We have been investigating the use of hardware available at the University of Minnesota Supercomputer Institute for speeding up data-base searches which are related to the interpretation of mass spectral data. We have concentrated on the Lutefisk/CIDentify package developed by Rich Johnson and Alex Taylor because it is freely available. A parallel version has been developed for running the program on a 32 processor Linux cluster and a multi-processor SGI Origin. Work is in progress on an IBM SP parallel version.

As mass spectrometry becomes of increasing importance in the proteomics area, it becomes desirable to do mass spectral data interpretation "on-the-fly" by adjusting instrumental parameters while the experiment is in progress in order to maximize the likelihood of a successful analysis.

A large Linux cluster is an ideal test-bed to investigate the feasibility of constructing a suitable small cluster for operation in a laboratory environment. We envision that a small, rack-mounted, cluster could be built which would have significant capabilities for using large databases for use in mass spectral interpretation as well as for direct instrument control.

Practical aspects of this work will also be discussed including performance and database issues, as well as a brief general introduction to mass spectrometry and its use in proteomics research.