Migration and/or separation of databases containing chemical structures and/or reactions
Migrations or separations of structure and reaction databases are pretty straight forward as long as the drawing rules and data models of the starting database(s) and destination(s) are identical or at least very similar. In this special case all structures and their related data that must be migrated can be stored in a flat SD file (or simple data record) that is used for the re-registration into the new database for example by tools like Accelrys’ Pipeline Pilot, Knime, or by Oracle processes. But this simple case is the exception in the field of mergers and separations.
One of the most popular transfer file types for reactions, the Reaction Data file (RD file) is hierarchically organized representing the fact that each reaction may have one or more variations. That makes RD files laborious to be handled by data pipelining tools because most of them do not handle hierarchical data formats as default. But even if the data model adaption is solved the drawing rules of the starting and destination databases may differ providing different representation for identical chemical structures or reactions on both sides of the transfer process-And last but not least for mergers between different DBMS each system provides properties that may not be transferrable and need special handling in order not to lose any information. (See Accord example).
StructurePendium supports the full migration and separation process including
- development of migration rules for the participating data models
- development of migration rules for chemical structures and/or reactions (for comparison see article about drawing rules)
- set up of fully automated migration processes for chemical structures/reactions and all related data (for comparison see article about automated Transformation)
- migration of biological and physicochemical data between different databases / DBMS.
- development of the requirements and project management
The member of StructurePendium gathered experiences with multiple tools starting with the export/import tools of MDL’s ISIS and Isentris, working with Oracle PL/SQL procedures, VB.NET, or using Accelrys’ Pipeline Pilot or Knime beside other tools that are provided by the DBMS vendors.