>> What we want within a dataset service is to manage a set of graphs
>> (every dataset is a graph) and to attach metadata (source, identifier,
>> title, ...) to each graph (dataset). So a collection of datasets could
>> look like this in N3:
>>
>> # data
>> dataset1 {
>> 	compound1, feature, value1.
>> 	compound2, feature, value2.
>> 	...
>> }
>>
>>     
Every RDF model is a graph.  One of the things missing above, are
statements, defining the relationship between dataset1 and 
(compound,feature,value) triples.   It's a little bit counter intuitive
to think in terms of statements (triples), and not objects that can
include other objects, but this is exactly what gives RDF its flexibility.

A guide (Defining N-ary Relations on the Semantic Web) how the above is
modeled by triples is available at
http://www.w3.org/TR/2006/NOTE-swbp-n-aryRelations-20060412/ . Following
the guide and modeling Features as resources will lead to a model very
close to the proposed. 

BTW, below is an excerpt of  how ChemAxiom describe properties of the
compounds (note the intermediate classes):

:benzene
a ChemDomain:NamedChemicalSpecies ;

ChemAxiomProp:hasProperty
:Density_1 , :BoilingPoint_1 ;

:BoilingPoint_1
a ChemAxiomProp:BoilingPoint ;

ChemAxiomProp:hasValue
"80.1"^^xsd:string ;
:measuredBy Metrology:Ebulliometry .

:Density_1
a ChemAxiomProp:Density ;
ChemAxiomProp:hasValue
"0.8786"^^xsd:string .

:benzeneMolecule
a ChemDomain:MolecularEntity ;

ChemDomain:hasIdentifier
:MolecularFormula_1 , :CASNumber_1 .

:CASNumber_1
a ChemDomain:CASNumber ;
ChemDomain:hasValue "71-43-2"^^xsd:string .

:MolecularFormula_1
a ChemDomain:MolecularFormula ;

ChemDomain:hasValue "C6H6"^^xsd:string .

SCOVO is another example, they have Datasets and Items inside datasets 
http://sw.joanneum.at/scovo/schema.html.

Best regards,
Nina