Bio-data Project

The data needed for the processing on the BioGrid reside in what people call ‘databases'. These are just flat files of various formats that are regularly updated by the maintainers of these ‘databases'.
In order to avoid confusion for those who think of databases in the RDBMS sense of the word, in this document the term ‘bio-databases' will be used. In order to be able to submit jobs on the distributed infrastructure of the BioGrid, the files of the various bio-databases need to be easily accessible from every node in the BioGrid. Of course there are many factors that need to be taken into account when addressing the distributed bio-database problem in the BioGrid.

The points above build the basis for a possible design of a distributed bio-database infrastructure in Switzerland. There are other relevant points to consider that are not directly but indirectly influencing the design:

• Swiss users preferred. Somehow it should be possible to assure that the Swiss academic users get preferred downloads over non-Swiss or non-academic users.
• Distributed Expertise. The people having the in-depth knowledge about the various bio-databases, their content, etc. are spread across all of Switzerland and beyond.
• Uniform Environment. The executables making use of a database need to be able to run at any site. Therefore it is necessary that localities are collected into a uniform BioGrid environment that can be configured individually at each site while providing uniformity and ‘ubiquitousness' to the application. o Agreement on a set of environment variables to be used in the Grid job is necessary o The environment needs to be extensible
• Distribution of applications. Not only the bio-databases but also the corresponding applications could be mirrored and made readily available to the community, if they come without licensing constraints and need to be updated regularly.

These requirements should all be taken into consideration. Some mechanisms already exist to mirror bio-databases at individual institutions.