Research For H3 Antigenic Evolution And Variation Rule Of Influenza A Virus

Influenza A virus remains an important pathogen which threatens health of human in long time. Deeply understanding of its evolution and variation rule has very important meaning for prevention of influenza. With the help of latest developed data mining technique and phylogenetic analysis method, evolution rule of H3 antigen were explored from bioinformatics point of view, the research detail were as follows:1. With regard to the dynamic updating characteristic of bioinformatics data, data split technique were enrolled in data collection, two-step cluster and self-organization map were applied to explore the basic rule of sequence evolution. The result shows that all sequence could be split into 6 clusters, host distribution between clusters reflect the inherit distance of different host strain, while time distribution of clusters shows the variation history of H3 subtype virus in human being.2. Based on result of cluster analysis, random forest were applied to search for key variation site in H3 sequence, with the help of random permutation, 27 sites were filter out. Except for site 2, other 26 sites were associated with the antibody-combining sites. This result indicate that the variation of human influenza were mainly driven by immunize barrier of host.3. Bayesian phylogenetic tree analysis was applied to study evolution rule of H3 strain in detail. The results were highly consistent with that of cluster. Beside provide evolution history detail of H3 sequence, two importance questions were also answered: First, swine mainly play a role of deposit for old variation strains, it is not important in antigenic drift of human influenza virus. Second, as to variation of human influenza virus, south China is not a sticking point for antigenic drift, it should not be concerned as origin of new strain for same subtype.4. Based on the trunk of tree structure, Yang's codon-substitution model was applied to search positive selected sites in each cluster separately. Results shows that immunization pressure on H3A1 sequences is not constant in different period, likelihood test results indicate that pressure increased with time, and reach the climaxafter 1995. Positive selected sites also various between clusters, 28 codons, while most of them were associated with the antibody-combining sites, were identified as positive selection site on the tree trunk in different time. However, these positive sites changed markedly between adjacent clusters.Integrate the information shows above, we could draw the conclusion that mainly due to enhancement of host immune barrier, the variation speed of H3 subtype influenza virus was accelerated in these years. The core function of present immunity strategy is protecting high risk population, however, it can not prevent the emergence of new variant strain. Based on farther bioinformatics research, as well as the coherent active immunity strategy in the world, it is possible to ceasing the prevalence of related subtype. Conclusions drawn from the study should be able to serve as a knowledge base of policy-making and intervening for these propose.While answer the question of epidemiology, the research also consummate the methodology for sequence evolution rule analysis, phylogenetic tree analysis method for large data were also provided, our result indicate these methodological system could satisfy the need of related analysis, it is worthwhile to be applied in similar research.