Parallelism in bioinformatics workflows

Abstract

Parallel processing is frequently used in bioinformatics programs and in Database Management Systems to improve their performance. Parallelism can be also used to improve performance of a combination of programs in bioinformatics workflows. This work presents a characterization of parallel processing in scientific workflows and shows real experimental results with different configurations for data and programs distribution within bioinformatics workflow execution. The implementation was done with real structural genomic and automatic comparative annotation workflows and the experiments run on a cluster of PCs. © Springer-Verlag Berlin Heidelberg 2005.