Open-Systems-Pharmacology / OSPSuite.ReportingEngine

R implementation of the reporting engine
https://www.open-systems-pharmacology.org/OSPSuite.ReportingEngine/
Other
5 stars 5 forks source link

MeanModel Sensitivity plot crashes for some PK-Parameters. #218

Closed Yuri05 closed 4 years ago

Yuri05 commented 4 years ago

Sensitivity plot crashes for some PK-Parameters. E.g. for 03.zip: Values 'AUC_inf_tD1_norm' are not included in parent values: 'C_max', 'C_max_tD1_tD2', 'C_max_tDLast_tEnd', 't_max', 't_max_tD1_tD2', 't_max_tDLast_tEnd', 'C_trough_tD2', 'C_trough_tDLast', 'AUC_tD1_tD2', 'AUC_tDLast_minus_1_tDLast', 'AUC_inf_tD1', 'AUC_inf_tDLast', 'MRT', 'Thalf', 'Thalf_tDLast_tEnd'."

Where does this parent values list come from? It containes only 15 out of 22 calculated PK-Parameters (s. Report\PKAnalysisResults\SimSet-PKAnalysisResults.csv)

msevestre commented 4 years ago

What's going on with these PK parameters? It's pretty simple: the list of PK parameters depends on the simulation. There is no global list that can be used.

Or is there one that is specified by the user?

Yuri05 commented 4 years ago

User specifies per Output which PK-Parameters he is interested in. Which of course must be a subset of all available PK-parameters for this simulation. But the code of calc sensitivity obviously uses for the check "is required PK-parameter available?" not the list of "all available PK-parameters for this simulation", but something else.

pchelle commented 4 years ago

The issue is the same as #214 , validateIsIncluded was added to improve the error message. parentValues are saResults$allPKParameterNames with saResults=importSensitivityAnalysisResultsFromCSV So it corresponds to all the pk parameters available in the csv file saved during the sensitivity analysis step.

Yuri05 commented 4 years ago

ok, that means that SA does not return values for ALL PK Parameters. PK-Parameters calculation returns here: grafik And SA-calculation returns: grafik all _norm parameters are missing in SA @msevestre An issue in OSPSuite.Core then? Probably related to https://github.com/Open-Systems-Pharmacology/OSPSuite.Core/issues/798

msevestre commented 4 years ago

It does not matter if the parameters are missing or not. The fact that the parameters are not in the list of available parameters means they should not be used programmatically by the code no?

As to why this is like that: Can you rerun with the latest ospsuite:r to see of this solves the issue.? That being said, since all _norm parameter are calculated, I don't see why they would not be calculated for SA. I think it's something else

How can I reproduce just with OSPSuite-R? How was the SA performed? Where all parameter varied or just a subset?

msevestre commented 4 years ago

@Yuri05 Not sure why parameters are not in the list. I just tried with the latest version and still the same issue. Is a bodyweight available for this simulation?

msevestre commented 4 years ago

@Yuri05 Please see comment here This is by design https://github.com/Open-Systems-Pharmacology/OSPSuite.Core/issues/800#issuecomment-642283061

Yuri05 commented 4 years ago

Please see my reply here :) https://github.com/Open-Systems-Pharmacology/OSPSuite.Core/issues/800#issuecomment-642308802