Yoshinori K. Okuji, Toshiaki Katayama, Mitsuteru S. Nakao

BioRuby

Thanks to the significant improvement of biotechnology and the advance
of computational biology, we now have a lot of biological resources,
such as genome sequences, expression profiles and signaling pathways,
mostly integrated into publicly available databases. Although that is
unquestionably very useful to study the large and complex systems of
organisms, that also brings up an unexpected problem: You cannot do
what you want to do without effort. This is sometimes because most
databases use their own data formats which are incompatible with each
other, sometimes because link information is missing or difficult to
extract between different kinds of databases, and sometimes because
bioinformatics tools are not necessarily adequate for batch processing
which is quite often required for bioinformaticians.

Thus, we believe that a kind of reusable framework for bioinformatics
is necessary. Such a framework may provide us with easily accessible
interfaces which should hide the difference of database formats and
user-unfriendly interfaces of computational tools. For this purpose,
the object-oriented model is appropriate, because it can build up an
unified interface more consistently than the old-style procedure
model, regardless of the difference of underlying data objects.

In this presentation, we introduce BioRuby, which aims to provide
"open source" resources for bioinformatics. The main difference
between BioRuby and the other projects, such as BioPerl, is that our
project is based on the pure object-oriented language Ruby. As BioRuby
is at the dawn, we have a few database classes and computational tools
at the moment, but we are already using the code for daily
researches. As an example, the implementation of database classes
(e.g. SwissProt class) will be shown.