Improvement of prediction of mutation positions in H5N1 hemagglutinins of influenza A virus using neural network with distinguishing of arginine, leucine and serine.

In this study, we attempt to improve the prediction of mutation position in H5N1 hemagglutinin from influenza A virus using the neural network model with distinguishing of arginine, leucine and serine. In comparison with previous logistic regression prediction model, the 3-6-1 feedforward back propagation neural network model reduced independents from seven to three without compromising the predictability, and the prediction is made according to the most similar parameterised hemagglutinin rather than the population means.