Is there a cobrapy function for gapfilling with required metabolic tasks?

[PkiwiBird]

Cobrapy user guide explains how to do gapfilling when we want to guarantee that the organism objective is to grow or guarantee a certain metabolite is produced. Is there a cobrapy function equivalent to the 'fitTasks' function on RAVEN toolbox in matlab https://github.com/SysBioChalmers/RAVEN/blob/master/core/fitTasks.m that performs gapfilling by providing a list of essential metabolic tasks? And is it possible to use the function from matlab on python?

[gregmedlock]

This is certainly possible in cobrapy, but there isn't a "convenience" function for performing gap filling given a list of tasks. If you can share the types of tasks you're trying to gapfill for, we can point you in the right direction; the constraints associated with common tasks are easily set it cobrapy.

There is some functionality for running MATLAB functions from python, but I would recommend a pure python or pure MATLAB approach if possible.

[PkiwiBird]

Thank you for replying. I basically created new reactions with existing metabolites and then added bounds to those reactions. For example, a task of the type: ATP[c] + H2O[c] <=> ADP[c] + Pi[c]. But I don't know how to proceed from there. I guess we start by adding one of those reactions to the model, but how do we "inform" the gapfilling function that we want that reaction to have flux while gapfilling?

[gregmedlock]

In that case, constraining the lower bound of the reaction of interest to be > 0 (if forward flux is desired) would be sufficient. So this would do it:

reaction_of_interest = model.reactions.get_by_id("your_reaction_id_here")
x = 0.05 # desired minimum flux for the task
reaction_of_interest.bounds = (x, reaction_of_interest.upper_bound)

After constraining this reaction, the constraint will be maintained when you create a gapfilling object in cobrapy (or run the gapfill function). You can add as many constraints as you'd like.

[PkiwiBird]

Makes sense. What is odd is the file with metabolic tasks has some reactions as you said with lb 1 or 0.0001 and ub 1000, but others have lb 0 and ub 1000 or even lb -1000 to 1000. So, required metabolic tasks can have flux 0 ? For metabolic task to happen doesn't have to have flux value different than 0?

[gregmedlock]

I agree that some of those bounds look odd--I can't imagine a scenario where a task with the bounds (-1000,1000) would be useful alone (e.g., unless maximizing or minimizing flux through that reaction is also set as the objective function, and there is some other step checking for a particular objective value). (0,1000) could be used to indicate that anything except reverse flux is allowed for a reaction that the creators wanted to keep reversible in the reconstruction, but the use case is unclear to me.

[cdiener]

This looks stale. Feel free to reopen if the issue persists.