Protein Secondary Structure Prediction Using Neural Network

Abstract

Proteins are biomolecules that play a very important role in the functioning of living organisms and have three dimensional structures that are fully specified by sequence of amino acids. The three-dimensional protein structure determines the functional properties of the protein. But the tertiary structure of the protein cannot be predicted directly from the sequence so secondary structure can be used to predict the tertiary structure of the protein because secondary structure prediction represents an intermediate step in this process and may be determined from sequence alone. The goal of prediction of protein structure is to drug discovery, to uncover the biological information and to use this information to enhance the standard of life for mankind. There are lots of methods for the prediction of secondary structure of protein from the sequences. In this paper neural network method has been used to predict the protein structure and by training the neural network the complex structure of the protein can be identified. There are different types of neural networks but in this back propagation neural network have been used to predict the protein structure. The result shows that a back propagation neural network provides 60% accuracy as compared to the NMR and X-ray diffraction. The NMR and X ray diffraction provides 100% accuracy and are best methods for protein structure prediction but these methods are expensive and time consuming. So many other methods have been developed and the result shows that neural network method provides better accuracy than other methods.