Multiple resampled genomic matrices provide mixed support for arachnid monophyly

Genome assemblies for thousands of species make a resolved tree of life achievable, though this goal is hindered by computational burden and poor modelling of large datasets. We have developed an approach to systematically identify robust and challenging nodes in ancient phylogenies using multi-protein, clade-specific matrices containing different taxa and different genes. In each matrix, we include one randomly selected species for each of the major clades, infer single-copy orthologs, and analyse the resulting concatenated supermatrices using mixture models. We assess node support for competing topologies under different strategies, such as removing distant relatives and recoding, using amino acid and nucleotide data. We applied this approach to chelicerates, a group with an unresolved phylogeny in which the position of marine horseshoe crabs as the sister of arachnids is contentious. Furthermore, we also analysed two other ancient animal clades: molluscs and vertebrates, as examples of better resolved groups. While their phylogenies show stable support, chelicerate phylogenies are method and dataset -dependent. Our results suggest that, despite varying levels of support for alternative hypotheses, the evolutionary history of chelicerates remains unresolved even with the genomic data currently available.