feat: new reactions based on latest gene annotation

[hongzhonglu]

@edkerk @BenjaSanchez , Dear both, in next phase, I want to focus on adding new GPRs into model. The followed is the detailed step:

1. Add new GPRs by merge of ise926, yeast7.6 and ito980;
2. Add new GPRs based on the yeast models and gene annotation in reactome, biocyc ,kegg, SGD and uniprotKB;

[BenjaSanchez]

@hongzhonglu looks good to me. Some additional comments:

• I would advise that the issues we create address more specific questions; see for instance issue fix: existing geneRules based on iSce926 #13 as it is specific to add GPRs from one specific model. This is easy to accomplish and then check from the list. On the other hand, an issue like this one (feat: new reactions based on latest gene annotation #38) might take long and it is a bit undetermined how to define that we have reached the goal (e.g. points 4 & 5 are quite generic and would be hard to call them "solved" once we address them, as new data could become available in the future).
• Also related to this, I would advise as you have already addressed a lot of other issues, that you start to send me pull requests to update the model so we can close said issues. For instance, we should already be able to close issues fix: subSystems field #11, fix: add missing annotation #18 and fix: metabolite annotation information #37 with the work you have done. Do let me know if you have any questions or if you need specific help with the pull requests (size of each request, format, etc).

[hongzhonglu]

@BenjaSanchez Very good suggestion! For issues #11, #18, #37, I have nearly completed them. I will share with you later. For issue #38, it mainly used to improve model coverage. To make the update clear, I put these items together, so that everyone could understand it clearly. Let us do it step by step.

[hongzhonglu]

@feiranl @BenjaSanchez @edkerk
Dears,
These days, I just further check the new gene-protein-reactions which should be added into the model. I find some interesting things, for example,
As for gene "YBR136W", it has EC number 2.7.11.1, but this EC number could be connected with so many reactions (https://www.brenda-enzymes.org/enzyme.php?ecno=2.7.11.1#NATURAL SUBSTRATES, the same result can also be found in MetaNetX). Then it will become difficult for us to choose. How do you think of this question?
Thanks in advance!

[edkerk]

EC 2.7.11.1 is a protein kinase, and these kind of enzymes would fall outside of the scope of a purely metabolic model (which is what Yeast7 is and Yeast8 still will be?). So you can leave out GPRs where the reaction is a protein-modifying reaction (kinase, phosphatase, ubiquitine ligase, etc.).

Any other questionable ones?

[hongzhonglu]

Great thanks! Yeast8 will be a purely metabolic model, so we can excluded some genes. Will let you know if I find other doubted genes.

[hongzhonglu]

@BenjaSanchez @edkerk @feiranl @zhengmingzhu
In next step, we will merge the new GPRs into the yeast model. So how we add them? The followed code is from yeast-ME model. I think that we can refer it.
https://github.com/hongzhonglu/Yeast-ME-GEM/blob/master/ComplementaryScripts/Reconstruction-Laptop/addNewRxns.m
Do you have any other suggestion?

[elsemman]

@BenjaSanchez @edkerk @feiranl @zhengmingzhu
This function is from RAVEN toolbox.

[hongzhonglu]

@elsemman Thanks for your remind. So we can use it directly.

[hongzhonglu]

@ALL, this issue will be focus on adding new GPRs based on model merge and latest genome annotation as @BenjaSanchez suggested. So it will be clear for us to tack it.

[edkerk]

As @elsemman mentioned, addRxns from RAVEN can be used for this. Please use release 2.0.0-rc.1 A few points:

• The model should be in RAVEN-format in Matlab. Therefore, make sure you load the model using importModel, not readCbModel from Cobra. Alternatively, ravenCobraWrapper.m can be used to convert Cobra model to RAVEN model in Matlab, but using importModel would be preferred.
• New genes should first be added with addGenes.m.
• New metabolites should preferably be added using addMets.m, instead of allowing addRxns.m to add new metabolites.
• If you have the reactions already completely ready in one model structure, you can also 'copy' reactions, genes and metabolites together using addRxnsGenesMets.m, but please read the note on when this function can be used (concerning metabolite matching).

[hongzhonglu]

@edkerk Thanks for the detailed comments! Presently, I have evaluate each reaction connected with new gene but not for each metabolite in these reactions. We can further check these metabolites and screen the new metabolites compared with present model.

[BenjaSanchez]

@hongzhonglu what do you mean here by "model merge"? I thought that was already included in issue #13?

[BenjaSanchez]

@hongzhonglu also the part of "genome annotation" I thought was already covered by issue #60. So it looks that this issue is redundant. If so could we close it?

[hongzhonglu]

@BenjaSanchez This issue can be focused on addition of new GPRs based on the 5 different database! This part is on going now.

[BenjaSanchez]

@hongzhonglu I see, then this issue focuses on adding whereas issue #60 focuses on correcting. I have renamed both issues to reflect this, just make sure of committing separately when you address each one of the issues. Cheers!

[BenjaSanchez]

update: PR #142 added all new rxns to devel based on annotation from 5 different databases. This issue will hence be closed on the next release.

[BenjaSanchez]

As discussed in #149, even though PR #142 added all new genes + rxns associated to them, it did not add new rxns connected to the old genes already present in the model. This issue will therefore be reopened until we include those.

[edkerk]

Solved in Yeast8