[Go] [Fwd: [Fwd: Re: [Annotation] annotating ribosomal proteins]]

[Valerie Wood]

This was the e-mail from keseler at ai.sri.com.
There may be other outstanding items on this list as I don't think 
anybody  is responsible for checking thst items are resolved on here.

---------------------

Hi,

I don't think I can contribute anything but more questions to this 
discussion, but would really like to know the answers.  First of all, I 
could absolutely swear that not too long ago, I saw a sentence somewhere 
in the annotation documentation saying something along the lines of, if 
a complex has a function, the function is transferred to its subunits.  
(Maybe it was about process, I don't recall.)  However, I can not find 
this anywhere any more.  It is very important to us to know the answer 
to this; in EcoCyc, I have the ability to annotate complexes themselves 
with GO terms, and I think those become attached to the individual gene 
products for the mapping file.

While poking around in the guide, I only found a couple of relevant 
references to protein complexes.  Under "Valid Function Terms" in 
http://geneontology.org/GO.function.guidelines.shtml, it says this: 
"Functions are not restricted to the activities of single gene products; 
multi-gene product complexes can also have functions."  That's generous 
:-) , but no word on how to transfer those functions to single gene 
products.  There is also a section on "Function Terms for Subunits", 
which refers to the GO annotation guide for advice on how to annotate 
subunits of a complex - in this case, referring to itself is not 
entirely useful.

 From the discussion so far, my take-home message was that one should 
never annotate components of a complex to a process or function of the 
complex with IDA.  Is this the correct impression?  If yes, then looking 
at this the other way, any function/process that is performed by a 
multi-subunit complex will never have an IDA annotation.  No more IDAs 
for translation, replication, transcription,...?

Aside from my IDA concerns, would it be possible to use IMP?  Granted, 
if the mutant phenotype is death, that's not very interesting (Botstein 
said so himself) nor very informative in and of itself, but if a protein 
is part of the ribosome and you can't knock out the gene without killing 
the cell, would it be a safe assumption that said subunit is involved in 
translation?  (I think there *are* non-essential ribosomal subunits, 
which would then not get the translation term.)

-Ingrid



-- 
 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. 
-------------- next part --------------
An embedded message was scrubbed...
From: Valerie Wood <val at sanger.ac.uk>
Subject: [Fwd: Re: [Annotation] annotating ribosomal proteins]
Date: Mon, 02 Mar 2009 13:09:55 +0000
Size: 14369
URL: <http://fafner.stanford.edu/pipermail/go/attachments/20090302/e9eb721e/attachment-0001.mht>