Exploring the phylogeny of the Leptanillinae (Hymenoptera: Formicidae) through high-throughput sequencing and Bayesian total-evidence inference, with accommodation of systematic bias

Ants of the subfamily Leptanillinae (Hymenoptera: Formicidae) are minute
and subterranean, with taxonomy impeded by the dissociation of disparate
male and female forms. The advent of phylogenomic inference can remedy
this dissociation while resolving leptanilline phylogeny with strong
statistical support. However, genome-scale molecular data are vulnerable
to systematic biases that may result in decisive statistical support for
erroneous conclusions, and previous phylogenomic study of the
Leptanillinae did not address this, nor attempted to resolve the
phylogenetic positions of the several aberrant leptanilline lineages for
which UCEs cannot be obtained. I query the phylogeny of the Leptanillinae
from ultra-conserved elements (UCEs) using a Bayesian framework and
mitigate the effects of several potential systematic biases by replicating
analyses using multiple curated UCE alignments. In addition, I implemented
Bayesian total-evidence inference from UCEs and male morphological data to
resolve the phylogenetic positions of the monotypic former
genus Scyphodon and three other enigmatic terminals for which
molecular data are unavailable. The phylogeny of the Leptanillinae
inferred herein is robust to compositional biases in phylogenomic data.
The synonymy of Scyphodon with Leptanilla is confirmed, with the
former being recovered within the Leptanilla havilandi species-group with
high Bayesian posterior probability, but the phylogenetic positions of the
remaining three morphology-only terminals were unresolved by Bayesian
total-evidence inference due to insufficient availability phylogenetic
information or inferential artifacts.