Bayesian evolutionary rate estimate and dated trees
Root-to-tip regression analysis of the temporal signal from datasets revealed a moderate association between genetic distances and sampling years.
The best fitting models, Correlation coefficient, the estimated mean value of the evolutionary rate and the most common recent ancestor (TMRCA) for each dataset are provided in Table 2. Due to the low Correlation coefficient, the phylodinamic analysis was not performed on dataset VII.
In the bayesian tree, built using complete genome (dataset II; Supplementary Figure 1), two main clades were found: one included HEV-3e and HEV-3f sequences separated back to 1952 (95% HPD: 1918-1988) and the other HEV-3a, HEV-3c, HEV-3l, HEV-3n and unclassified strains originated back to 1936 (95% HPD: 1890-1971). The classification of Italian strains based on Maximum likelihood tree correlated with the Bayesian tree.
In dataset III (Supplementary Figure 2), built with HEV-3e and HEV-3f ORF2 short genomic regions, the WB110VT2017 clustered with wild boar strains previously reported (WB161VT2017, MK889004; WB125VT2017, MK889005, WB02VT2016, MG582608) in the same time period and with two human strains reported in 2018 and 2019 (INMI_1902_2019, MN537876; ISS_ID_258/2018, MZ274246). The WB03VT2016 clustered with a correlated wild boar strain WB02VT2016 (MG582608). Both wild boar and human strains were detected in the same region (Lazio).
Similarly, the 55863/2018 strain clustered with other wild boar strains reported from the same region (Umbria region, e.g.: 59068.5/UM/2018, LR699563). However, 55863/2018 clustered also with a human strain from Lazio (ISS_ID_305/2019, MZ274253) and the cluster also included a swine HEV-3 strain (14RS333-8_50549-4, MK532935) detected in Northern Italy in 2014.
In dataset IV (Supplementary Figure 3), built with the few sequences of HEV-3a ORF2 available for Italian strains, two clusters were apparent. The WB01VT2016 strain clustered with two strains sequenced from wild boar hunted in the same area in 2012 (Hepatitis_E_virus_isolate_5, MH836549; Hepatitis_E_virus_isolate_116, MH836544) and with the strain INMI_1736_2017 (MN444837) isolated from a patient in 2017; all strains found in the same region (Lazio). The second HEV-3a cluster comprised all of the sequences isolated from wild boar in northern Italy (Supplementary Figure 3).
In dataset V (Supplementary Figure 4), built with HEV-3c ORF2, two main clusters were apparent: the first grouping Italian swine sequences reported in northern Italy and the second including Italian human and wild boar sequences collected in central and northern Italy. The strain WB119VT2017 clustered with several wild boar and human strains reported in central Italy (Lazio, Abruzzo and Umbria).
In the bayesian tree built with HEV-3n ORF2, dataset VI (Supplementary Figure 5), two main clades were observed: the first one only formed by wild boar strains reported in northern Lazio (Viterbo province) between 2016/2017 and including the wild boar strain WB171VT2017 from this study; the second clade included strains isolated from wild boar in Umbria (land border with northern Lazio) in 2016/2017 and wild boar strains from southern Italy collected in 2017.