I concur - I think that the added power of OWL is vital for any SW toolkit. Sesame has a number of the features I would find valuable, but OWL is a big thing.
I've started working on a simple coupling where a SAIL implementation also implements the org.semanticweb.owl.model.OWLOntology interface.
Now checked in and available is a rule file for use with the custom inferencer that supports OWL entailment, specifically the OWL DLP subset as identified in [Grosof and Horrocks, WWW2003].
Supported OWL constructs are: equivalentClass, sameAs, TransitiveProperty, SymmetricProperty, inverseOf, hasValue restrictions, intersectionOf.
Note that this is a very naive solution that uses forward chaining. The approach will not be scalable enough for expressive OWL ontologies with large A-boxes, but may suffice for 'light' use of OWL primitives in mostly RDFS-only ontologies (for example, occassional use of sameAs to enable concept identity matching).
Is the paper you refer to thins one: http://www2003.org/cdrom/papers/refereed/p117/p117-grosof.html
They don't talk about OWL but about DAML+OIL. In any case, the question I have and I'm not sure is answered is if the inferences supports owl:TransitiveProperty.
I think that is an old version of the paper I meant, I'm sure there is an OWL-specific update somewhere. I'll look around.
To answer your question: yes, TransitiveProperty is covered by Sesame's OWL DLP custom inferencer.