[Ontology-editors] DASH complex - two part_of parents

[Valerie Wood]

Hi Chris,

The problem appears to be caused by  the use of "microtubule associated 
complex" for complexes which are clearly sub-complexes of larger units 
(e.g. kinetochore). Taking this to its logical conclusion "kinetochore" 
would also be  a "microtubule associated complex" .

I'm trying to think of an analogy. Sorry this is a really rubbish one, 
but if the DASH complex was a tow bar on a car (a kinetochore)
and you attached a trailer (spindle cytoskeleton) you wouldn't say that 
the tow bar or the car were part of the trailer.

Intuitively you might say that DASH was a microtubule associated complex 
but maybe there  does need to be some  additional  to the def that a 
microtubule associated complex in GO, based on its position as a child 
of microtubule cytoskleleton, that it  must be associated *primarily* 
with the microtubule and not be an intrinsic  part of some other large 
cellular body which is NOT part of the cytoskeleton ?

The fact that DASH subunits associate with microtubules at some level, I 
would capture by binding annotations (if this was shown experimentally), 
and the process of "kinetochore-microtubule attachment", but I wouldn't 
expect this association to be hardwired into the ontology.

Sorry this is thinking whist typing. I just read the rest of your 
comments and I didn't fully understand the making MF annotatiosn as 
proxies for CC annotations

As an aside, at the jamboree yesterday a number of us were talking about 
whether we should get rid of colocalizes_with.
I recently evaluated my annotations and reduced for 200 to about 50, I 
am sure all of the remaining annotations could be better captured with a 
different or less specific annotation, but they need some more work.


VAl







Chris Mungall wrote:

>
> On Feb 23, 2009, at 6:50 AM, Valerie Wood wrote:
>
>> I'm still not sure about this.
>>
>> The DASH complex is usually considered part of the kinetochore. It  
>> associated with spindle microtubules, when they attach to the  
>> kinetochore..
>> I'm not sure it is correct to position under 'spindle microtubule'  
>> as this would imply it was a constituent of the microtubule.
>> According to the ref genome discussion recently microtubule should  
>> only be used to annotatate tubulins, and other things which  
>> assoicate with the microtubule should be annotated with microtubule  
>> binding, or the appropriate microtubule cytoskeleton component term.
>
>
> Should this decision not be reflected in the comment tag? Seems like 
> a  very specific rule.
>
> Would colocalizes_with to microtubule be allowed for non-tubulins?
>
> When we have the MF->CC links in the ontology we will be able to make  
> inferences of the form: a gene product that is executing a function 
> of  X binding must, at the least, be in the vicinity of some X. Here 
> I  would think vicinity_of would translate to colocalizes_with, and 
> would  be treated appropriately in AmiGO
>
> I think it's especially important to be thinking about this if people  
> are making MF annotations as proxies for CC annotations
>
>>
>> In this case I don't think  the DASH complex would be considered as  
>> part of the microtubule, or  microtubule cytoskeleton?
>>
>> I'm also not sure that the 2 children microtubule associated DASH  
>> complex and kinetochore associated DASH complex are required?
>> As far as I'm aware DASH complex is always associated with the  
>> kinetochore when it exists. When the spindles attach it is also  
>> associated with the microtubules. Is there a precedent for having  
>> separate componet terms in this situation when we are referring to  
>> an identical complex ? (It seems  the terms are being used to  
>> describe temproal aspects of the complexes interactions)
>>
>> VAl
>>
>> Midori Harris wrote:
>>
>>> Hi,
>>>
>>> This is just to record that Val and I have looked at the literature  
>>> on the DASH complex (GO:0042729), and agreed that it's OK for it to  
>>> have two part_of parents for now.
>>>
>>> Background: the DASH complex term was originally added with two  
>>> part_of parents, GO:0000778 condensed nuclear chromosome  
>>> kinetochore and GO:0005828 kinetochore microtubule. It was duly  
>>> flagged by Chris' QC report, and at a meeting involving Jane, David  
>>> and me, Jane created two
>>> child terms to avoid true path violations. After that work we had:
>>>
>>>  DASH complex GO:0042729
>>>  [i] kinetochore-associated DASH complex GO:0043926
>>>  [i] microtubule-associated DASH complex GO:0043925
>>>
>>>  condensed nuclear chromosome kinetochore
>>>  [p] kinetochore-associated DASH complex GO:0043926
>>>
>>>  kinetochore microtubule
>>>  [p] microtubule-associated DASH complex GO:0043925
>>>
>>> But Val and I confirmed that, as far as is known in the field,  
>>> every DASH complex that forms is part of a kinetochore and  
>>> associated with kinetochore microtubules. I think using a different  
>>> relationship type, perhaps associated_with or connected_to or  
>>> adjacent_to, between DASH complex and kinetochore microtubule would  
>>> improve our representation of the biology, but until we add new  
>>> relationship types, allowing the original DASH complex term to have  
>>> two part_of parents is the most biologically accurate thing we can  do.
>>>
>>> Accordingly, I've merged the newer terms into the parent, restoring  
>>> the part_of relationships (and removing the newer term names from  
>>> the synonym list -- looks like nobody uses them).
>>>
>>> midori
>>>
>>>
>>>
>>
>>
>>
>> -- 
>> The Wellcome Trust Sanger Institute is operated by Genome Research  
>> Limited, a charity registered in England with number 1021457 and a  
>> company registered in England with number 2742969, whose registered  
>> office is 215 Euston Road, London, NW1 2BE.  
>> _______________________________________________
>> Ontology-editors mailing list
>> Ontology-editors at geneontology.org
>> http://fafner.stanford.edu/mailman/listinfo/ontology-editors
>>
>
>
>
>



-- 
 The Wellcome Trust Sanger Institute is operated by Genome Research 
 Limited, a charity registered in England with number 1021457 and a 
 company registered in England with number 2742969, whose registered 
 office is 215 Euston Road, London, NW1 2BE.