Nucleotide Substitution Model Selection Is Not Necessary for Bayesian Inference of Phylogeny With Well-Behaved Priors

Model selection aims to choose the most adequate model for the statistical analysis at hand. The model must be complex enough to capture the complexity of the data but should be simple enough not to overfit. In phylogenetics, the most common model selection scenario concerns selecting an adequate substitution and partition model for sequence evolution to infer a phylogenetic tree. Previously, several studies showed that substitution model under-parameterization can bias phylogenetic studies. Here, we explored the impact of substitution model over-parameterization in a Bayesian statistical framework. We performed simulations under the simplest substitution model, the Jukes-Cantor model, and compare posterior estimates of phylogenetic tree topologies and tree length under the true model to the most complex model, the $\text{GTR}+\Gamma+\text{I}$ substitution model, including over-splitting the data into additional subsets (i.e., applying partitioned models). We explored 4 choices of prior distributions: the default substitution model priors of MrBayes, BEAST2, and RevBayes and a newly devised prior choice (Tame). Our results show that Bayesian inference of phylogeny is robust to substitution model over-parameterization and over-partitioning but only under our new prior settings. All 3 current default priors introduced biases for the estimated tree length. We conclude that substitution and partition model selection are superfluous steps in Bayesian phylogenetic inference pipelines if well-behaved prior distributions are applied and more effort should focus on more complex and biologically realistic substitution models.