#11 regulates, negatively_regulates and positively_regulates

open
nobody
5
2008-01-17
2008-01-17
No

placeholder definitions included below.

Example file:
ftp://ftp.geneontology.org/pub/go/scratch/gene_ontology_with_regulates_relations_test.obo

[Typedef]
id: OBO_REL:regulates
name: regulates
def: "A relation between a process and a process or quality. A regulates B if the unfolding of A affects the frequency, rate or extent of B. A is called the regulating process, B the regulates process" []
transitive_over: OBO_REL:part_of

[Typedef]
id: OBO_REL:positively_regulates
name: positively_regulates
def: "A regulation relation in which the unfolding of the regulating process *increases* the frequency, rate or extent of the regulated process" []
is_a: OBO_REL:regulates
transitive_over: OBO_REL:part_of

[Typedef]
id: OBO_REL:negatively_regulates
name: negatively_regulates
def: "A regulation relation in which the unfolding of the regulating process *decreases* the frequency, rate or extent of the regulated process" []
is_a: OBO_REL:regulates
transitive_over: OBO_REL:part_of

Discussion

  • Chris Mungall

    Chris Mungall - 2008-01-23

    Logged In: YES
    user_id=254723
    Originator: YES

    More details:

    These relations are non-transitive (i.e. it is not the case that X reg Y reg Z => X reg Z). They are transitive over part_of (i.e. X reg Y part_of Z => X reg Z) - this can be represented in OWL1.1 using a SubObjectPropertyChain:

    <SubObjectPropertyOf>
    <SubObjectPropertyChain>
    <ObjectProperty URI="&OBO_REL;regulates"/>
    <ObjectProperty URI="&OBO_REL;part_of"/>
    </SubObjectPropertyChain>
    <ObjectProperty URI="&OBO_REL;regulates"/>
    </SubObjectPropertyOf>

     
  • Alan Ruttenberg

    Alan Ruttenberg - 2008-01-23

    Logged In: YES
    user_id=138355
    Originator: NO

    I think these definitions need to be made in term of more primitive relations, probably by making reference to what happens to the participants. Unfolding isn't much to hang on - pretty much synonymous with "happens"

    I think more work needs to be done to nail down what these mean. Is "regulates" synonymous with "influences"? How do we distinguish regulation processes from other process parts?

     
  • Chris Mungall

    Chris Mungall - 2008-01-24

    Logged In: YES
    user_id=254723
    Originator: YES

    unfolding is indeed synonymous with happens. It's about as primitive as "process".

    I agree the definition should take into account participants. There should be some participant in common. It is difficult to avoid agency or causality here I think, otherwise we cannot distinguish regulation from coincident changes.

    An additional difficulty is that the definition should refer to frequency, rate or extent of a process. We can use PATO here, but PATO defines these as qualities of processes, which some regard as controversial.

     
  • Alan Ruttenberg

    Alan Ruttenberg - 2008-01-24

    Logged In: YES
    user_id=138355
    Originator: NO

    How to avoid agency or causality is exactly the issue, IMO. Agency is clearly inappropriate for molecules, and causality is pretty problematic without carefully thinking about how to talk about it in a way that doesn't admit everything in the light cone. Previous discussions include

    http://ontolog.cim3.net/forum/ontolog-forum/2007-06/msg00379.html

    http://sourceforge.net/mailarchive/forum.php?thread_name=20070518235446.281670%40gmx.net&forum_name=obo-relations

    One might try also to define regulation in some way that makes reference to evolution - Perhaps taking as evidence a sequence of organisms which show a progression from exhibiting a process that unfolds in pretty much the same way always, to one which gains participants and shows more variation in how it happens, and where that variation can be shown to confer a selective advantage.

    As you may have inferred by now, I'm not sure that the addition of these relations is an obvious step in the right direction...

     
  • Chris Mungall

    Chris Mungall - 2008-01-24

    Logged In: YES
    user_id=254723
    Originator: YES

    I think processes that arise through design can regulate too.

    At the end of the day I don't think we can avoid causality

    Are you saying that what we colloquially know as regulation exists, it's just hard to define - or that we should dispense with any notion of regulation altogether?

     
  • Aqualung

    Aqualung - 2008-01-24

    Logged In: YES
    user_id=1316559
    Originator: NO

    To "regulate" is a core cybernetic/ST notion, hence it involves the notion of feedback. At any rate, I think greater clarification can be expected by (gently) suggesting that GO developers consult a Dynamical Systems manual: it will certainly spare them the grief of having to reinvent the wheel.

    Otherwise, leaving the sneers aside, the standard construal of the feedback loop involves reference to *systems* connected in a reaction topology (positive or negative), as opposed to--as I have learned it is fashionable in the biomedical ontologies world--*processes* that regulate each other. I find it healthier to shake off this talk about processes, however entrenched/chronic it may be, and switch the focus to (talk about) *systems* that implement/run/undergo/participate in the corresponding processes. This should definitely give a clearer picture of the underlying phenomena (if anything, systems are certainly easier to pinpoint to, on paper and elsewhere), and should open the gates to a more overt adoption/import of core ST results and techniques.

    In short, while I applaud the decision to include such relations, I also deplore their process (occurrent)-centric flavor. I am also hoping that, in establishing their syntax and behavior, GO developers will try to align them with the corresponding ST notions.

     
  • Alan Ruttenberg

    Alan Ruttenberg - 2008-01-24

    Logged In: YES
    user_id=138355
    Originator: NO

    Chris:
    > I think processes that arise through design can regulate too.
    Yes. But design involves agency. However evolution sometimes seems to have agency, and that's what was in the back of my mind when I suggested making reference to evolution.

    > At the end of the day I don't think we can avoid causality

    Ok. Let's define it then. We need some definition that can be reliably used for annotation.

    > Are you saying that what we colloquially know as regulation exists, it's
    just hard to define - or that we should dispense with any notion of
    regulation altogether?

    I think there are a bunch of phenomena that are called regulation and that it is a better strategy to more explicitly represent the phenomena. Perhaps, then, regulation is defined in terms of those. However in common usage I think "regulation" is used inconsistently and that before using it in languages meant for computation we should carefully work out definitions that are workable and from which we can reliably derive inferences.

    It's sort of like the case of the word "interaction". It's pretty meaningless, outside being a conceptual term, when it is applied to both protein binding interactions and genetic interaction interactions.

    We have regulation by transcription factors, regulation of processes that depend resources in limited supply, regulation in the sense of homeostatic processes.....

    I'm not sure what definition to use that covers all of these, and which can provide useful information that isn't represented by simply representing the larger process in which both the "regulating" and the "regulated" process are both part of.

     
  • Werner Ceusters

    Werner Ceusters - 2008-01-24

    Logged In: YES
    user_id=1601734
    Originator: NO

    Chris wrote:

    The GO is planning to make the implicit regulates relations in the
    > > ontology explicit. There will be 3 new relations:
    > >
    > > regulates
    > > negatively_regulates
    > > positively_regulates

    Chris,

    Are you sure about this ?

    One of the BFO principles is that a dependent
    entity cannot depend on another dependent entity. For that reason, there can
    be no dependence relation between processes or between processes and
    qualities. If it is not a dependency relation, what is it then ?

    Furthermore, it doesn't seem to work at the instance level. If some process
    A would "change" process B, then process B ceases to exist, and another
    process would come into existence.

    Next, can processes change anything ? I would expect only continuants to be
    able to change things.

    I believe this should be tought over much better.

     

Log in to post a comment.

Get latest updates about Open Source Projects, Conferences and News.

Sign up for the SourceForge newsletter:





No, thanks