Phylogenetic Analysis

The phylogenetic analyses were conducted with the ML framework using IQ-TREE (1.6.12) (Minh et al., 2020). For each alignment, we used the best substitution model selected by ModelFinder (Kalyaanamoorthy et al., 2017) of IQ-TREE. The selected models are described in the figure legends. The branch support values were computed based on the non-parametric bootstrap method with 100 bootstrap replicates, as well as transfer bootstrap expectation (TBE) with BOOSTER-WEB (Lemoine et al., 2018). TBE replaces the branch presence frequency in non-parametric bootstrap proportion (i.e., the expectation of a 0/1 function) by the expectation of a nearly continuous function. TBE supports acknowledge the presence of unstable taxa (single taxa that tends to move in and out of clades) and provide an alternative view of the robustness of a tree.

In the reconstruction of the phylogenetic tree of PolBs, Poxviridae sequences have been shown to be difficult to confidently position and tend to reduce the global resolution of phylogenetic trees, and were therefore subsequently removed as suggested by Guglielmini et al. (2019) to enhance the resolution of the phylogenetic trees.

For myosin sequences from viruses and eukaryotes, we also built phylogenetic trees using RAxML (8.2.12) (Stamatakis, 2014) on both “gappy” and “strict” datasets. We tried four substitution models (i.e., PROTCATAUTO, PROTCATIAUTO, PROTGAMMAAUTO, and PROTGAMMAIAUTO) and selected the tree with the highest likelihood score. Therefore, we obtained four trees for these myosin sequences (i.e., gappy/IQ-TREE, strict/IQ-TREE, gappy/RAxML, and strict/RAxML).