Hi everyone,
This is following up from ticket #10717 request
GO:1990328 RNA polymerase II, RPB4-RPB7 subcomplex:
- part_of GO:0005665 DNA-directed RNA polymerase II, core complex AND
- part_of GO:0000932 cytoplasmic mRNA processing body
Tanya pointed out that multiple part_of relationships are NOT allowed if they don't happen simultaneously. In the above example the complex has two functions in two different compartments independent from each other (as far as I can tell - see UniProt annotation!). At present, we cannot capture this but I would like to do it in some form (apart from the def).
Let's hear your thoughts :)
Birgit
Actually, in the above example it's two PROCESSES the complex is involved in so we cannot use the capable_of annotation-extension either.
One way to handle this is to make two subclasses of the complex, one for each location, and assign the capable_of in the relevant place.
Logically sound, but a little unsatisfying as it leads to ontology inflation, also curators may make annotations at different levels.
Birgit, I think the best way to do this is to pre-compose in the ontology as Chris suggests:
RNA polymerase II, RPB4-RPB7 subcomplex
---[i] RNA polymerase II core complex, RPB4-RPB7 subcomplex -capable_of/capable_of_part_of-> ??
---[i] RNA polymerase II RPB4-RPB7 subcomplex, mRNA processing body -capable_of_part_of-> mRNA processing?
(sorry not sure what the actual functions/processes are but you see what I mean!)
Then you can annotate to the specific term.
Jane,
Why can we not have capable_of and capable_of_part_of extensions to different function/processes? Where does it break the logic (my naivety comes to light here!)?
Is capable_of_part_of the relationship to process?
Birgit
You can have capable_of and capable_of_part_of relations to different function/processes - that's fine.
What you can't do is have a complex be part of two different physical entities - e.g. cytoplasm and nucleus - because no instance of that complex can be part of those two entities at any one time.
So the restriction is to cellular component relations.
A complex can be capable of multiple functions/processes.
Last edit: Jane Lomax 2014-03-31
Ok, so I have two choices in this specific case:
either have one GO term for the complex that has a part_of relationship at a high level, e.g. cell, and several capable_of_part_of extensions, or
two complexes in two compartments with only the relevant capable_of_part_of extension for the process in this compartment - as suggested by Chris and Jane.
I let you decide which is better and then change the current term and IntAct annotations accordingly.
Cheers,
Birgit
I think we have to go for the latter, otherwise we lose the fact that 'RNA polymerase II, RPB4-RPB7 subcomplex' is part of RNA polymerase II which would be a fairly major omission.
Ok, in that case I think we shall have:
(i)
GO:1990328 RNA polymerase II, RPB4-RPB7 subcomplex, involved in DNA-templated transcription:
- part_of GO:0005665 DNA-directed RNA polymerase II, core complex
- capable_of_part_of GO:0006352 DNA-templated transcription, initiation
- capable_of GO:0003899 DNA-directed RNA polymerase activity
Def:
A protein complexes that mediates DNA-templated, RNA polymerase II dependent transcription.
(ii)
[NEW] RNA polymerase II, RPB4-RPB7 subcomplex, involved in translation initiation and mRNA decay pathway:
- capable_of_part_of GO:0000956 nuclear-transcribed mRNA catabolic process
- capable_of_part_of GO:0006413 translational initiation
Def:
A protein complexes that mediates the two major cytoplasmic mRNA decay pathways and is required for efficient translation initiation in association with RNA polymerase II (Pol II).
Please modify as needed!
Thanks, Birgit
Okay I think I misunderstood here...I thought RPB4-RPB7 subcomplex was part of different supercomplexes when it performed its functions. Is that not the case?
Hi,
Could we please try to avoid creating two different instances of the same subcomplex, one part of x and the other part of y. I think this just becomes overly complicated and confusing for curators.
In some previous cases where a subcomplex is part of two different complexes we have discussed the idea that we may be able to use the has_part relationship to indicate that two different complexes contain the same subcomplex. For the "core TFIIH complex", Eurie Hong even submitted a SF item to use has_part relationships so we can get rid of the two horrible terms, though I don't think it was ever acted upon:
In this case, for the 4/7 subcomplex, while it is of course not true that a single 4/7 subcomplex is part_of the RNAP II complex and part_of the "cytoplasmic mRNA processing body" at the same time, it is true that a single 4/7 subcomplex participates in both of these things. The 4/7 subcomplex cycles on and off of RNAP II and back and forth between the nucleus and the cytoplasm.
For the "DNA-directed RNA polymerase II, core complex" term, it would definitely be true to say that it has_part the "RNA polymerase II, RPB4-RPB7 subcomplex".
-Karen
This type of has_part causes big problems for maintaining consistent curation.
Here, if you annotate Rpb4 to "RNA polymerase II, RPB4-RPB7 subcomplex", and the relation is via "has_part" the geen product will not be automatically annotated to
"DNA-directed RNA polymerase II, core complex"
...this means we will accumulate annotation 'holes' everywhere, because nobody is going to remember to do this (or indeed, know when they need to).
Has_part is great for ontology maintenence, but it isn't so good for annotation consistency.........
I proposed the integral_to qualifier which solves this. However, after presenting at a few GO meetings, folks still had issues with it.
Also we still would like a way (I think) to indicate that the different forms of the complex have different functions.
It may be easier to go with the pre-composed class in this case. To avoid curator inconsistency we could place a do-not-annotate-directly constraint on the generic form, forcing a distinction to be made (assuming it's always possible to tell)
Thank you for all your comments! FYI, I will be on hols 4-14th April.
I'm looking forward to a resolution of this problem as I will have many more of these cases once I start on super-complexes (and the IntAct schema is nearly ready for this!).
Birgit
I still need an answer on this question, Birgit:
Is RPB4-RPB7 subcomplex part of different supercomplexes when it performs its different functions? Or always part of RNA Pol II?
Hi Jane,
The 4/7 subcomplex is definitely not part of RNAP II when it goes to the cytoplasm and becomes localized to cytoplasmic mRNA processing bodies. It is well known that the 4/7 subcomplex separates from RNAP II and goes to the cytoplasm where it has roles in regulation of mRNA degradation. Basically 4/7 appears to cycle between the nucleus and cytoplasm and it is only associated with RNAP II in the nucleus.
However, I've never seen the "cytoplasmic mRNA processing body" described as a complex, or supercomplex, but rather a location within the cytoplasmic region that can be identified by the presence of certain proteins. From what I know of it at this point in time, it seems unlikely that it is a complex with a defined stoichiometry, but rather a structural organization containing mRNA and proteins involved in mRNA metabolism in the cytoplasm, kind of like the nucleolus is a substructure within the nucleus.
I hope that helps
-Karen
Thank you, Karen! You beat me to answer Jane's question - and so much better than I would have done it! However, that still doesn't help our dilemma as it's definitively involved in two completely separate cellular compartments!
Birgit
Okay, how about this:
RPB4-RPB7 complex GO:new
---[i] RNA polymerase II, RPB4-RPB7 subcomplex GO:1990328
---[i] cytoplasmic mRNA processing body, RPB4-RPB7 complex GO:new
cytoplasmic mRNA processing body GO:0000932
---[p] cytoplasmic mRNA processing body, RPB4-RPB7 complex GO:new
DNA-directed RNA polymerase II, core complex GO:0005665
---[p] RNA polymerase II, RPB4-RPB7 subcomplex GO:1990328
RNA polymerase II, RPB4-RPB7 subcomplex GO:1990328
-capable_of_part_of-> GO:0006352 DNA-templated transcription, initiation
-capable_of-> GO:0003899 DNA-directed RNA polymerase activity
cytoplasmic mRNA processing body, RPB4-RPB7 complex GO:new
-capable_of_part_of-> GO:0000956 nuclear-transcribed mRNA catabolic process
-capable_of_part_of-> GO:0006413 translational initiation
Hi Jane,
I think (!) the logic works even though it is the 'inflating' option some did not favour. Can I then annotate my Intact complex to BOTH children of the 4/7 complex so that we don't lose their respective relationships with the RNAPII and the mRNA processing body?
Can you please check with Tony and Rachael that that won't break the export they are currently working on?
Thanks,
Birgit
Hi Birgit
I know it's not ideal you have to do this sort of double annotation, but even if we went with the has_part solution Karen favours, you still wouldn't get the connection to both the RNAPII and the mRNA processing body because annotations don't propagate over has_part. So this is probably the best solution for now.
Rachael says annotating to both won't cause any problems for them.
Jane
Thanks, Jane.
The best of a bad solution :) Good news it won't break the export. I will fix the annotation in Intact. You can close this ticket once you have created the new terms.
Birgit
The new terms discussed above have not been created after all and the complex term is still a direct child of 'protein complex' with no hard-wired part_of relationships at all.