A distributed algorithm to align distantly related protein sequences using profile analysis

The explosive growth of biological sequence information has underscored the vital importance of integrating the computer and life sciences. Multiple sequence alignments, both DNA and protein, can be used to study different groups of related genes, to infer evolutionary relationships between them, and to discover patterns that are shared among functionally or structurally related multiple sequences. By combining the power of desktops, high performance clusters and supercomputers, grid communities can solve such heavy-computation problems affordably and efficiently. This paper focuses on parallelizing the process of finding and aligning distantly related sequences using information tables called "profiles". A dynamically scalable, heterogeneous grid of workstations implementing an intelligent job allocation mechanism is used to distribute the process. This approach focuses on both accuracy and complexity.