commented 5 years ago

will add a description in the documentation, but short answer here:

* What exactly is the concept behind the sigma_... and noise_... parameters in the sbml file (again :) )?
  • sigma_observableId specifies the full noise model, could for example also be something like: sigma_observableId = noiseParameter1_observableId + noiseParameter2_observableId * observable_observableId, but usually we have just a simple noise model with sigma_observableId = noiseParameter1_observableId
    • How to handle condition-specific noise parameters / models?
  • which conditions/observables/... share noise parameters is be defined in the measurementData sheet
    • What is supposed to be filled in the noiseParameters column in the measurement file unless specified as a float?
  • sorry I do not understand the question? in this column we have the string for the noise parameter
yannikschaelte commented 5 years ago

Thanks for clarifying. Concerning question 1: That's what I guessed. So in the Zheng model, should that one noise parameter be mapped to all noiseParameter1_observableId parameters? I.e. we have a mapping rule ``${noise} -> ${noise}Parameter1_${observableId}$ \forall ${observableId}?

Concerning question 3: By noise parameter you mean the sigma parameter in the wording of the sbml file? Or noise=sigma?

Ok, let's maybe discuss tomorrow. Probably I'm just confused. :)

LoosC commented 5 years ago

Thanks for clarifying. Concerning question 1: That's what I guessed. So in the Zheng model, should that one noise parameter be mapped to all noiseParameter1observableId parameters? I.e. we have a mapping rule ``${noise} -> ${noise}Parameter1${observableId}$ \forall ${observableId}?

yes exactly

Concerning question 3: By noise parameter you mean the sigma parameter in the wording of the sbml file? Or noise=sigma?

I mean the noiseParameter in the wording of the SBML file but yes lets discuss tomorrow :)

yannikschaelte commented 5 years ago

resolved.