Noroviruses cause acute gastroenteritis. Since multiple genotypes of norovirus co-circulate in humans, changing the genotype composition and eluding host immunity, development of a polyvalent vaccine against norovirus in which the genotypes of vaccine strains match the major strains in circulation in the target season is desirable. However, this would require prediction of changes in the genotype composition of circulating strains. A fitness model that predicts the proportion of a strain in the next season from that in the current season has been developed for influenza A virus. Here, such a fitness model that takes into account the fitness effect of herd immunity was used to predict genotype compositions in norovirus seasons in Japan. In the current study, a model that assumes a decline in the magnitude of cross immunity between norovirus strains according to an increase in the divergence of the major antigenic protein VP1 was found to be appropriate for predicting genotype composition. Although it is difficult to predict the proportions of genotypes accurately, the model is effective in predicting the direction of change in the proportions of genotypes. The model predicted that GII.3 and GII.4 may contract, whereas GII.17 may expand and predominate in the 2015-2016 season. The procedure of predicting genotype compositions in norovirus seasons described in the present study has been implemented in the norovirus forecasting system (NOROCAST).
Keywords: genotype composition; herd immunity; norovirus; prediction; vaccine.
© 2016 The Societies and John Wiley & Sons Australia, Ltd.