Reaction flux control does not consider individual enzyme pool constraints

[voidsailor]

Hello @Paulocracy,

I was using autoPACMEN to create a model with quantitative proteomics data, i.e., I want to integrate individual enzyme pools into the model (in addition to the global protein allocation constraint).

Before calibration in Matlab, the script "reaction_flux_control_by_scenario.py" needs to be executed to determine whether removing protein allocation constraints affects the objective value (section 6.1 in the manual). The output is a list of reactions whose kcat values will be calibrated.

I ran this script, but the relative change of the objective value was always 1, i.e., removing the protein allocation constraint had no effect on the objective and the list of reactions to be calibrated was empty.

However, I noticed that the reactions in my model were not restricted by the global protein allocation but by the individual enzyme pools. Looking into the script "reaction_flux_control_by_scenario.py" and at function "reaction_flux_control", I also noticed that only the global protein allocation constraint is removed but not individual enzyme pool constraints.

From my intuition, individual enzyme pools should be considered in the calibration and I would extend the respective function. What is your opinion on this?

Thanks! :)

[Paulocracy]

Hello @voidsailor :D
It's great to hear from you :-)
And yes it makes total sense that the calibration method typically won't look at any subcompartmental protein pools. That might be because your mentioned selection of k_cats that are going to be changed depends on whether or not they have an effect on one particular scenario optimization if they would not be enzyme-constrained, which might just disregard additional protein pools.
So feel free to expand the calibration functions for multiple enzyme pools :-) Unfortunately, I don't have enough time for this right now, but I hope that the code changes are not too complicated :3
The main reason why I don't have time is another major modeling package I'm working on (it will be released this year), which might be a better starting point than AutoPACMEN for any new calibration routines, but I cannot promise a launch date :-/ while AutoPACMEN is ready to be changed ;-)

[voidsailor]

Thanks for your feedback, I will create a pull request as soon as I have something working :)
Great to hear about your new package, I'm looking forward to it!