I'm developing a Neo4j database that will contain bioinformatics content. In reviewing similar graph databases for design ideas, I came across one that models one-to-many relationships with an interspersed collection node which then points to the child nodes. For example, Protein -> ReferenceCollection -> Reference(s). This pattern is used for most of the one to many relationships in the database. I was wondering what, if any, advantage this design provided.
