Interpreting infeasible models in singleGeneDeletion (growthRate Nan)

[Uri David Akavia]

Hi All,

I've been trying to use singleGeneDeletion on a feasible model. When some of the genes are deleted, the model becomes infeasible and can't be solved, which is kind of what I'd expect.

At that point, singleGeneDeletion (lines 103-108) sets growth rate as NaN. I've been trying to think what this means biologically - it seems to me that if a model becomes infeasible upon deleting the gene, the gene is essential. Therefore, it is a gene worth deleting in vitro.

Do you guys agree?

How do you treat genes that their deletion leads to infeasible models (and therefore NaN growth rate/ratio)?

Cheers,

Uri David

[Siu Hung Joshua Chan]

Hi Uri,

I agree with what you said. Just to add more to that. Usually a model has a non-growth-associated maintenance (NHAM) cost, represented by a positive lower bound for an ATP hydrolysis reaction (named ATPM in the BiGG convention). Infeasibility would imply that deleting the gene will cause the model unable to generate ATP, if there are no other reactions with +ve lower bound or -ve upper bound. If you get growth = 0 but still feasible, that would mean some biomass precursors cannot be produced but ATP can still be produced.

Best,
Joshua