FASTCORE finds model inconsistent after correction with FASTCC

[Peter Großmann]

Dear COBRA team,

i am trying to build diet specific metabolic models with FASTCORE on Recon2.2. I implement the diet by setting the lb of some exchange reactions to a fixed value, say -1000, and i set the lb of all other exchange reactions to zero. Then i run FASTCC on it and reduce the model to the unblocked reactions and i use transcriptome data and the GPRs to determine which reactions should be active.

For a relatively large number of my cases, FASTCORE reports an inconsistent input model, although FASTCC does not remove any more reactions when i rerun it on the presumably consistent model. The tolerance for both methods is set to 1e-4.

The error is: fastcore.m Error: Inconsistent irreversible core reactions.
In many cases i also get the following warning:

0 = sol.stat
INFEASIBLE = sol.origStat
Warning: LP solution may not be optimal
> In LP9 at 56
  In findSparseMode at 46
  In fastcore at 68


I have attached an example data set that causes this error on my system, and the system report. Note that i deliberately deactivated Java to accelerate the startup in no-GUI matlab. The issue persists also with Java enabled. My Gurobi version is 7.5.2.

Please tell me what i can change to make that work!

Best
Peter

[Ronan M.T. Fleming]

Hi Peter,

I have seen previously in perhaps a similar situation that the numerical tolerance on what one considers zero or nonzero flux (default 1e-4) needs to be reduced, e.g., to 1e-6, (sometimes 1e-8 with a poorly scaled model and a quad precision solver) when a model is numerically ill scaled.

These tutorials might help:
https://github.com/opencobra/COBRA.tutorials/blob/master/reconstruction/recon2FBAmodel/
esp findFluxConsistentSubset.m

Check the scaling of your model:
https://github.com/opencobra/COBRA.tutorials/tree/master/reconstruction/numCharact

Regards,

Ronan

--

---
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-toolbox+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

--
Mr. Ronan MT Fleming B.V.M.S. Dip. Math. Ph.D.
----------------------------------------------------------------------------
Senior research associate (EN) == Chercheur (FR),
Principal investigator,
Systems Biochemistry Group,
wwwen.uni.lu/lcsb/research/systems_biochemistry
Luxembourg Centre for Systems Biomedicine,
University of Luxembourg,

Campus Belval,
6, avenue du Swing,
L-4367 Belvaux.

&
National Centre of Excellence in Research on Parkinson’s disease
www.parkinson.lu

&

Adjunct Assistant Professor,

Division of Analytical Biosciences,

Leiden Academic Centre for Drug Research,

Faculty of Science,

University of Leiden.

http://analyticalbiosciences.leidenuniv.nl
----------------------------------------------------------------------------
Mobile:  +352 621 175 112
Office: +352 466 644 5528
Skype: ronan.fleming
----------------------------------------------------------------------------
(This message is confidential and may contain privileged information. It is intended for the named recipient only. If you receive it in error please notify me and permanently delete the original message and any copies.)