On Construction of Cluster and Grid Computing Platforms for Parallel Bioinformatics Applications

On Construction of Cluster and Grid Computing Platforms for Parallel Bioinformatics Applications

Chao-Tung Yang (Tunghai University, Taiwan) and Wen-Chung Shih (Asia University, Taiwan)
DOI: 10.4018/978-1-4666-0879-5.ch405
OnDemand PDF Download:
$30.00
List Price: $37.50

Abstract

Biology databases are diverse and massive. As a result, researchers must compare each sequence with vast numbers of other sequences. Comparison, whether of structural features or protein sequences, is vital in bioinformatics. These activities require high-speed, high-performance computing power to search through and analyze large amounts of data and industrial-strength databases to perform a range of data-intensive computing functions. Grid computing and Cluster computing meet these requirements. Biological data exist in various web services that help biologists search for and extract useful information. The data formats produced are heterogeneous and powerful tools are needed to handle the complex and difficult task of integrating the data. This paper presents a review of the technologies and an approach to solve this problem using cluster and grid computing technologies. The authors implement an experimental distributed computing application for bioinformatics, consisting of basic high-performance computing environments (Grid and PC Cluster systems), multiple interfaces at user portals that provide useful graphical interfaces to enable biologists to benefit directly from the use of high-performance technology, and a translation tool for converting biology data into XML format.
Chapter Preview
Top

1. Introduction

Bioinformatics is a combination of biology and information technology and includes any computational tools and methods for managing, analyzing and manipulating large sets of biology data. Thus, computing technologies are vital for bioinformatics applications (Konishi et al., 2002; Trelles et al., 1998). For example, biology problems often require repeating the same task millions of times such as when searching for sequence similarities in existing databases or comparing groups of sequences to determine evolutionary relationships. In such cases, the high-performance computers to process this information are indispensable. Biological information is stored on many computers around the world. The easiest way to access this information is to join these computers together through networking. Such activities require high-performance computing infrastructures (Prodan & Fahringer, 2002) with access to huge databases of information.

The major advances in computer technology and computer science over the past 30 years have dramatically changed much of our society. Currently, many parallel versions of bioinformatics applications can be used to conduct computing tasks on Linux PC Cluster or Grid systems, including, HMMer (http://www.rcsb.org/pdb/index.html), and many of these genome websites distribute large datasets as flat files (e.g., tab-delimited files). Flat files are text files lacking any form of markup language. The tasks of automating the processes of information retrieval and integration of heterogeneous biological data are difficult due to their unstructured formats. The data involved may range from nucleic acid and protein sequences, to three-dimensional protein structures, and relationships among various metabolic pathways. Furthermore, different approaches are used for data modeling, storage, analysis, and querying purposes. Therefore, Molecular Biology databases have only a few widely accepted schemas.

Complete Chapter List

Search this Book:
Reset