Are interphylum spiralian relationships resolvable?

While the membership of the major animal clade of Spiralia has been
relatively stable since the introduction of molecular sequence data, the
relationships between the constituent phyla are less clear. Focusing on
the five largest phyla (Annelida, Brachiopoda, Mollusca, Nemertea, and
Platyhelminthes), we find previous analyses have supported all 15 possible
unrooted trees that could relate them, suggesting a hard-to-resolve node.
We have used two recent phylogenomic data sets to explore this remarkable
example of taxonomic instability. Using a combination of taxon-jackknifing
and bootstrapping on empirical and simulated data, we explored the support
for all 105 rooted and 15 unrooted 5-phylum topologies under
site-homogeneous and site-heterogeneous substitution models. Rooted
analyses suggested that the preference for rooting Spiralia on
Platyhelminthes is due to artefactual long-branch attraction between this
clade and Gnathifera. Most unrooted analyses showed a marginal and
non-significant preference for the same 5-phylum topology. In all unrooted
trees, the branches relating to spiralian phyla were shorter, on average,
than the short (and contested) Deuterostome branch. Our results suggest
that spiralian phyla likely emerged in rapid succession, in a difficult to
resolve adaptive radiation. Resolving these very short branches will
require large data sets to overcome stochastic errors, as well as efforts
to address systematic errors arising from both branch-length and
site-compositional heterogeneity. The lack of clarity has implications for
our understanding of the clade’s history, the interpretation of Cambrian
fossils, and for clarifying the evolutionary history of traits, including
biomineralization, segmentation, and larvae.