byrdjb
commented
2 years ago Thank you for these observations. I was never able to get to the bottom of these issues when I looked at the code. I, too, would be very interested in working code that does what the paper describes.
From: stfrkniz @.> Sent: Friday, October 14, 2022 5:36:11 AM To: hallowkm/RenalModel @.> Cc: Subscribed @.***> Subject: [hallowkm/RenalModel] Not the parameters described in the publication (Issue #3)
External Email - Use Caution
I downloaded the code from here and also compared it with the textfiles from the supplement on the website where the paper ('A Quantitative Systems Physiology Model of Renal Function and Blood Pressure Regulation: Model Description') about this model was published. There are a few issues with the program and the model.
First, like others already discussed, the code does not work in R. There is an error that some parameters are missing although they obviously have a value set to them.To avoid this problem I implemented the code in python so that I can run the program and try out the simulations of the model.
Second, I noticed that a lot of parameters set in the file 'calcNomParams_human' do not match with the values for them described in the paper. If you implement the values described in the paper, simulations will not run appropriately (errors when you try to simulate even a rather short time span (e.g. 2-3 days) and some variables can not be computed since you divide through zero etc.).
Third, if you run simulations without changing the parameters according to the paper, the simulations will run, but they can not be the results differ from the results described in the paper. The following simulation shows the influence of the choice of controller gains for proportional-integral feedback controllers on the response of cardiac output and Na concentration to a perturbation (double the Na intake): [Simulation_HallowPaper]https://user-images.githubusercontent.com/115784216/195812267-fff637e1-d4ba-4fc6-8295-bc4bb480f57b.png I replicated this simulation but got a noticably different output: [Simulation_default_model]https://user-images.githubusercontent.com/115784216/195811160-5bcb7a11-4140-40d2-8557-93bfd5b81428.png Not only is for example the Na Concentration reduced instead of increased, it also takes ten times longer to get back to the steady state Na conecentration of 140.0 mEq/L.
I am really interested in this model but the fact that the values set for the parameters in the programm are different from the values described in the paper and no matter which parameterization you choose, you can not replicate simulation output and therefore can not validate this model is frustrating.
Furthermore a lot of the parameters in the model are not described and there is no indication where they are from so it is not possible to find out or compare the values for myself.
Did somebody manage to replicate simulations and values described in the paper and if yes, what parameters have to be changed and to what value?
— Reply to this email directly, view it on GitHubhttps://github.com/hallowkm/RenalModel/issues/3, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AG6BHPMRJXH3C4XQX7O7GKLWDESQXANCNFSM6AAAAAARFB743A. You are receiving this because you are subscribed to this thread.Message ID: @.***>
Electronic Mail is not secure, may not be read every day, and should not be used for urgent or sensitive issues
Zhaoyxcpu
I replicated this simulation but got a noticably different output: 
Not only is for example the Na Concentration reduced instead of increased, it also takes ten times longer to get back to the steady state Na conecentration of 140.0 mEq/L.
I downloaded the code from here and also compared it with the textfiles from the supplement on the website where the paper ('A Quantitative Systems Physiology Model of Renal Function and Blood Pressure Regulation: Model Description') about this model was published. There are a few issues with the program and the model.
First, like others already discussed, the code does not work in R. There is an error that some parameters are missing although they obviously have a value set to them.To avoid this problem I implemented the code in python so that I can run the program and try out the simulations of the model.
Second, I noticed that a lot of parameters set in the file 'calcNomParams_human' do not match with the values for them described in the paper. If you implement the values described in the paper, simulations will not run appropriately (errors when you try to simulate even a rather short time span (e.g. 2-3 days) and some variables can not be computed since you divide through zero etc.).
Third, if you run simulations without changing the parameters according to the paper, the simulations will run, but they can not be the results differ from the results described in the paper. The following simulation shows the influence of the choice of controller gains for proportional-integral feedback controllers on the response of cardiac output and Na concentration to a perturbation (double the Na intake):
I replicated this simulation but got a noticably different output:
Not only is for example the Na Concentration reduced instead of increased, it also takes ten times longer to get back to the steady state Na conecentration of 140.0 mEq/L.
I am really interested in this model but the fact that the values set for the parameters in the programm are different from the values described in the paper and no matter which parameterization you choose, you can not replicate simulation output and therefore can not validate this model is frustrating.
Furthermore a lot of the parameters in the model are not described and there is no indication where they are from so it is not possible to find out or compare the values for myself.
Did somebody manage to replicate simulations and values described in the paper and if yes, what parameters have to be changed and to what value?