High resolution, high throughput protein structure prediction using IBM Blue Gene supercomputers: predicting CASP targets in record time

High-resolution 3-dimensional structures of proteins can help in functional annotation and also explain the underlying mechanisms of protein interactions. The ability to quickly and accurately predict a protein's structure from its sequence is an important challenge in computational structural biology. Critical Assessment of Techniques for Protein Structure Prediction (CASP) is a biannual community-wide competition for protein structure prediction and an important benchmark for judging the performance of the various protein structure prediction algorithm.CASP 7 ran from May 11th to August 10th 2006 and for the first time we have been able to make use of the unprecedented computational power provided by Blue Gene systems at San Diego Supercomputer Center, Argonne National Lab and IBM TJWatson with the Rosetta protein structure prediction code. We describe the results of our CASP predictions and detail how access to Blue Gene systems has impacted our protein structure design approach.