These are chat archives for opencobra/cobrapy

24th
Mar 2017
Peter St. John
@pstjohn
Mar 24 2017 14:16
So i noticed that pfba returns the absolute flux as the f attribute, rather than the optimal objective value
so cobra.flux_analysis.pfba(model).f vs. model.optimize().f will be different
were we still planning on allowing pfba to be called via a kwarg to optimize? if so, we’ll probably want to standarize the f value
Henning Redestig
@hredestig
Mar 24 2017 15:23
I think I changed my mind on calling pfba via optimise, mainly for the 'one way to do it' and question of what happens if you have another fba method in the future, should they then all be added to optimize(?)...
But anyway, with pfba changing the objective, I felt that it more unsurprising that the objective value no longer is only that of the flux through the reaction you had before doing pfba, than the previous behavior
Henning Redestig
@hredestig
Mar 24 2017 15:33
...but, not strong opinion on the topic, I can be convinced otherwise
if one can figure a good way to implement it...
João Gonçalo Rocha Cardoso
@joaocardoso
Mar 24 2017 17:22
I think pfba should return that. The objective of pfba is to to minimize the sum of the fluxes. We can discuss that topic sometime, maybe make an issue.
Peter St. John
@pstjohn
Mar 24 2017 18:44
That’s true, but the model objective gets reset back to the original objective after the call to pfba. I’m also not super picky as to which way we do it, I just wasn’t sure if the returned value was an oversight or concious decision