The Institute for Systems Biology (ISB) has selected IBM to provide its infrastructure technology. ISB will use IBM servers, storage and data integration products to support its research on protein-protein interactions to better understand and predict diseases, and identify potential preventions and treatments.

Through the agreement, announced today at the BIO 2002 Conference, IBM and ISB will also explore research collaborations in systems biology. ISB is a leader in this emerging field, which uses computer technology to model not just the functions of individual genes and proteins, but their complex interactions within a cell, tissue, organ or whole organism.

"This research is the essence of what ISB is all about," said Co-founder and President Dr. Leroy Hood. "You can't learn about systems by studying one gene or protein at a time. The information from genes and proteins is complex and requires tremendous computational firepower. The technology framework from IBM will give us the power we need to quickly analyze and integrate the data, and accelerate our research efforts."

ISB is replacing products from non-IBM vendors with the IBM systems, including a 64-node IBM eServer xSeries 1300 Cluster, with two micropressors per node. This highly scalable, prepackaged cluster will be used to process data from a network of mass spectrometers, which identify and analyze proteins -- critical steps in the process of determining protein-protein interactions.

"As we looked over the landscape of various computer companies to determine which ones we should be working with, IBM was clearly our first choice," said Professor George Lake, ISB faculty member. "It was critical to have a technology partner who understood our research goals. IBM does everything -- they are a leader in high-performance computing, Linux clusters, storage and hierarchical data management, and all of this ties back to people in IBM's research labs working on basic technologies, data federation and bioinformatics."

ISB researchers will use IBM's DiscoveryLink software to quickly and seamlessly integrate proteomics data in disparate formats and file types and from a variety of public and private data sources.