We describe challenges in and approaches for modelling multiword expressions in machine-readable dictionaries. OntoLex is a widely used community standard for lexical resources on the web, and the predominant RDF vocabulary for the purpose. The current challenge is for OntoLex users to figure out the correct modelling strategy, as different use cases require the application of different OntoLex modules. This chapter serves as an orientation point for researchers and practitioners, and for a number of real-world use cases it will describe modelling strategies and compare their advantages and disadvantages.