Resolving incongruences in insect phylogenomics: A reply to Boudinot et al. (2023)

Abstract

Over the last two decades, advances in molecular phylogenetics have established a new understanding of beetle phylogeny. However, some historically contentious relationships, particularly among early-diverging beetle clades, remain to be resolved. In a recent paper (Cai et al., 2022), we identified model-dependent signals in beetle phylogeny and showed how the removal of the most compositionally heterogeneous sites, in combination with the use of across-site compositionally heterogeneous models leads to results that are more congruent with the distribution of morphological characters and the beetle fossil record. In their reply, Boudinot et al. (2023) suggested that our analyses are affected by a range of shortcomings, encompassing almost every aspect of our study. Unfortunately, the arguments presented by Boudinot et al. (2023) are based on misinterpretation of the results of statistical tests, as well as misconceptions concerning substitution models, model testing and its role in phylogenomics. Here we clarify these misconceptions and show that the critiques raised by Boudinot et al. (2023) have no merit.