Genome Integrated Supercomputing Toolkit - GIST (July 2000)

The high performance biological computing resource at ORNL has reached an unprecedented level of throughput and sophistication since the release of the Joint Genome Institute chromosomes 5,16,19, and the release of the public Human Draft Genome. The main vehicle of high performance analysis is the Genomic Integrated Supercomputing Toolkit (GIST), which uses the IBM SP3 at ORNL, and integrates seamlessly genome annotation tools developed with DOE support. The tools available within GIST are MPP-BLAST for sequence alignment, protein families with MPP-Pfam, Gene Modelling GRAIL-EXP and protein structure prediction with PROSPECT.

Analysis of five microbial genomes conatining some 22,000 genes (N. punctifrome, P. Marinus, R. palustris, N. europea, E.faecium ) now takes just under 48 hours compared to weeks using conventional computing resources. Genome annotation tasks have used a total 64120 processor hours in May 2000, 54011 hours in June 2000 and 117553 hours in July 2000. The ORNL genome annotation website receives over 200,000 hits per month. (Contact: Phil Locascio, locasciop@ornl.gov, 865-574-4567; Funding Source: DOE/SC/OBER)