Chem2Bio2RDF is deployed as a Virtuoso triple store available via Sparql endpoint. To build the Virtuoso triple store, each dataset is first deployed as a PostgreSQL dataset (set of tables), and converted to RDF by D2RQ, deployed via Tomcat and a custom mapping file. Virtuoso is loaded from D2RQ RDF in batch; at query time, PostgreSQL and D2RQ are not involved. This workflow is complex but well suited to the complexities of the source data. Source databases are typically available as tables from and compatible with relational databases. Thus each PostgreSQL dataset can reflect the source database accurately. The D2RQ mapping file can easily be modified to convert each database to intercompatible RDF. Virtuoso allows cross database queries (which the multiple D2RQ servers cannot) and provides high performance.