RE: RDF for molecules, using InChI

Egon,

My question (without knowing the true scale of what you had) was simply to see if a mechanism existed that told me "which molecules at OpenMolecules are URI-resolvable as RDF?".

I guess having a master list that points to each (your second option) would be easiest to use. ANother way to put it, would be to offfer an InChi (or InChI URI) query service that either returned a compound if you had a record of it, or said "Not Found".

The curious thing of Compounds and InChI is that one can actually create new URIs for any imaginable compound (i.e., not even stored) that are resolvable as long as 1) the InChI syntax is correct, and 2) the structure defined by the InChI is plausible. This has some interesting consequences regarding virtualizing entities on the Semantic Web with structure content simply by giving them a InChi URI...

Eric

-----Original Message-----
From: Egon Willighagen [mailto:egon.willighagen@gmail.com]
Sent: Tue 8/21/2007 3:40 AM
To: Eric Neumann
Cc: public-semweb-lifesci@w3.org
Subject: Re: RDF for molecules, using InChI
 
Eric,

On 8/18/07, Egon Willighagen <egon.willighagen@gmail.com> wrote:
> On 8/17/07, Eric Neumann <eneumann@teranode.com> wrote:
> >  Thanks for the pointer-- is there a list of all the molecules you store
> > something about?
>
> Not at this moment. That would be a rather lengthy RDF doc. The number
> of molecules of which something is know is currently in the order of
> 10M. I have not taken up the challenge of hosting that.

Actually, I could make a list of some 250 molecules [1]. Should I make
one RDF file listing all triples for all molecules, or make one master
file, which points to the current RDF 'files'?

Egon

1.http://cb.openmolecules.net/inchis.php

-- 
----
http://chem-bla-ics.blogspot.com/