[go] What are evidence codes for?

[Catherine Ball]

Ben,

The answer to the top-down question is that associations between  
genes and GO terms are more or less "trustworthy" if they come from  
different types of evidence.  The most important consumers of the  
evidence codes are the users of GO -- they use those codes understand  
which annotations have experimental evidence behind them and which  
are based on automated annotation pipelines or sequence similarity.

In real life, what mostly happens is that the savvy GO user will  
discard all IEA (inferred from electronic annotation) gene-GO term  
annotations and use the rest.

Cheers,

Cathy

On Sep 14, 2007, at 1:25 PM, Benjamin Hitz wrote:

>
> On Sep 14, 2007, at 11:47 AM, David Hill wrote:
>
>> Exactly. We spend a great deal of time arguing about exactly which  
>> evidence code to use in an annotation. Can we come up with a list  
>> of where we have seen the research community using GO and  
>> determine how they have used the codes?
>
> I think this great from a bottom-up approach, but I was actually  
> interested in the top-down question:  Why were they invented in the  
> first place?  And do they serve this purpose?  Is this purpose the  
> same as it was when they were invented?
>
> Also, Tanya and Jim's points beg the following questions:
>
> Are the evidence codes for the GOC?  For the users of GO?   Is it  
> the best solution to have a single set of evidence codes for both?
>
> Ben
>
> --
> Ben Hitz
> Senior Scientific Programmer ** Saccharomyces Genome Database ** GO  
> Consortium
> Stanford University ** hitz at genome.stanford.edu
>
>