Clarification/specification of order of condition changes required

[matthiaskoenig]

Depending on the order of the changes applied for a condition the model will be in a different state. I.e. the order of the columns matters in the condition table. E.g. a model with species S in compartment C with an initial concentration of 1.0. If you change first the Volume C and then the initial concenctration [S0], the model has different values for [S0] compared to first changing concentration and then volume.

Currently it states

Row- and column-ordering are arbitrary, although specifying conditionId first may improve human readability.

This should be changed to:

Row- and column-ordering do not have any restrictions, although specifying conditionId first may improve human readability. The parameterOrSpeciesOrCompartmentIds changes for a condition are applied in order of the columns.

See also for instance: https://github.com/AMICI-dev/AMICI/issues/2062

[FFroehlich]

Hmm, interesting point. I would prefer to assume that volume changes always take precedence over concentration changes.

[dweindl]

Agreed, that needs to be clarified.

I would prefer to assume that volume changes always take precedence over concentration changes.

This is also what would happen if an SBML model has an InitialAssignment for both the initial concentration and the respective compartment size, right? Sounds good then.

[FFroehlich]

SBML level 3 version 2 release 2:

The ordering of InitialAssignment objects in a model is not significant. The collection of InitialAssignment, AssignmentRule and KineticLaw objects forms a set of assignment statements that must be considered as a whole. The combined set of assignment statements should not contain algebraic loops: a chain of dependency between these statements should terminate.

[FFroehlich]

would propose to handle this the same way as for simultaneous assignment of volume size & concentration:

Upon such size changes, the value of the concentration or density must be recalculated from the simple relationship concentration = amount/size if the value of the concentration is needed (for example, if the species identifier appears in a mathematical formula or is otherwise referenced in an SBML construct). There is one exception: if the species’ quantity is determined by an AssignmentRule, RateRule, AlgebraicRule, or an EventAssignment and the species has the attribute value hasOnlySubstanceUnits=“false”, it means that the concentration is assigned by the rule or event; in that case, the amount must be calculated when the compartment size changes.

[matthiaskoenig]

This is the same issue as in SBML Events with multiple EventAssignments based on a trigger (i.e. this has been solved already for events and the same solution should be used here).

4.12.5 EventAssignment

An “event assignment” has effect when the event is executed. (As noted above, the operation of event is divided conceptually into two phases: the first phase when the event is triggered and the second phase when the event is executed.)

In SBML all event assignments are performed simultaneously. The same should happen in PEtab for all the model changes for a given condition. I.e. there is no order of execution. All changes are applied at once.