Distributed BLAST in a grid computing context

The Basic Local Alignment Search Tool (BLAST) is one of the best known sequence comparison programs available in bioinformatics. It is used to compare query sequences to a set of target sequences, with the intention of finding similar sequences in the target set. Here, we present a distributed BLAST service which operates over a set of heterogeneous Grid resources and is made available through a Globus toolkit v.3 Grid service. This work has been carried out in the context of the BRIDGES project, a UK e-Science project aimed at providing a Grid based environment for biomedical research. Input consisting of multiple query sequences is partitioned into sub-jobs on the basis of the number of idle compute nodes available and then processed on these in batches. To achieve this, we have implemented our own Java-based scheduler which distributes sub-jobs across an array of resources utilizing a variety of local job scheduling systems.