Secondary Structure Prediction Using ANN Learning
In this paper, the authors present two methods for predicting the secondary structure of proteins based on artificial neural network learning. Two variations of NN learning rule are employed using feedforward backpropogation architecture for predicting secondary structure of proteins from their primary sequences of amino acids. About 500 sequences and more than 10000 patterns were trained with variable size of patterns. After initial level of training, an accuracy rate of about 70%-75% is obtained through first learning rule whereas 80-85% of accuracy is obtained using the second variation of the learning rule. Both the methods are implemented within a software tool by the name NNSec developed on Visual.NET flatform.