Phylogenetic analyses
Phylogenetic trees were reconstructed using maximum likelihood (ML)
implemented in IQ-TREE v. 1.5.5 (Nguyen, Schmidt, Von Haeseler, & Minh,
2015) under the general time-reversible (Tavaré, 1986), four class
mixture model (GTR+FO*H4) on the unpartitioned matrix. This is a General
Heterogeneous evolution On a Single Topology model (GHOST) that infers
separate base frequencies per class and accounts for heterotachy or rate
variation across sites and lineages (Crotty et al., 2020). Bootstrap
support values (bs) were estimated via the ultrafast bootstrap algorithm
with 1500 replicates (Minh, Nguyen, & Haeseler, 2013). Bayesian
inference (BI) was assessed in ExaBayes v 1.5 (Aberer, Kobert, &
Stamatakis, 2014) with the implemented GTRGAMMA model with four coupled
Markov chain Monte Carlo (MCMC) runs, each with 10 million generations,
and sampling every 500 generations. Convergence was checked based on the
average standard deviation of split frequencies (ASDSF
<0.2%). The first 25% of the trees were discarded as
burn-in for each MCMC run prior to convergence. Topological robustness
was assessed using posterior probabilities (pp). Trees were visualized
in FigTree v. 1.4.4 (Rambaut, 2014) and edited in Adobe Illustrator CC
2018 (Fig. 1).