Three Levels of the Predicting of the Influenza Vaccine Strains

Abstract

Relevance. The influenza vaccine can reduce the incidence and mortality from influenza if that are antigenically the same or related to the viruses. There are at least 3 levels of prediction: vaccine strains for the upcoming epidemic season; the trend of influenza viruses for 2–3 years ahead, the emergence of pre-pandemic virus strains. The aim of the work was to analyze the potential of bioinformatics to implement prediction at these 3 levels for subtypes H1N1 and H3N2. Materials and methods. For the computer analysis, the database of the hemagglutinin (HA) primary structures of the H1N1 and H3N2 strains isolated in the influenza epidemiological season 2016–2017 – 2018–2019 was used from the Internet. At the first prediction level, an adapted hidden Markov model was used, at the second, invariants were searched for НА H1N1 and its antigenic sites, and at the third level, prediction was based on identifying invariants in structural proteins of pandemic strains. Results. The circulation of several dominant strains in the epidemiological season 2018–2019 was predicted, the existence of invariants in НА Н1 and its antigenic sites НА H1N1 was shown, and it was concluded that the threat of an influenza pandemic caused by avian influenza viruses was unlikely. Conclusion The bioinformatics approach can be considered as a valuable tool in predicting, at different levels of circulation, certain strains of the influenza virus in the epidemic season.