As sponges and ctenophores are such disparate animals¹³, the nature of the first diverging animal lineage has implications for the evolution of fundamental animal characteristics. Adult sponges are generally sessile filter-feeding organisms with body plans organized into reticulated water-filtration channels, structures built out of silica or calcium carbonate, and specialized cell types and tissues used for feeding, reproduction and self-defence, but they lack neuronal and muscle cells¹⁵. By contrast, ctenophores are gelatinous marine predators that move using eight longitudinal ‘comb rows’ of ciliary bundles^16,17; they are superficially similar but unrelated to cnidarian medusae^13,18 and possess multiple nerve nets¹⁹. Thus, whereas the sponge-sister scenario suggests a single origin of neurons on the ctenophore–parahoxozoan stem, the ctenophore-sister scenario implies either that either ancestral metazoan neurons were lost in the sponge lineage, or that there was convergent evolution of neurons in the ctenophore and parahoxozoan lineages^3,6. Similar considerations apply to other metazoan cell types¹⁸, gene regulatory networks, animal development^13,18 and other uniquely metazoan features.

Despite its importance for understanding animal evolution, the relative branching order of sponges, ctenophores and other animals has proven to be difficult to resolve². The fossil record is largely silent on this issue as verified Precambrian sponge fossils are extremely rare²⁰ and putative fossils of the soft-bodied ctenophores are difficult to interpret²¹. Morphological characters of living groups (for example, choanocytes of sponges) are not sufficient to resolve the question because true homology is difficult to assign, and such characters are easily lost or can arise convergently^13,22. The ctenophore-sister hypothesis is supported by a pair of gene duplications shared by sponges, bilaterians, placozoans and cnidarians but not ctenophores²³. Although sophisticated methods for sequence-based phylogenomics have been developed and applied to increasingly large molecular datasets, there is still considerable debate about the relative position of sponges and ctenophores as results are sensitive to how sequence evolution is modelled¹¹, which taxa or sites are included^24,25, and the effects of long-branch artifacts and nucleotide compositional variation²⁶. New approaches are needed.

We reasoned that patterns of synteny, classically defined as chromosomal gene linkage without regard to gene order²⁷, could provide a powerful tool for resolving the ctenophore-sister versus sponge-sister debate. Chromosomal patterns of gene linkage evolve slowly in many lineages^12,28,29,30, probably because it is improbable for interchromosomal translocations to be fixed in populations with large effective population sizes^28,31,32. Notably, some changes in synteny are effectively irreversible. For example, when two distinct ancestral synteny groups are combined onto a single chromosome by translocation, and subsequent intrachromosomal rearrangements mix these two groups of genes, it is very unlikely that the ancestral separated pattern will be restored by further rearrangement and fission, in the same sense that spontaneous reduction in entropy is improbable¹². Such rare and irreversible changes are particularly useful for resolving challenging phylogenetic questions as they give rise to shared derived features that unambiguously unite all descendant lineages^33,34,35. Deeply conserved syntenies observed between animals and their closest unicellular relatives¹² suggest that outgroup comparisons could be used to infer ancestral metazoan states and polarize changes within animals to address the sponge-sister versus ctenophore-sister debate. Yet, chromosome-scale genome sequences of the unicellular or colonial eukaryotic outgroups closest to animals (choanoflagellates, filastereans and ichthyosporeans) have not been reported.