consider what datasets at NCBI (not in datastore) might be representable in DSCensor

[adf-ncgr]

seems like having stats for genomes and annotations at NCBI could be useful; they also have datatypes that we might not have that would expand the interest of the interface (e.g. methylation data)