[Psidev-ms-dev] RE: [Fuge-devel] Article about data model development

[Miller, M. D (Rosetta) <Mic...@Ro...>]

Hi Andy,
 
Great review, gives a good perspective on the standards.
 
The only thing I wanted to add is about the intent of the NVT.  
 
As I mentioned previously at the meeting for developing the FuGE milestone,
one of our key use cases is the ability to use MAGE in pipeline situations.
Typically the use of the format in these circumstances is in no-way complete
in terms of repository export for any particular pipeline but we have been
very successful in getting a diverse amount of data from various
instruments, wet labs and third party annotation that end up being connected
together in our application.
 
The NVT is documented as "Allows specification of name/value pairs.  Meant
to primarily help in-house, pipeline processing of instances by providing a
place for values that aren't part of the specification proper."  So best
practice says don't use NVTs for anything that has biological significance.
But we have indeed used these for many processing purposes, one of the
primary ones being for specifying what action to take if a record already
exists.
 
cheers,
Michael

-----Original Message-----
From: fug...@li...
[mailto:fug...@li...] On Behalf Of Andy Jones
Sent: Friday, October 14, 2005 6:16 AM
To: fug...@li...; psi...@li...;
psi...@li...
Subject: [Fuge-devel] Article about data model development



Hello all,

 

Norman Paton and I have written an article that reviews the current models
in functional genomics and the kinds of tasks that must be performed over
data formats. It then presents a set of guidelines about the kinds of
modelling structures that are suitable for particular tasks. I think the
guidelines may be of interest to people currently involved in data model
development. The article is in BMC Bioinformatics
(http://www.biomedcentral.com/1471-2105/6/235
<http://www.biomedcentral.com/1471-2105/6/235> ).

 

Apologies for the blatant self-publicity!

Best wishes,

Andy