Using Grey Neural Network to Predict Protein Primary Structure

Recent advances in large-scale genome sequencing have led to the rapid accumulation of amino acid sequences of protein. Because there are some undetermined amino acids in these proteins, it is vitally important to develop an automated method as a high-throughput tool to timely identify these amino acids. By corresponding amino acid residues with its electrostatic charge with high coefficient on isoelectric point and net charge one by one, the protein sequence can be represented by a series of real numbers. In this paper, we construct a grey neural network, integrating the gray theory and neural network, to estimate the undetermined amino acid's value of electrostatic charge based its pre-sequence and reach the aim of predicting residues indirectly. The feasibility of the method is indicated by the actual calculation. Finally, we analyze this method's relative error.