Building and search system for a large-scale DNA database

A flat-file system is inadequate for building, integrating and searching a large-scale database in which the number of entries is increasing in an explosive manner. The DNA database of DDBJ (DNA Data Bank of Japan) was built in a flat-file system until 2 years ago. GenBank in the United States helped DDBJ to convert the flat-file data constructed by DDBJ to a relational format and to install the Annotator's Workbench, AWB. AWB is a tool for building the DNA database with the so-called GenBank-schema based on a relational database management system, SYBASE. However, AWB with the GenBank-schema does not have a simultaneous processing function for a large amounts of DNA data such as Expressed Sequence Tags (ESTs). In addition, both integrating and searching were not carried out in the relational database at DDBJ. Recently, we newly developed a hierarchical relational schema for effectively building, integrating, and searching the DNA database on the relational database management system, SYBASE, at DDBJ. The schema is named the 'DDBJ-schema'. The schema allowed us to implement window interfaces to easily build the DNA database. The DNA database built in the GenBank-schema using AWB is converted into the relational format of the DDBJ-schema using a restructuring tool at DDBJ. We proposed a virtual-structured SQL-programming method to implement the restructuring tool. The method is developed through the 'view' function and the control flow language (CFL) of SYBASE. Moreover, we proposed two methodologies by which we can execute these software tools on UNIX based workstations connected by a computer network.