deleteModelGenes command not effective

[Prashant Pokhriyal]

The documentation ( https://opencobra.github.io/cobratoolbox/latest/modules/deletion.html ) says that 

deleteModelGenes(model, geneList, downRegFraction)[source]

Deletes one or more genes and constrain the reactions affected to zero and appends ‘_deleted’ to the gene(s)

Usage

[model, hasEffect, constrRxnNames, deletedGenes] = deleteModelGenes(model, geneList, downRegFraction)

Input

• model – COBRA model with the appropriate constrains for a particular condition

Optional inputs

• geneList – List of genes to be deleted (Default = all genes in model)
• downRegFraction – Fraction of the original bounds that the reactions corresponding to downregulated genes will be assigned (Default = 0 corresponding to a full deletion)

Outputs

• model – COBRA model with the selected genes deleted
• hasEffect – True if the gene deletion has an effect on the model
• constrRxnNames – Reactions that are associated to the genes in geneList
• deletedGenes – The list of genes removed from the model.

 So I just try it for the Ecoli_core model-

Input-

>> load('ecoli_textbook.mat');

>> geneList={'b3916'};  % Here I have selected the gene coding for phosphofructokinase (pfkA), which is one of the essential genes in the pathway.

>> [model, hasEffect, constrRxnNames, deletedGenes] = deleteModelGenes(model, geneList, 0)

Output-

model = 

        modelVersion: [1x1 struct]

                rxns: {95x1 cell}

                mets: {72x1 cell}

                   S: [72x95 double]

                 rev: [95x1 double]

                   c: [95x1 double]

            metNames: {72x1 cell}

         metFormulas: {72x1 cell}

                  lb: [95x1 double]

                  ub: [95x1 double]

           metCharge: [72x1 int32]

               rules: {95x1 cell}

               genes: {137x1 cell}

          rxnGeneMat: [95x137 double]

             grRules: {95x1 cell}

          subSystems: {95x1 cell}

    confidenceScores: {95x1 cell}

       rxnReferences: {95x1 cell}

        rxnECNumbers: {95x1 cell}

            rxnNotes: {95x1 cell}

            rxnNames: {95x1 cell}

          metChEBIID: {72x1 cell}

           metKEGGID: {72x1 cell}

        metPubChemID: {72x1 cell}

      metInChIString: {72x1 cell}

                   b: [72x1 double]

         description: 'ecoli_textbook.xml'

hasEffect =

     0

constrRxnNames = 

     {}

deletedGenes = 

    'b3916'

Now according to the documentation, the constrRxnNames should account for the deleted reaction associated with the gene, but it is showing an empty string. Also, the hasEffect output also shows that there is no overall effect on the model because of the gene deletion.

I have tried it for other models and genes also but every time it shows the above result. Even other similar commands like singleGeneDeletion and doubleGeneDeletion are showing the erroneous results regarding the deletion of the reaction.

As I am trying to work on a mutated strain, it may pose problems in the model predictions.

So, kindly help me in this regard.

Thanks and Regards,

Prashant Pokhriyal

Institute of Chemical Technology, Mumbai

pokhr...@gmail.com

[Thomas Pfau]

Hello Prashant,

The example you give has no effect, as PFK (the reaction associated with b3916) has the following GPR:

(b3916) or (b1723)

i.e. either b3916 or b1723 are sufficient for the reaction. Thus if you only delete one of them, nothing will change.
It is not unlikely, that b3916 is the gene coding for the enzyme that is commonly catalysing this reaction, and that the contribution of b1723 is minimal. However, the model does not know this distinction. IF you want to make sure, that only b3916 can catalyse for PFK, you will have to alter the model (and I assume, the same applies for other models as well). However, this is likely to not be precise neither.

One option I would suggest is to manually reduce the allowed flux through the PFK reaction by reducing the upper bound on the reaction via changeRxnBounds, to reflect the reduced activity.

Best

Thomas

--

---
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.

-- 
Université du Luxembourg
Faculté des Sciences, de la Technologie et de la Communication
Campus Belval, Biotech II 115
6 avenue du Swing
L-4367 Belvaux
Tel: (+352) 46 66 44 5309
Email: thoma...@uni.lu