DNA currently has a circular dependency, in that the JCR project depends on the repository and graph projects (and others) as the backing store while the repository project depends on the existence of a JCR layer for the SequencingService and ObservationService. Removing the JCR dependency is tentatively planned for the ObservationService as part of DNA-252. It would be useful to remove the final piece of the circular dependency and make the SequencingService interact directly with the graph layer.

As I see it, this would have the following benefits:
1. Removing a circular dependency in the architecture
2. (Ever so slightly) increasing performance by removing the expense of converting DNA nodes and properties to and from their JCR equivalents