Distributed Protein Sequence Alignment

Given the explosive growth of biological sequence databases and the computational complexity of aligning large sequences over extremely large databases most researchers have opted for utilizing the BLAST algorithm. While BLAST is completely appropriate for some purposes, the more rigorous and more computationally expensive Smith-Waterman algorithm is preferred for certain purposes. This work presents an implementation of the mathematically optimal SmithWaterman protein sequence alignment algorithm using a collection of distributed computers. We also present the use of the Unicon programming language as an alternative for writing biological search algorithms and other applications in bioinformatics rather than the most commonly found C or Perl approaches. The system has fault tolerant capabilities and can dynamically add and remove nodes to deal with worker node failure. The system currently operates on up to 87 P4 3.0 GHz machines. The system is called UDPS for Unicon Distributed Protein Searcher