Export to snapshot / reimport in PK-Sim alters protein reference concentration

[AndreDlm]

The attached PK-Sim file contains a European Standard individual with 2 CYP3A4 enzymes (CYP3A4 and CYP3A4 1), both loaded from the gene expression database with a reference concentration of 4.32 µmol/L each.
When exporting the file to Snapshot (attached) and re-importing it in PK-Sim, the reference concentration of CYP3A4 1 is set to 1 µmol/L.

Alprazolam.zip

[Yuri05]

The problem here is the enzyme naming. A snapshot saves only parameters which differ from default.
When adding an Enzyme, it's reference concentration is set to a default value, which is <>1 for the set of "known" enzymes, otherwise set to 1.
CYP3A4 is known, CYP3A4 1 is unknown - thus when reopening from snapshot, it's RefConc is set to 1.

The behaviour can be reproduced even easier:

• Add an individual
• Add Enzyme CYP3A4 (even not from the data base) (RefConc is set to 4.32 µmol/l)
• Rename CYP3A4 to e.g. CYP3A4_renamed
• Save to snapshot and reload: RefConc of CYP3A4_renamed is now 1 µmol/l

@msevestre The problem here is renaming of an enzyme. As soon as we rename an enzyme, all it's parameters, which

• have the default values (= defaults for the old enzyme name) and
• have different default values for the new enzyme name

must be set to changed. Otherwise reimport from snapshot will create wrong values.

[msevestre]

Because we do not save the relexp query (rather the relexp expression value), I'd suggest to always save the global parameters in snapshot (RelExp and T1/2s). We hardly have any default value anyways and this would make the whole experience much cleaner and predictable

[Yuri05]

I'd suggest to always save the global parameters in snapshot (RelExp and T1/2s).

@AndreDlm @sfrechen @TWendl do you agree?

[sfrechen]

Agree to Michael.
Might be worth to re-evaluate once we would come up with another database concept for relExp and refConc.

[TWendl]

Agree to Michael as well!
For future releases: Would it be possible to technically enable the matching of two processes to the same enzyme? E.g. in AndreDlm's example two metabolic processes to the same enzyme (CYP3A4) such that there is no need to create an additional Enzyme (CYP3A4 1)?

[sfrechen]

@TWendl That is already possible.

[TWendl]

@sfrechen OK! Didn't know that! Thanks!

[msevestre]

Ok majority rules here (sorry @AndreDlm as you did not respond). I'll implement it as described here