Protein Structure Prediction System Based on Artificial Neural Networks

J. Vanhala and K. Kaski

Methods based on the neural network techniques are among the most accurate in the secondary structure prediction of globular proteins. Here the same principles have been used for the tertiary structure prediction problem. The map of dihedral dp and V angles is divided into 10 by 10 squares each spanning 36 by 36 degrees. By predicting the classification of each residue in the protein chain in this map a rough tertiary structure can be deduced. A complete prediction system running on a cluster of workstations and a graphical user interface was developed. Keywords: artificial neural networks, protein structure prediction, distributed computing.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.