Fig. 9

Partial eukaryotic organism tree showing widespread losses of ST-like and/or Aquatic GDPD-like SMaseD/PLD genes, according to a model of ancient duplication followed by vertical descent. Colors indicate nodes and branches retaining both paralogs (purple), ST-like only (red), Aquatic only (blue), or neither (black). Branch labels indicate important clades: M, Metazoa; C, Cnidaria; D, Deuterostomia; P, Protostomia; S, Spiralia; E, Ecdysozoa; A, Arthropoda; H, Hexapoda (see Materials and Methods). Each taxon (phylum or class in most cases) is annotated with the number of genomes containing a PLD gene, divided by the total number of NCBI representative genomes assembled at scaffold level or higher (May 2018). Taxa marked with * also include additional genomes in which PLD genes were detected in unassembled (or assembled but not NCBI-deposited) genomic data. Taxa marked with ^ showed no PLD genes in NCBI representative genomes but did have hits in transcriptomes from multiple genera. Taxa marked with ~ showed one or several hits but with contamination suspected. This tree topology and evolutionary model imply 15 losses of ST-like and 18 losses of Aquatic genes, and additional losses would be necessary to explain incomplete conservation within certain taxa. While such extensive loss seems unlikely, the insets (dashed boxes) show distributions within Anthozoa and Rotifera that are consistent with at least three ST-like or Aquatic gene loss events in these phyla alone. The Anthozoan tree includes genera with transcriptome data only (italicized). The presence of scattered proteobacterial homologs in the Aquatic clade supports a role for lateral gene transfer in contributing to the observed sparse species distribution