tRNA 3'-trailer cleavage, endonucleolytic

[gocentral]

I have 2 tRNAse Z's wtih 3'-tRNA processing endoribonuclease activity one is mitochondrial an one is nuclear.

What is the best way to curate these? Should we add new process terms, mitochondrial tRNA 3'-trailer cleavage, endonucleolytic or should they be with an annotation extension

Reported by: ValWood

Original Ticket: geneontology/ontology-requests/8044

[gocentral]

Good question" There's certainly precedent for mitochondrial process terms, but from time to time someone moans that we should put the CC info in annotation extensions instead. (And Chris always just points out that they're logically equivalent so we should decide what makes more sense to annotators ... which is exactly what we usually have trouble figuring out.)

Anyway, seems to me you could either poll the annotators and see if a clear winner emerges, or just decide what you want to do and hope it doesn't cause too much grief ;)

m

Original comment by: mah11

[gocentral]

Yes, I do this a bit inconsistently at present. Sometimes I request the terms, and sometimes I don't. In general I prefer the terms because they are immediately visible. I don't know when we will be able to make the annotations extensions I am stockpiling visiblem so I tend to reserve these for things which which could not be captured a GO term (i.e xps to other ontologies). Certainly in PomBase I expect it to be >12 months before we can handle XPs. I just wondered if we should have some guidelines, or if it was annotator preference. Perhaps this is one to discuss during an annotation call....

Original comment by: ValWood

[gocentral]

yeah, for me it's easy -- I just add the terms or don't add the terms, depending on how annotators want to annotate

Original comment by: mah11

[gocentral]

Hi,

You should talk to Chris and David about creating location specific terms. THere was a discussion on this topic y'day at the annotation call- transmembrane receptor activity. Should transmembrane be part of the term or should it be a cross product. As Midori says we have precedence for creating location specific BP terms (mitochondrial transcription). In the past we allowed this if the process was very specific for that location. Having said all that, I think it is a good idea to bring this up on the next Annot. conf. call. Becky is going to check with Chris on the transmembrane issue. We can wait to hear from her and take it from there?

Rama (lot of rambling, sorry!)

Original comment by: rbalakri

[gocentral]

Hello,

Chris said: // I don't think that if the CC is intrinsic to the function it's fine to have the compositional term. E.g. transmembrane receptor kinases should stick around surely? These should of course be accompanied by the correct logical definition.

I recall a lot of the existing "transmembrane X" terms were fairly inconsistent and I've nothing against going through and cleaning these up //

Original comment by: rebeccafoulger

[gocentral]

Good point about yesterday's call, but I think it's worth bearing in mind that we don't necessarily have to do the same thing in MF and BP; for a long time we've accepted location-based distinctions in BP even as we tried to avoid them in (most of) MF.

Original comment by: mah11

[gocentral]

I prefer the specific term otherwise I would need to do mitochondrial RNA processing and tRNA metabolic process to represent the same annotation

I always prefer adding the specific term ...... less chance of missing annotations, so better annotation consistency, and enrichment results.

val

Original comment by: ValWood

[gocentral]

I prefer adding a more specific process term because it prevents the need for redundant annotation in different branches (maybe this would be a rule for adding a term if it would have multiple parents?) . It is also much easier if annotators can see the terms pre-composed as they will be encouraged to annotate more correctly, and this improves annotation consistency and the results of analysis.

I am doing some enrichment analysis at the moment and I can see genes which are clearly enriched in certain mitochondrial (and other) processes which are not statistically overrepresented because my annotation is not specific enough. I am going through these now and trying to find common specific terms to link the gene sets.

I would go for more specific term every time if it is a different process.

The transmembrane receptor activity is different. This is a function and a component.

Original comment by: ValWood

[gocentral]

Just clarifying that the 'mitochondrial' and 'nuclear' terms here refer to the location where the activity is taking place, not the genome that encodes these gene products. Is that right? That means it is tRNA 3'-trailer cleavage that occurs in the mitochondrion NOT 3'-trailer cleavage that happens to mitochondrial genome encoded tRNAs.

Original comment by: tberardini

[gocentral]

It is referring to the cleavage that occurs in the nucleus and the mitochondria respectively. However the activity which occurs in the mitochondria is specific for mitochondrial tRNAs, and the nuclear process presumably occurs in the nucleus before the nuclear tRNAs are exported to the cytoplasm.

However, I think in some species the same enzyme might perform both the activity which acts on cytoplasmic tRNAs (in the nucleus) and the activity which acts on mitochondrial tRNAs), so maybe in this cases it does need to be a single process I GO? If so in these cases it bothers me that the mitochondrial processing events cannot be distinguished from the nuclear ones, because the phenotypic consequences are different. For many experiments (microarray, proteomics, phenotype analysis) if you are using GO to analyse the data it isn't helpful not to have these, and many similar cases "lumped".

Original comment by: ValWood

[gocentral]

I would say add precomposed terms. We have plenty of these in BP don't we? Of course there is the usual worry about redundancy between MF and BP but that holds regardless of whether we pre-compose in MF (although I suppose this could make the problem worse, if we have large numbers of MFxCC terms duplicating BPxCC terms).

We can set up a TG template if these are common

Original comment by: cmungall

[gocentral]

Looking at the comments, I went ahead and added the term for this one:

mitochondrial tRNA 3'-trailer cleavage, endonucleolytic GO:0072684

Original comment by: mah11

[gocentral]

• labels: --> New term request
• milestone: --> 224410
• assigned_to: nobody --> gomidori
• status: open --> closed-accepted

Original comment by: mah11

[gocentral]

• milestone: 224410 --> PomBase

Original comment by: mah11