Development of Neural Network for BLSOM Clustering of HA Genes of Avian Influenza Viruses Isolated in Guangdong Province

摘要

A neural network classification method,and a batch-learning self-organizing map(BLSOM),was established using trinucleotide and tetranucleotide in the hemagglutinin gene sequences of 25 avian influenza viruses isolated in Guangdong Province. Statistical analysis and normalization of the fragment number were done and MATLAB function was used to simulate the human brain thinking for self-organizing learning. When the number of training steps was 100 and above,the strains could be successfully clustered. H_1,H_3,H_5,H_7 and H_9 subtype strains fell within different classes,respectively,and the HA gene cluster map of H_3N_2 and H_7N_9strains was quite similar,suggesting that these strains shared the same origin; H_5N_1 strain was quite different in different years; H_1N_1 and H_9N_2 strains could be clustered into one group,indicating the natural recombinant variation in the two kinds of viruses,thereby providing a reference for high-risk strain screening and traceability.