Hi Egon,

Thanks for replying!

Egon Willighagen <egon.willighagen at gmail.com> wrote:

> Hi Nina,
> 
> On Fri, Oct 2, 2009 at 8:24 AM, Nina Jeliazkova <nina at acad.bg> wrote:
> > I am tempted to propose RDF as a standard exchange format for OpenTox
> > resources. Instead of custom XML schemas and custom name spaces, we can
make
> > use of existing RDF name spaces. This should also work nice with
ontologies
> > and integrating data from different sources.
> 
> I can strongly second this suggestion, and help/advise here in these
> matters if you like. Last week I have become official Invited Expert
> to the W3 Health Care and Life Sciences (HCLS) Interest Group, and
> working on RDF in relation to cheminformatics (in the CDK and
> Bioclipse, mostly).
> 
Congratulations! It will be really good if you could help us with some
pointers to start with? OpenTox would be interested in RDF for

1) Chemical compounds 
2) Properties of chemical compounds (any kind of; existing ontologies )
3) Sets of chemical compounds 
4) Sets of chemical compounds + properties
5) Models
6) Algorithms (machine learning algorithms as a start)
7) Model validation

I am pretty sure there is lot of work already done 1) and 2) (e.g. RDF name
space for InChI exists).


> > From technical point of view, RDF/XML can be considered just a XML with
> > specific name spaces.  It seems there are plenty of libraries for
different
> > languages and frameworks.
> 
> RDF/XML is only one serialization; others exist too, like N3, but even JSON.
> 

Yes, I pointed to XML serialization, because it might be perceived as a
shorter step from current custom XML schemas. I didn't knew about JSON
serialization (many things to learn these days).

> > Last but not least RDF offers a way to denote multiple URI point to the
same
> > object ;)
> 
> I assume your refer to owl:sameAs ?
> 

Yes, we had the issue of non-unique URIs discussed several times.

> Egon

Best regards,
Nina