both Jena and RDF4J are great and obvious choices for a Clojure users

RDF4J feels a bit more modern but in the end it doesn't matter if you are using it from clojure

I ended up switching to jena mostly because I wanted to use https://github.com/TopQuadrant/shacl

btw, SHACL seems like a must for a traditional web service if you plan to replace the traditional relational database design with RDF

but I'm just learning about this stuff myself... working through :

http://www.learningsparql.com/

and

https://book.validatingrdf.com/

both great resources

I can't remember whether that book mentions protege:

https://protege.stanford.edu/

My understanding is Protege is used for ontology development, it doubtless has other uses too as there’s a plugin ecosystem too. But yes, you can write an ontology in protege and export it in various formats (including RDF, e.g. as .ttl) which you should then be able to load into a Jena OntModel or whatever… Though I’m really no expert on this; I’ve only ever used it for very basic ontology work — using the reasoner to essentially flush out basic implications and check the domains/ranges all align etc.

I’d personally recommend going easy on the reasoning/OWL stuff… as there are a lot of sharp edges for the uninitiated to cut themselves on. I’d include myself amongst the uninitiated, and I’ve probably got more experience of this stuff than most. I’m not saying this stuff isn’t valuable or useful, it is; it’s just very easy to sink a lot of time into it for very little reward. So just be careful, because it might not be able to infer what you hope… and the inferences may not support useful business cases. Stick to the simple practical benefits, like making certain queries easier to write, and just using simple stuff like domains/ranges, subclasses etc… essentially rdfs+; and you’ll be happier… and even that isn’t normally necessary.