#95 Sideproduct and sidesubstrate

open
Nick Juty
new term (59)
3
2013-05-15
2013-02-12
Andreas Dräger
No

In order to better support the SBML layout extension, it would be helpful to have specialized SBO terms to indicate the roles of reaction participants by introducing SBO terms for "sideproduct" and "sidesubstrate. See the specification of the SBML Layout extension for details (http://otto.bioquant.uni-heidelberg.de/sbml/level2/20050425/SBMLLayoutExtension-20050425.pdf)

Discussion

  • Nick Juty
    Nick Juty
    2013-02-13

    thanks for the suggestion - will check specification for more info..

     
  • Nick Juty
    Nick Juty
    2013-02-13

    • assigned_to: nobody --> njuty
     
  • Nick Juty
    Nick Juty
    2013-02-13

    Relevant snippets from Layout spec':
    "The role attribute is used to specify how the species reference should be displayed. Allowed values are substrate, product, sidesubstrate, sideproduct, ..."

    ".. sidesubstrate and sideproduct are used for stuff like ATP, NAD+, etc. that some renderers might choose to display as side reactions...."

    alternative name ideas: co-product, by-product; co-substrate,

     
  • This makes no sense to me. Those sidestuff do not correspond to anything in SBML. Either they are part of the reaction (reactant, product, modifier) or they are just annotation. I do not even understand how people would use them at all. I would like to see examples.

    When it comes to SBO, maybe we can create something like "non-essential reactant" or seomthing like that? Weird though.

     
  • Oh, it does make sense for how to render a reaction arrow: With the help of the "sidestuff" it is possible to determine the main axis of a reaction. Consider, for instance, the reaction Fructose + ATP -> ADP + Fructose-1-phosphate. With the help of sidesubstrate and sideproduct it can easily be specified that the main arrow goes directly from Fructose to Fructose-1-phosphate, whereas there could be a halfpipe arrow connecting the reaction with ATP and ADP. Particularly, in the case that only the bounding box of a speciesReferenceGlyph is specified (and not its curve), this information is crucial for a rendering tool.

     
  • Pablo Moreno
    Pablo Moreno
    2013-04-17

    I will like to support Andreas petition. As we spoke the other day with Nicolas and Nick before Prof. Schrieber's talk, we need such a sboterm to be able to mark currencies and chemical compounds that are not in the main flow of a reaction. Nicolas suggested to call the sbo term "currency", however, there are cases in which small molecules that are not normally considered currencies need to be treated as side compounds (that is cloned/aliased and left "to the side" of the reaction). Examples could be glutamate, protons (well, one could argue that glutamate could be considered a nitrogen currency and protons a redox currency), or water. For that reason I would rather have something as main compound or side compound, as Andreas suggests. Still, currency is better than having nothing. Currently I'm doing something very dirty to pass this information to the layout algorithm (in the model annotation):

    <sidecompounds xmlns="http://sbml.org/annotations/sidecompounds">
    <compound id="WID12048871_DWID86_c"></compound>
    <compound id="WID12050463_DWID86_c"></compound>
    <compound id="WID12050890_DWID86_c"></compound>
    <compound id="WID12053085_DWID86_c"></compound>
    <compound id="WID12050461_DWID86_c"></compound>
    <compound id="WID12048877_DWID86_c"></compound>
    <compound id="WID12054304_DWID86_c"></compound>
    <compound id="WID12050453_DWID86_c"></compound>
    <compound id="WID12053839_DWID86_c"></compound>
    <compound id="WID12050451_DWID86_c"></compound>
    </sidecompounds>

    So I hope that you can agree on a suitable sboterm to avoid this :-).

    Thanks!
    Pablo

     
  • Hi Pablo and all others,

    Thanks for your answer. There is still one more problem: In the SBML Layout extension it is possible to define reaction glyphs for which no SBML core reaction exists. In this case, the role of the element should also tell if it is a reactant or a product. If I correctly understand your current proposal, there would be just one term, say "sidecompound" for both (reactant and product). I would like to suggest that this sidecompound term would have two children; one for reactant, one for product.

    Cheers and thanks
    Andreas

     
  • Nick Juty
    Nick Juty
    2013-04-18

    Thanks for the feedback.

    I think we could place the appropriate terms in the tree here (yyy):

    SBO:0000236 - physical entity representation
    SBO:0000240 - material entity
    SBO:0000247 - simple chemical
    SBO:0000xxx - yyy

    The name is harder to come up with. These 'currency compounds' are not really 'side products' as they are necessary for the main reaction, but just are not the focus of the human interpretation. I thought about it a bit and my best suggestion is 'ancilliary' compound/substrate/product. This implies it is 'necessary to support the primary activity or operation'[1]. (I discounted auxiliary as it implies something that is not necessary and can be provided supplementally).

    Let me know what you think....

    1. http://oxforddictionaries.com/definition/english/ancillary?view=uk

     
  • I think this is perfectly justified. What I am unsure about is the use of this SBO term? Where would it be used? Presumably it would not be the value of an sboTerm attribute on species? Would-it annotate a Layout construct?

     
  • OK, I realise now that there is an easy solution. They should be children of

    SBO:0000015 - substrate
    and
    SBO:0000011 - product

     
  • Nick Juty
    Nick Juty
    2013-04-29

    I'll create these once we decide on a name. Should we go with 'side' product and [reactant|substrate]? Or 'ancillary' product and [reactant|substrate]?

     
  • Pablo Moreno
    Pablo Moreno
    2013-05-09

    I would vote for side product / side reactant names. I think reactant is more widely used in APIs/DBs than substrate (SBML, KEGG, and BioCyc they all use reactant generally instead of substrate).

     
  • Pablo Moreno
    Pablo Moreno
    2013-05-09

    This by-product and side product distinction is widely used in the chemical engineerin/process world. I agree that from a chemical engineering point of view by-product would be better suited for what we are referring to in the products part. What I struggle with however is that, to the best of my knowledge, there is no such a thing as a by-reactant (I am ~70% chemical engineer ;-) ).

     
  • Well, we can create the terms ... my fear is that using side-X, we will trigger confusion.

     
  • Pablo Moreno
    Pablo Moreno
    2013-05-09

    When you design ontologies, do you do want to have "single-responsibility" for your leaf terms? Or that doesn't really matters?

    Having a "side-product" term means that the term has two responsibilities: saying that it is a product and it is secondary/side/ancilliary. Shouldn't we have a term that, independent of whether something is product or reactant, denotes that the compound is not considered a main/central to the reaction? This would also mean that you would create only one additional term instead of two.

     
  • That leads to multiple parentage, which is frown upon. The preferred course of action is to choose the main criteria and then subclasses. In the case of SBO, we only use classes, and not properties, so this sometimes triggers a multiplication of terms. In better structured ontology, we would have terms refined with properties.

    Here, since we do not need to have "primary product" and "primary reactant", we can just subclass the secondary ones.

    BTW I think "secondary" is a winner adjective. "secondary product" will have "by product" as a synonym.

     
  • Pablo Moreno
    Pablo Moreno
    2013-05-09

    I agree that the side word might cause confusion. In chemical synthesis, a side reaction is an undesired reaction that happens in parallel to the desired set of reactions.

    I like secondary, but it might produce confusion as well due to its use for secondary metabolite or secondary metabolism.

     
  • Nick Juty
    Nick Juty
    2013-05-10

    It's hard isn't it!? With ancillary I was trying to avoid the connotations that we are encountering now. I think we just need to decide on the names, and make sure that the definition makes clear what our intent is for their use.

     
  • Nick Juty
    Nick Juty
    2013-05-15

    I've created the terms as there is a clear need for them:
    SBO:0000604: side substrate
    SBO:0000603: side product

    We can rename them and refine definitions as necessary when/if we conclude this discussion.
    Thanks all for your continued input!

     
  • Nick Juty
    Nick Juty
    2013-05-15

    • priority: 5 --> 3