ion channel gene annotation

[David Hill]

Hi Doug,

The east coast folks are awake now. I have had the same issue with the 
contributes_to qualifier and I think only the details of certain 
experiments will allow us to use it. Let's take an ion channel and a 
classic experiment of injecting a Xenopus oocyte with the RNA for it to 
test its function. So we inject the oocyte and we patch clamp and show 
that the injection of the RNA confers an ion channel to the oocyte. I 
think that THIS experiment leads the authors to conclude that the gene 
has ion channel activity. But the channel may need to work as a 
homo-dimer or a heterodimer. This experiment doesn't address this. In 
another experiment, they may show that the channel works as an obligate 
homodimer by mutating a dimerization domain and eliminating function. In 
this case I think that then we could add the contributes_to qualifier.

In the same vein, I think many mutational studies that eliminate the 
function of one element of a functional heterodimer simply show loss of 
function. In this case, we have no evidence that the gene product did 
not have the function. I have made many annotations for the reference 
gene project where in some cases the annotation contains the 
contributes_to qualifier and in some cases an annotation to the same GO 
term does not. It all depends on what the experiment tells us.

David

Doug howe wrote:
> If a paper shows that introducing a gene product into a cell line 
> confers ligand-stimulated ionic currents upon the cell, and the 
> authors say (but don't show) that the gene product forms a 'homomeric 
> ion channel ", should this gene be annotated directly to 'ligand-gated 
> ion channel activity', or should this be a 'contributes_to' annotation?
> -Doug