A simulation which was running in V10 does not run in V11 anymore

[Yuri05]

DDI_Treatment-Rifampicin_Midazolam-Backman1998.zip
Loading this snapshot works in V10:

In V11 the error message "Variables became negative" comes:

Also interesting: when loading the pkml exported for this simulation by the qualification runner in MoBi and running the simulation,
the same error message comes - but at different time point

After reducing the absolute tolerance from 1E-10 to 1E-11: simulation runs in V11.
However the simulation result is also slightly different.

Attached is a MoBi project with 2 simulations:
DDI_Treatment-Rifampicin_Midazolam-Backman1998.zip

• Simulation _DDI... exported from V10 (after importing the snapshot in PK-Sim 10)
• Simulation DDI... was created by the Qualification Runner (as part of the CYP3A4-DDI Quali workflow) and in MoBi AbsTol was reduced from 1E-10 to 1E-11 to be able to run the simulation

V10-Simulation results (_DDI...):

V11-Simulation results (DDI)

[Yuri05]

P.S. just to be sure: I rerun the V10 simulation with reduced AsTol: 1E-11
Did not change the result

[msevestre]

Looks like the new changes in SimModel might have introduce the change no?

Or maybe this is due to the new svg icons?😃

[Yuri05]

Or maybe this is due to the new svg icons?

Yep, that must be the root of the problem :)

[Yuri05]

Looks like the new changes in SimModel might have introduce the change no?

Don't think so. Both simulations above (V10 and V11) were imported and run in MoBi 10.

[msevestre]

Are you saying it runs in Mobi but not in PKSim?? What?

[msevestre]

And Mobi and pksim returns different values? I can't quite believe this

[Yuri05]

no no.

• Snapshot=>PK-Sim 11: cannot import simulation because it does not run
• Simulation 1: Snapshot=>PK-Sim10=>Export to PKML => Import in MoBi 10 => Runs AS IS
• Simulation 2: Snapshot=>QualificationRunner 11 =>Export to PKML => Import in MoBi 10 => Does not run AS IS but runs with reduced AbsTol (and result is sligthly different from Sim1)

MoBi project with Simulation 1 and Simulation 2 is what is attached above.

[Yuri05]

OK... so now I imported both simulations (Sim1_Sim2_pkmls.zip) into MoBi 11 project: MoBi11_DDI_Treatment-Rifampicin_Midazolam-Backman1998.zip

And here the V11-simulation runs with the predefined AbsTol=1E-10 (??). But the result is the same as in MoBi 10 with absTol=1e-11 (and again, different from the result of V10-simulation).

Hmm, hmm, hmm...

[msevestre]

I am so confused right now. Let's discuss this tomorrow :)

[Yuri05]

OK, one reason is: some protein start amounts are different between PK-Sim 10 and PK-Sim 11.

Path	Start Amount PK-Sim 10	Start Amount PK-Sim 11
Bone|Interstitial|P-gp	0,40414	0
Gonads|Interstitial|P-gp	0,00099	0
Heart|Interstitial|P-gp	0,02469	0
Stomach|Interstitial|P-gp	0,00620	0
SmallIntestine|Interstitial|P-gp	0,23021	0
LargeIntestine|Interstitial|P-gp	0,05136	0
Lung|Interstitial|P-gp	0,11259	0
Muscle|Interstitial|P-gp	0,59072	0
Pancreas|Interstitial|P-gp	0,00383	0
Spleen|Interstitial|P-gp	0,02167	0

All other protein start amounts are the same between V10 and V11

[Yuri05]

And the reason for the difference of the start amounts is: fraction expressed basolateral is 1 in V10 and 0 in V11
So the fraction expressed basolateral is obviously wrong in V11.

Here how the formulas for e.g. Bone|Interstitial|P-gp|Initial concentration look like in V10 and V11

Parameter: Sim1_V10Organism|Bone|Interstitial|P-gp|Initial concentration
	Value: 0.34195064649 [µmol/l]
	Formula: f_int > 0 ? RC * rel_exp_org * f_exp_org_basolatateral * 1 / f_int : 0
		f_exp_org_basolatateral: Sim1_V10Organism|Bone|Interstitial|P-gp|Fraction expressed basolateral(1)
		f_int: Sim1_V10Organism|Bone|Fraction interstitial(0.1)
		RC: Sim1_V10P-gp|Reference concentration(1.41 [µmol/l])
		rel_exp_org: Sim1_V10Organism|Bone|Intracellular|P-gp|Relative expression(0.0242518189)

Parameter: Sim2_V11Organism|Bone|Interstitial|P-gp|Initial concentration
	Value: 0 [µmol/l]
	Formula: f_int > 0 ? RC * rel_exp_org * f_exp_org_basolateral * 1 / f_int : 0
		f_exp_org_basolateral: Sim2_V11Organism|Bone|Interstitial|P-gp|Fraction expressed basolateral(0)
		f_int: Sim2_V11Organism|Bone|Fraction interstitial(0.1)
		RC: Sim2_V11P-gp|Reference concentration(1.41 [µmol/l])
		rel_exp_org: Sim2_V11Organism|Bone|Intracellular|P-gp|Relative expression(0.0242518189)

[Yuri05]

Here a very simple example to reproduce (just individual with a transporter, no simulations):
Individual_with_Transporter.zip

When importing the snapshot in V10: initial concentrations in interstitial are correct:

And in V11: