High throughput Web services for life sciences

This paper describes a collaborative project between IBM and the North Carolina BioGrid: building, deploying, and using high-throughput Web services for bioinformatics applications. Bioinformatics applications can highly benefit from grid computing power, but the challenges of exposing these applications to the grid often cause reluctance. Typical bioinformatics applications are "quick-and-dirty" Perl scripts, which usually run at small homegrown laboratory clusters. Our work intends to fill the gap between grid power and the bioinformatics world by utilizing Web services and by extending the Open-Bioinformatics Foundation's software package BioPerl. Our extensions to BioPerl libraries allow bioinformatics applications to be exposed to Web services seamlessly. Our development tool WSDL2Perl integrates these applications with the vast computational power of the NC BioGrid. We present a case study in building Web services for bioinformatics applications, namely BLAST (a well-known standard homology search application), and a process for deploying grid-enabled Web services to the NC BioGrid.