Phylogeny and evolutionary dynamics of bee phenomes: An integrative approach to semantic-aware morphological phylogenetics

Morphologically diverse groups such as bees offer excellent opportunities to study the dynamics of complex phenotypic evolution. Bees are a clade of more than 20,000 species with an evolutionary history of more than 120 million years showing an impressive morphological and behavioral diversity. Although the group has received considerable attention in the last two decades, a comprehensive account of the phylogeny and morphological evolution of the whole clade is still lacking. In this study, we capitalize on recent advances in Bayesian phylogenetics and anatomy ontologies to infer a phylogeny and reconstruct the dynamics of the morphological evolution of bees. We produced a large data set of 124 species and 394 characters representing all families of bees including information from external and internal morphology to infer a phylogenetic hypothesis. Then, to study the dynamics of morphological evolution, we employed new developed tools from the R packages rphenoscate and OntoPhylo to annotate character statements with anatomy terms from the Hymenoptera Anatomy Ontology and build appropriate evolutionary models. Anatomy ontologies are formal representations of knowledge about anatomical entities thus allowing information on dependencies to be incorporated in models of trait evolution. This approach allows one to investigate how complex phenotypes or the entire organismal morphology evolves. The main phylogenetic results obtained are as follows: (Stenotritidae + Colletidae) + (Andrenidae + Halictidae + (Melittidae + (Apidae + Megachilidae))). Finally, we discuss how evolutionary dynamics of complex anatomies, such as mouthparts and tentorium, could potentially impact phylogenetic inferences with morphological data for bees.