How to associate enzymatic reactoins with metabolic pathways?

[0m1n0]

Dear Biolink team,

Let me start by thanking all your colleagues who do such a wonderful job. I work on metabolic pathways and that's why I'm very interested in Biolink Model, and I considered it a brilliant approach!

I'm trying to understand how to link Biolink Model's vocabularies to the different participants in metabolic pathways.

We have a pathway PWY containing two successive reactions RXN1 (CMP1 as substrate, CMP2 as product) and RXN2 (CMP2 as substrate, CMP3 as product) :
CMP1 -> RXN1 -> CMP2 -> RXN2 -> CMP3

# URIs of pathway, reaction, compound are simple examples
@prefix bl: <https://w3id.org/biolink/vocab/> .
@prefix comp: <http://example.org/comp/> .
@prefix path: <http://example.org/pathway> .
@prefix reaction: <http://example.org/reaction/> .

path:PWY bl:category bl:Pathway .

reaction:RXN1 bl:category bl:MolecularActivity ;
    bl:has_input comp:CMP1 ;
    bl:has_output comp:CMP2 .

reaction:RXN2 bl:category bl:MolecularActivity ;
    bl:has_input comp:CMP2 ;
    bl:has_output comp:CMP3 .

comp:CMP1 bl:category bl:SmallMolecule .
comp:CMP2 bl:category bl:SmallMolecule .
comp:CMP3 bl:category bl:SmallMolecule .

Now I want to link RXN1 and RXN2 to PWY and I saw that several associations are proposed:

• Class: GeneToPathwayAssociation: between gene coding enzyme and pathway?
• Class: ChemicalToPathwayAssociation: between pathway and chemical entity including enzyme, metabolites (substrate/product), cofactors?
• Class: MacromolecularMachineToBiologicalProcessAssociation
• Class: MacromolecularMachineToMolecularActivityAssociation
• Relation: in_pathway_with: only between enzymes (unspecified pathway)?
• Class: ReactionToCatalystAssociation: this one seems the most relevant to link the enzyme and the reaction since it can contain attributes such as reaction direction

All these associations do not directly link reactions to pathways: we have to go down to the molecular level (gene or enzyme).

Is there any association of type <reaction> bl:in_pathway <pathway>?

Or should I use generic vocabularies like Relation: participates_in or Relation: actively_involved_in?

I apologize for overwhelming you with so many questions.
Your answers will greatly help me.

Regards,
Min

[nlharris]

Hi Min, thank you for your interest in Biolink Model! We will get back to you soon with some answers. You happen to have contacted us during a week when we have an all-week all-hands meeting for the Translator project, so all the Biolink Model architects and developers are super-busy right now. Hopefully someone will get back to you next week.

[sierra-moxon]

Hi Min, thanks so much for reaching out and for such a well crafted discussion topic! Adding a biolink:MolecularActivity to biolink:Pathway association makes sense to me. PR: #1036 edit: spoke to Chris M offline, and he agrees. Let's add it! :)

[0m1n0]

Thank you very much for your quick answers despite being very busy. And I am glad you are going to add this "reaction-pathway" association (can't wait!).

Let me ask you one more questions.
I would like to group the pathways into super-pathways. I have seen that there is Relation: superclass_of but not all pathways participate in super-pathways. So it would be better to use Relation: subclass_of in the other direction? Or do you have any other advice?

[cmungall]

do your superpathways have the same semantics as biocyc superpathways?

…

On Thu, Jun 9, 2022 at 5:29 AM Min ***@***.***> wrote: Thank you very much for your quick answers despite being very busy. And I am glad you are going to add this "reaction-pathway" association (can't wait!). Let me ask you one more questions. I would like to group the pathways into super-pathways. I have seen that there is Relation: superclass_of <https://biolink.github.io/biolink-model/docs/superclass_of.html#relation-superclass_of> but not all pathways participate in super-pathways. So it would be better to use Relation: subclass_of <https://biolink.github.io/biolink-model/docs/subclass_of.html#relation-subclass_of> in the other direction? Or do you have any other advice? — Reply to this email directly, view it on GitHub <#1035 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAMMOPO3SGFQH5QZAVOMNLVOHPRFANCNFSM5YGJAGVQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.*** com>

[0m1n0]

@cmungall Indeed, you have caught the idea. But I would especially like to include bigger superpathways such as KEGG pathway maps (e.g. Carbohydrate metabolism, Energy metabolism... etc). I should probably represent them at different levels, i.e. BioCyc like superpathway level and KEGG like superpathway level. Possibly other semantics for grouping pathways will come up in the future.