search

pnnl-predictive-phenomics / eflux

Other
1 stars 0 forks source link

Zero fluxes in Eflux #29

Open djinnome opened 2 months ago

djinnome commented 2 months ago

Hi @djinnome ,

I'm trying to set up Eflux2 for the putida model.

I ran FBA on the model and got rid of the reactions with zero fluxes. However, when I ran FBA on the model whose zero-flux reactions had been removed, there were more reactions whose fluxes were zero.

Is this normal? This is throwing me off when setting up Eflux2 because, as you know, Eflux2 gives v_star values, none of which can be zero.

Here is the notebook I was using to make the new putida model. There is a section called "Removing 0-flux reactions"

Thank you,

@janisshin

janisshin commented 2 months ago

I've figured out that it's because I was creating the model with one conda env and then running eflux2 with another conda env. Running the command model.optimize() yields different results between the two envs.

Weirdly enough, in both conda envs, I have the same version of cobra, 0.29.0, but different versions of numpy, 1.24.0 vs. 1.23.4 It is not dependent on the use of model.solver = 'gurobi'