Parameter Identification for OSP Inhalation Model simulation.

[pyrokinesis244]

Dear All,

I trust this inquiry meets you well.

I have been stuck on trying to utilize the surrogate absorption pathway (the OSP inhalation model) to access the fate of a sublingually administered molecule of interest. I succeeded (not with absolute certainty) in manipulating, using the R codes in the model, to tweak the particle bin numbers to capture deposition in the lung generations as well as other useful physiological & pharmacokinetic parameters. Also i successfully ran a simulation using the published tutorial directives (in Mobi) to get the prediction at first before overlaying the observed data (though there was discrepancies with the model predicted values & observed values from the study but it was well within the ).

Furthermore, i started the parameter identification [P.I] (this was done in two stages after i speculated i could vary the minimum and maximum values for some metrics) for optimization of useful PK & physiological metrics (in mobi) after the first simulation with one of the doses. The first parameter identification schedule yielded results that were in varying agreement with the observed data although this was more in the elimination phase as there were two absorption peaks in the absorption phase.
The second P.I had one absorption peak and showed necessities for further optimization.

My question is this:

1. Is the parameter identification meant to be run in MOBI initially without the population of interest in the conducted clinical study without taking into account the enzymes that are responsible for the compound's metabolism or does it have to be run in PK-Sim after sending the initial simulation in Mobi and characterising it with the population that was captured in the clinical study.

2. Does the max. number of evaluations (in the P.I configuration) for the Nelder Mead's Unconstrained optimization option, as well as the convergence tolerance need to be altered if the parameters for identification are few or much ( I noticed the P.I consumed time >2hours due to this iteration setting)?

Please, I implore the clemency of your experience and knowledge to assist in helping to answer this question.

I've attached two images for the respective P.I simulation predictions in Mobi although this is yet to be applied to the simulation

Thanks in advance.
P..