Problem with FBA in network evaluation&debugging

[Emily Schina]

Dear all hello!

I am creating a model( I have constructed it in excel) and recently loaded it into Matlab following the Protocol of Thiele and Palsson.I am currently trying to test if biomass precursors can be produced in standard medium and for each biomass component i perform the test:

model=changeObjective(model,'DM'); where DM is the demand reaction of the component

FBAsolution = optimizeCbModel(model,'max')

The results I am getting are:

FBAsolution = 

           f: 0

           x: []

        stat: 0

    origStat: 110

      solver: 'glpk'

        time: 0.0600

I have read previous posts on similar issues and the papers suggested but haven't found what is happening. What confuses me most is that x(the flux value for reaction in the model) is []. It is infeasible and I don't understand why.I tested the function with a similar model(with less reactions) and although I get f=0 for some components,x exists.I based my model on another model so I reconstructed it in the same way and also created the metabolic maps to make sure each biomass metabolite is produced.

I would be very grateful if someone could help me!

Thank you in advance,

Emily

[Hulda S. Haraldsdóttir]

Dear Emily,

The most likely reason is that your flux bounds (model.lb and model.ub) are inconsistent. Consider the simple network -> A -> B ->. If you constrain A -> B to have negative flux (e.g. between -1 and -10) and constrain B -> to have positive flux (e.g. between 1 and 10), then your problem will be infeasible since you can't satisfy both constraints at steady state. To test whether your problem is caused by inconsistent flux bounds, you can try setting the lower bound to -1000 and the upper bound to 1000 for all your reactions and see if you get a feasible solution.

Best,

Hulda

--

Hulda S. Haraldsdóttir, Ph.D.

-----------------------------------------------------------------

Research associate,

Systems Biochemistry Group,

Luxembourg Centre for Systems Biomedicine,

University of Luxembourg,

Campus Belval,

7, avenue des Hauts-Fourneaux,

Esch-sur-Alzette,

Luxembourg,

L-4362.

-----------------------------

Mobile:  +352 661 228 191

Office: +352 466 644 6228

wwwen.uni.lu/lcsb/people/hulda_haraldsdottir

wwwen.uni.lu/lcsb/research/systems_biochemistry

-----------------------------------------------------------------

---

Date: Sat, 28 Mar 2015 00:22:17 -0700
From: emil_pie...@hotmail.com
To: cobra-...@googlegroups.com
Subject: Problem with FBA in network evaluation&debugging

--

---
You received this message because you are subscribed to the Google Groups "COBRA Toolbox" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cobra-toolbo...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Emily Schina]

Dear Hulda!

Thank you very much for you answer! I've already set  the bounds for irreversible reactions to 0,1000 (lb and ub respectively) and for reversible reactions -1000,1000.I will try changing all lower bounds to -1000 to see if something changes.Is it possible that many intracellular transport reactions that I added(which are all reversible) caused this?

Best regards,

Emily

[Emily Schina]

I'tried what you suggested and it worked as far as solution.x is concerned.It still gives me f=0 but I think that has to do with the connectivity of the network.Then I tried applying my simulation constraints for autotrophic conditions,for example for biomass reaction model=changeRxnBounds(model,'BIO_ls',0,01375,'b'); then run it again for the first demand I have and it gives me x=[].

Thank you again, if you have any suggestion it would be amazing!

Best,

Emily

[Emily Schina]

After many efforts and tests ,which included building biomass reaction from beggining and deleting each biomass metabolite I've found out that the problem is with one specific metabolite-eicosapentaenoic acid.When I delete this metabolite from biomass reaction FBA runs okay and I get results.When I add it,flux is zero.Does anyone have any idea why this happens? I need to solve this issue because this metabolite is a major biomass component for my model and specifically test it's production in different conditions and I'm stuck!

Thank you in advance,

Emily  :)

[Emily Schina]

Hoorey people!

I solved it, so I am posting in case i can help someone else!I used a tool called Structural Analysis Tool for Biochemical Networks.In this tool you can load your model in .xml and then  press to see summary.It has a list of independent and dependent entities in your model.The problematic ones are those in dependent entities. This tool lists the problematic metabolites with which other problematic ones are connected.For example I had a problem with a fatty acid and it was listed with other metabolites from biotin biosynthesis. From what I realised generic terms create huge problems,this was my problem, so I replaced them with non generic or made them boundary species or deleted them.
I am new to this tool so maybe I didn't explain well, but there are papers that can help understand better.

Best regards,

Emily