Abstract: GOBASE is a relational database that integrates data associated with
mitochondria and chloro-plasts. The most important data in GOBASE, i. e.,
molecular sequences and taxonomic information, are obtained from the public
sequence data repository at the National Center for Biotechnology Information
(NCBI), and are validated by our experts. Maintaining a curated genomic
database comes with a towering labor cost, due to the shear volume of available
genomic sequences and the plethora of annotation errors and omissions in
records re-trieved from public repositories. Here we describe our approach to
increase automation of the database population process, thereby reducing manual
intervention. As a first step, we used Unified Modeling Language (UML) to
construct a list of potential errors. Each case was evaluated independently,
and an expert solution was devised, and represented as a diagram. Subsequently,
the UML diagrams were used as templates for writing object-oriented automation
programs in the Java programming language.
