Book of Abstracts: Albany 2009
June 16-20 2009
© Adenine Press (2008)
Pushing the Frontiers of Atomic Models for Protein Structure Prediction
Protein folding considered as the holy grail of molecular biology continues to remain elusive even after six decades of the discovery of secondary structures. While significant advances have been made in tertiary structure prediction via knowledge-base driven Bioinformatics methodologies, all atom models, which promise a physico-chemical understanding of the folding and detection of new folds, have yet to mature to be predictive.
We describe here an energy based computer software suite for narrowing down the search space of tertiary structures of small globular proteins. The protocol comprises eight different computational modules that form an automated pipeline. The software suite initially predicts the secondary structure starting from the sequence and generates multiple trial structures by varying the dihedrals of the residues in the loops. It combines biophysical filters (1) with physics based potentials (2) to arrive at five plausible candidate structures. The methodology has been validated here on 50 small globular proteins (< 100 amino acids) consisting of 2-3 helices and strands with known tertiary structures. For each of these proteins, a structure within 3-7 Å RMSD (root mean square deviation) of the native has been obtained in the five lowest energy structures within 1-3 hours on a 64 processor cluster. The protocol has been web enabled and is accessible at http://www.scfbio-iitd.res.in/bhageerath (3). Further developments in the trial structure generation protocol are in progress at present to reduce the computational times involved and to improve the prediction accuracy for proteins with higher complexity both in terms of sequence length as well as number of secondary structure units. The accuracies and limitations of the server along with some new developments will be presented and discussed.
References and Footnotes
Dept of Chemistry & Supercomputing Facility