Robustness Analysis

[Neha Rohatgi]

Hi all,

I did robustness analysis to see the effect of uptake of a particular metabolite on objective function. The objective function firsts increases, becomes constant and then at some point drops to zero. I want to check what are the exchanges that cause this drop or what are the limiting factors. I thought I would change the bounds on all exchanges one by one and see the effect on objective function. I was wondering if there is a easier way to do this.

Thank you

 :)

[Brandon Barker]

Palsson's 1994 mentions shadowprices in this context, which may be interesting to you - it is a fast method for checking this. I believe the cobra solver returns these as the variable y:

>> optimizeCbModel(rec2)

ans =

           x: [7440x1 double]

           f: 3.1981

           y: [5063x1 double]

           w: []

        stat: 1

    origStat: -99

      solver: 'gurobi5'

        time: 0.3528

You are probably seeing things go to zero because the model is not able to sink all that flux you are requiring it to uptake, so you wind up with infeasible constraints. Although I would not necessarily interpret that forced uptake as being biologically lethal, it probably isn't usually healthy.

[Neha Rohatgi]

Thanks Brandon. :)