Utility of various genome lengths in diversity and evolution analyses of Hepatitis E virus

The aim of this study was to investigate to what extent fragments of the HEV genome could be used for accurate diagnostics and inference of viral population-scale processes. For this, we selected all the published whole genome sequences from the NCBI GenBank and trimmed them to various fragment lengths (ORF1,2,3, ORF1, ORF2, ORF3, 493 nt in ORF2 and 148 nt in ORF2). Each of the fragment lengths was used to infer the richness and diversity of the viral sequence types, typing accuracy, and potential use in phylodynamics. The results obtained from the different fragments were compared. We observed that, generally, the longer the nucleic acid fragment used in typing, the better the accuracy in predicting the viral subtype. However, the dominant HEV subtypes circulating in Europe were relatively well classified even by the 493 nt fragment, with false negative rates as low as 8 in 1000 typed sequences. Most fragments also give comparable results in analyses of population size, albeit with shorter fragments showing a broader 95 % highest posterior density interval and less obvious increase of the viral effective population size. The reconstructed phylogenies of a heterochronous subset indicated a good concordance between all the fragments, with the major clades following similar branching patterns. Furthermore, we have used the HEV sequence data from the Netherlands available in the HEVnet database as a case study for reconstruction of population size changes in the past decades. This data showed that molecular and epidemiological results are concordant and point to an increase in the viral effective population size underlying the observed increase in incidence of acute HEV infection cases. In the absence of whole genome sequencing data, the 493 bp fragment can be used for analyzing HEV strains currently circulating in Europe, as it is informative for describing short term population-scale processes.