Figure 4: Phylogeny of hepatitis B viruses and two scenarios on the origin of mammalian HBVs.

(a,b) Consensus tree (grey branches not to scale) of maximum likelihood and Bayesian phylogenetic analyses (Supplementary Fig. S2a,b; for full alignment, see Supplementary Data 3) of the polymerase protein from extinct eZHBV HASs and extant HBVs. Note that this tree topology is corroborated by our whole-genome nucleotide sequence analysis (Supplementary Fig. S2c). Rooting the tree to mammalian HBVs (a) places the origin of Orthohepadnaviridae before the endogenization of eZHBV_C and implies ancient codivergence of avian and mammalian HBVs. Alternatively, the same tree rooted to the Mesozoic eZHBV_C EVE (b) implies the emergence of Orthohepadnaviridae after a host switch from birds to mammals at some point after the endogenization of eZHBV_C. The representatives of Orthohepadnaviridae are extant HBVs from human (HBV), woolly monkey (WMHBV), and woodchuck (WHV), complemented by representatives of Avihepadnaviridae from duck (DHBV), crane (CHBV) and heron (HHBV). Rectangles are coloured according to the respective geological epoch in the International Stratigraphic Chart and denote the temporal ranges of insertion events. Nodes are labelled with maximum likelihood bootstrap values (in %) and Bayesian posterior probabilities (italicized).