Fig. 1

Overview of the interoperable chemical structure search implementation, separated into logical layers. Clients submit the SPARQL queries to the frontend HTTP server in Apache Tomcat. The incoming queries are passed to the SPARQL engine, which uses the connected RDF/SQL mapping to translate them into equivalent SQL queries. The resulting queries are evaluated by backend PostgreSQL DBMS, using the Sachem extension (which executes chemical structure search) and pgSPARQL extension (which evaluates SPARQL-specific parts of the query and gathers metadata). Stored data include the relational data in PostgreSQL and specialized chemical indexes in Sachem. Resultsets of the evaluated queries are passed back to the SPARQL engine, translated into a SPARQL result in the requested format, and sent back to the client