geneontology / go-ontology

Source ontology files for the Gene Ontology
http://geneontology.org/page/download-ontology
Creative Commons Attribution 4.0 International
223 stars 40 forks source link

NTR: kinetochore/centromere assembly complex #11541

Closed gocentral closed 7 years ago

gocentral commented 9 years ago

Hi Paola,

Can you please create the following term as per our email discussion:

centromere assembly complex

Def: A protein complex essential for the assembly of the kinetochore on the centromeric region of the chromosome.

Relationships: is_a GO:0043234 protein complex capable_of_part_of GO:0034508 ! centromere complex assembly capable_of_part_of GO:1903099 ! positive regulation of CENP-A containing nucleosome assembly part_of GO:0000775 ! chromosome, centromeric region

Then, please make GO:1990568 MIS18 complex a direct child of the new term. Also adjust the relationships of MIS18 complex as discussed (remove capable_of_part_of GO:0034508 ! centromere complex assembly)

Thanks, Birgit

Reported by: bmeldal

Original Ticket: geneontology/ontology-requests/11368

gocentral commented 9 years ago

I think this GO:1990568 MIS18 complex is the same thing as

Term InformationAncestor ChartChild TermsProtein AnnotationCo-occurring TermsChange Log ID GO:0098654 Name CENP-A recruiting complex Ontology Cellular Component Definition A protein complex that includes Mis16(Yippee family) and Mis18 (WD repeat) subunits that is involved in the deposition of centromere specific (CENP-A containing) nucleosomes at the centromere. PMID:24774534

just needs MIS18 complex adding as a synonym

val

Original comment by: ValWood

gocentral commented 9 years ago

Thanks Val.

The MIS18 complex we just curated had subunits MIS18A:MIS18B:MIS18BP1.

I guess if we change the def a bit to make it "MIS16 and/or MIS18 subunits" it would be fine :)

Paola,

What do you think?

Birgit

Original comment by: bmeldal

gocentral commented 9 years ago

Absolutely! Thanks Val... Birgit, I'll do the edits soon, I have a long list of SF tickets from my last week on duty :-(

Original comment by: paolaroncaglia

gocentral commented 9 years ago

No stress, Paola. I will change the AC in the portal entry to the existing term.

The existing complex also needs the relationship part_of GO:0000775 ! chromosome, centromeric region and any of the capable_of_part_of AEs that are applicable.

Birgit

Original comment by: bmeldal

gocentral commented 9 years ago

Hi, Birgit, Paola,

A little more on this (Could you copy in the e-mail discussion so we can see the rationale?)

1. I’m not sure that capable_of_part_of "positive regulation of CENP-A containing nucleosome assembly" rather than just capable_of_part_of "CENP-A containing nucleosome assembly" (this complex is probably directly involved in cenp-A loading rather than rate limiting/regulating?)

2. Also, do we need "centromere assembly complex" .... as a grouping term

I see the problem though...if there was some structure here you probably would have found "CENP-A recruiting complex", and I don't know why I didn't add "Mis18 complex" as a synonym.

-but why not do this by the existing component hierarchy rather than by process?

Centromeric complexes (including Mis18 complex/GO:0098654 CENP-A recruiting complex is a resident part of the centromere-) should be housed under this part of the CC graph: http://www.ebi.ac.uk/QuickGO/GTerm?id=GO:0061638#term=ancchart “chromosome, centromeric region" is the GO term for centromere ( Tangent 1- I would prefer that “centromere complex” which is an exact synonym was the primary name here, can we do this?)

In fact, all of the centromere associated complexes should move under this term (or one of its descendents). This would include, for example: GO:0031511 Mis6-Sim4 complex GO:0031617 NMS complex GO:0000818 MIS12/MIND complex etc Would this be sufficient for the structure you require? This was sort of on my radar but I hadn’t got around to opening a ticket for it……

(Tangent 2, should we rename “CENP-A recruiting complex” to “Mis18 complex” and make CENP-A recruiting complex exact synonym?)

If we did have the proposed term “centromere complex assembly “ grouping term, I am worried that :

  1. It mirrors the process ontology- but the this link between the process and the component is probably something which should be captured at annotation time, particularly in areas where the presses are only beginning to be fleshed out and the annotations are in flux. Most of these “centromere complex assembly” annotations are phenotype based, and they may change when the processes are better fleshed out (some of the CENP-loading for example, could be indirect/upstream processes i.e if one part of the centromere isn't present the next layer cannot form, these are not involved in assembly per se they are just required for assembly to occur or they could be maintaining the CENP-A chromatin once it is assembled). These are difficult to dissect with phenotypes and the annotation may change eventually, but it is difficult to keep track of any independent assertions which is hardcoded in the cellular component hierarchy if the process annotations change.
  2. On this basis, many complexes could be classified as “centromere assembly complexes” in some way (its a big complex and it includes kinetochore assembly), some of which are part of the centromere and some which aren't. It might not be useful to introduce this new (and complicated) axis of classification into this part of the CC graph, as to maintain the hierarchy would be difficult, not least because things are moving so fast in this area.
  3. It means maintaining the hierarchy in 2 places, and if we do it this way, why choose this level (this complex is specifically involved in centric heterochromatin assembly at the central core?

I'm interested in this area, because we (with David H) recently did some ontology revisions for higher level structure for centromeric component terms, but we didn’t yet address the centromere sub complexes, and also because I do quite a lot of annotation of the fission yeast the centromere papers. I would prefer to keep it simpler with component-based grouping terms if possible......

Happy to discuss this further by Skype or at the EBI if required.

Val

Original comment by: ValWood

gocentral commented 9 years ago

Hi Val, Thank you, that is super helpful!

The main reason why I wanted to create the centromere assembly complex is due to the obsoletion comment on GO:0005698 centromere "because it is a genetically defined region and not a specific subcellular location."

And I missed the synonym on chromosome, centromeric region. Plus The above is NOT a child of protein complex.

So yes, please keep existing terms and make following changes:

  1. Obsolete MIS18 complex and add as synonym to CENP-A recruitment complex
  2. Change capable_of_part_of according to Val's suggestion.
  3. Make centromere complex term name and add is_a protein complex (or protein-DNA complex?)

As centromere complex is quite overarching should kinetochore be is_a centromere complex?

Will paste email when back in the office. Writing on my phone right now :(

Birgit

Original comment by: bmeldal

gocentral commented 9 years ago

Emails that triggered the ticket:

Hi Paola,

Another tricky term:

The MIS18 complex is essential for the assembly of the centromere protein CENP-A. I would like to make a generic parent, but I don't know whether it should be 'kinetochore assembly complex' or 'centromere assembly complex' as centromere is not a GO CC (obsolete as not a true subcellular location).

If we added such a parent, it would still be capable_of_part_of GO:0034508 ! centromere complex assembly (The naming is ok here because where GO:0034508 ! centromere complex assembly has been given the logical def 'cellular component assembly' results_in_assembly_of 'chromosome, centromeric region' and 'chromosome, centromeric region' has exact synonym 'centromere complex' ('centromere', instead, is a related synonym)).

The point is, the new parent term would have the same logical def as the MIS18 complex and this wouldn't work. Unless we took the logical def away from MIS18 complex. Are there other children terms that you could envisage for the hypothetical parent term? Is it really needed - would it add information?

Cheers, Paola

Original comment by: bmeldal

gocentral commented 9 years ago

OK... I will read all comments very carefully next week when I'm back from leave, and will edit accordingly :-)

Original comment by: paolaroncaglia

gocentral commented 9 years ago

Hi Birgit and Val,

I did the following:

I’d leave the other pending questions up for discussion with David H as he was already working with Val in this area - I’ll assign this ticket to him but, David, feel free to provide feedback and assign back to me for further edits.

Bye for now :-)

Paola

Original comment by: paolaroncaglia

gocentral commented 9 years ago

Original comment by: paolaroncaglia

gocentral commented 9 years ago

Hi David,

I'm assigning this to you as it relates to previous work you did with Val on centromeres. Please take a look when you have a chance. I did some 'immediate' edits but we'd welcome your feedback on some more general issues. Feel free to assign back to me for edits.

Many thanks, Paola

Original comment by: paolaroncaglia

gocentral commented 9 years ago

Thanks Paola!

Original comment by: bmeldal

gocentral commented 9 years ago

It sounds as though this is all in hand. One other thought I had afterwards. I don't know if the entire "centromere complex" and all of its sub complexes should be defined as a "protein-DNA complexe". It might seem odd for some of the outer kinetochore subcomplexes? I wondered if only the inner kinetochore should be a protein-DNA complex, and the rest should be is_A protein complex? thoughts?

Original comment by: ValWood

ukemi commented 7 years ago

@ValWood @bmeldal

Reading through this ticket, it looks to me like everything here is done. Since you both were the main people in the conversation, can you double-check for me. If there are additional edits required, can you please summarize them and I will make them. Thanks.

-D

bmeldal commented 7 years ago

Hi David,

There was one more suggestion (pt 3) in one of my earlier posts (3 Dec 2014):

Make "centromere complex" term name

There's no discussion as to why this should or shouldn't happen.

Birgit

ukemi commented 7 years ago

Thanks Birgit. I'm trying to clear away some of these old tickets and some require my getting up to speed on the discussions again. I will also wait to hear from Val, but looking at the results of previous edits, 'centromere complex' is an exact synonym of 'chromosome, centromeric region' and the logical definition of 'centromere complex assembly' includes results_in_the_assembly_of 'chromosome, centromeric region'. So it seems to me that swapping the synonym with the term name is at least consistent with what is in the ontology. Any objections Val?

bmeldal commented 7 years ago

Yes, that was considered above, to swap the term name and the synonym.

I hadn't even realised this was still open but I know there are still a few open tickets from when I first started 4 years ago!

ValWood commented 7 years ago

I'm not sure about making 'centromere complex' and exact synonym of 'chromosome, centromeric region'.

From its annotations, it seems much much broader. 'chromosome, centromeric region' is really a rather "fuzzy" legacy CC term.... we used it when things were identified in the vicinity of the centromere locality. It includes a lot of things which are really spindle pole body, like DASH and Aurora OR which exist at the centromere and other places (like topoisomerase), or the SHEC complex which also occur at other heterochromatic regions....

I don't think that centromere complex should be an exact synonym/ its narrow so should probably be a term in its own right.....

bmeldal commented 7 years ago

Ok, two terms cross-wired here :(

I agree with Val that "centromere complex" should not be EXACT synonym of "GO:0000775 chromosome, centromeric region". It's should be at least NARROW if at all a synonym.

I think what I was suggesting 2 years ago was to call the "GO:0098654 CENP-A recruiting complex" "centromere complex", and maybe make "CENP-A recruiting complex" NARROW synonym... That would allow for a slight variation in components while the function is conserved.

ValWood commented 7 years ago

I think that would be confusing. The purpose of the complex will always be to recruit CENP-A (this is a completely generic name for the centromere specific histone), and CENP-A is already in all of the related GO terms (CENP_A containing chromatin assembly etc).

(Also centromerne complex is sometimes used to refer to the CENP-A containing chromatin complex itself PMID: 16716197)

I would say this is one of many centromere complexes....

bmeldal commented 7 years ago

Ok, I have no recollection of the details from 2 years ago! Just read what we said above.

Sounds like in that case, nothing left to do...?

ValWood commented 7 years ago

I haven't seen any problems with how it is now recently ;)