A large number of bioinformatics analysis tools available today are processor intensive. Keeping in mind that the amount of biological data to be analyzed is growing steadily, and these tools are not easily deployable on Beowulf clusters, we propose BioFilter. BioFilter is a software implementation of an architecture that provides a framework for the deployment of these tools on a cluster thereby providing large performance increases. Another facet of bioinformatics analysis is that the tools, in general, are not used in isolation, but are used collaboratively during an analysis. Our architecture provides a way to create pipelines of different analysis tools chained together in a cluster environment, passing the output from one analysis tool as the input to another analysis tool.
Thomas S. Brettin, Avinash Kewalramani