Abstract

The problem of predicting protein structure based on amino acids sequence is one of the most interesting problems in molecular biology. Determination of the protein structure based on the experimental methods is very expensive, requires significant time and needs experts from different fields. For some types of proteins, theoretically based method for structure prediction is one of the alternatives. In this paper we describe the method and results of using CB513 as a dataset suitable for development of artificial neural network algorithms for prediction of secondary protein structure with MATLAB-Neural Network Toolbox.

Key words: artificial neural networks, secondary structure prediction, MATLAB environment