2.4 Phylogenetic analyses
We conducted a Bayesian time calibrated analysis on our Sanger data set (c-mos and cyt b ) in Beast v. 2.5.2 (Bouckaert et al., 2019). We used a relaxed log-normal clock and a Yule tree prior assuming a constant lineage birth rate. The species Farancia erytrogramma ,Micrelaps muelleri , and Contia longicaudae were used as outgroups to allow the use of two fossils for calibration, one at the Elapoidea + Colubridae node (minimum age: 30.9 Mya), and one at theHeterodon + Farancia node (minimum age: 12.08 Mya), with fossil ages and placement based on Head, Mahlow, and Müller (2016). Two runs of 100,000,000 generations were conducted and logged every 10,000 generations. Convergence was assessed using Tracer v. 1.7 (Rambaut, Drummond, Xie, Baele, & Suchard, 2018). A burn-in of 10% was used to create a maximum clade credibility tree. Node ages are based on median tree heights.
We analyzed our SNP data set, including all samples of both species ofToxicodryas , using both species-tree summary quartet and maximum likelihood phylogenetic methods. The quartet method was implemented through SVDquartets (Chifman & Kubatko, 2014) in PAUP* v. 4.1a166 (Swofford, 2003). We sampled all possible quartets and assessed support using 100 nonparametric bootstraps and species tree topology was summarized with DendroPy v. 4.4.0 (Sukumaran & Holder, 2010). We ran a maximum likelihood analysis of our genus-wide SNP data set in IQtree v. 1.6.12 (Nguyen, Schmidt, von Haeseler, & Minh, 2014) using 10,000 ultrafast bootstraps (Hoang et al., 2018) and the ModelFinder function to choose the best substitution model (Kalyaanamoorthy, Minh, Wong, von Haeseler, & Jermiin, 2017).