The phylogenetic affinities of Chaetognathifera, with considerations of systematic error and the robusticity of macrosyntenic results

Chaetognathifera, a superphylum comprising Syndermata (Rotifera including Acanthocephala), Micrognathozoa, Gnathostomulida and Chaetognatha, is a complex grouping generally recovered as the sister to all other Lophotrochozoa. However, phylogenetic relationships within this group are controversial, in part due to poor sampling, resulting in two key questions. The first is whether Gnathostomulida or Chaetognatha represent the sister group to Syndermata+Micrognathozoa. The second is the phylogenetic position of the former phylum Acanthocephala within Syndermata. Here, we present the first study of the phylogenetic affinities of Chaetognathifera with genomic representation from all major phyla, and explore the potential of macrosynteny to better understand these relationships. For this latter aspect, we also developed a new jackknifing procedure to assess the robustness of linkage groups inferred by macrosyntenic analyses. We show that the phylogenetic relationships between these clades are corroborated through a variety of gene selection and analysis methodologies. This provides clear evidence of Acanthocephala as a derived clade within Syndermata as sister to Seisonidea, and that Gnathostomulida is sister to Syndermata+Micrognathozoa, with Chaetognatha as the earliest diverging clade within Chaetognathifera. On the other hand, we found that macrosyntenic patterns cannot resolve this question. Moreover, almost all possible linkage groups involving chaetognathiferan species lack robusticity and hence, should not be considered reliable. As a consequence so far, in Chaetognathifera none of the bilaterian ancestral linkage groups can be reliably found and independent massive chromosomal rearrangements occurred. We therefore strongly suggest that studies of macrosynteny should not only assess the significance of possible linkage groups, but also the robusticity of these linkage group inferences. Furthermore, we also present a script for this purpose, which can be found at: https://github.com/JFFleming/MacrosyntenicJackknife