Choosing the right programming model can play a decisive role in framing the kinds of problems that can be efficiently solved - and as a result the scope of a system. I like how Marko Rodriguez' presentation (below) helps clarify those situations in which a graph traversal model excels over a relational table model.
Chemistry is full of graph relationships at all scales. I'm fascinated by the idea of re-interpreting some common cheminformatics programming and database problems in the context of graph traversal. Fortunately, powerful and free tools now exist for doing so.