David Nickerson wrote:
> Hi all,
> Just wanted to see what people think about using the BioModels 
> qualifiers (http://biomodels.net/index.php?s=Qualifiers) in CellML 
> models? i.e., following the SBML annotation specification (see section 6 
> of the SBML level 2 version 2 specification).
> The reason I ask is that I am starting to look at how to reference 
> external data, for example to justify a parameter value or provide 
> experimental data for use in making a graph. At least for these examples 
> the http://biomodels.net/model-qualifiers/isDescribedBy qualifier seems 
> quite appropriate - where the referenced resource could be a journal 
> publication to justify a parameter value or a reference to some 
> experimental data.
> Before progressing too far with this, I thought I'd better check. One 
> problem I can see immediately is that the 
> http://biomodels.net/model-qualifiers/is and/or 
> http://biomodels.net/biology-qualifiers/is qualifiers are pretty much 
> the same as the cmeta:bio_entity already defined in the CellML Metadata 
> Specification. So is there already something in the CellML metadata 
> specification that will let me reference arbitrary (possibly external) 
> resources? Is it ok to just use the BioModels qualifiers that we want 
> without supporting them all? Or is there some other way to achieve the 
> same result?
The cmeta specification does allow you to give an identifier as a URI:

<rdf:Description rdf:about="#myVariable">
      <rdf:li rdf:parseType="Resource">
        <cmeta:identifier rdf:parseType="Resource">
> I know Carey did some initial work looking at using the XML Resource 
> Directory Description Language (RDDL, www.rddl.org) to reference 
> external data. To make full use of this (especially in regard to model 
> curation), I think we would need to define our own natures (roles) and 
> purposes (arcroles) - which would essentially result in the same 
> metadata as using the BioModels qualifiers but with different namespaces.
If we are using the exact same semantics as they are, we could the 
BioModels namespaces. Of course, it is not clear that the semantics used 
are sufficient for useful machine interpretation, and so it would be 
worth reviewing the existing set, coming up with a range of examples, 
and possibly using these to create a richer set of semantics with their 
own URIs.

Best regards,

cellml-discussion mailing list

Reply via email to