Transcriptomics provides a robust framework for the relationships of the major clades of cladobranch sea slugs (Mollusca, Gastropoda, Heterobranchia), but fails to resolve the position of the enigmatic genus Embletonia

Background: Cladobranch sea slugs represent roughly half of the biodiversity of soft-bodied, marine gastropod molluscs (Nudibranchia) on the planet. Despite their global distribution from shallow waters to the deep sea, from tropical into polar seas, and their important role in marine ecosystems and for humans (as bioindicators and providers of medical drug leads), the evolutionary history of cladobranch sea slugs is not yet fully understood. Here, we amplify the current knowledge on the phylogenetic relationships by extending the cladobranch and outgroup taxon sampling using transcriptome data. Results: We generated new transcriptome data for 19 species of cladobranch sea slugs and two additional outgroup taxa. We complemented our taxon sampling with previously published transcriptome data, resulting in a final supermatrix covering 56 species from all but one accepted cladobranch superfamilies. Transcriptome assembly using six different assemblers, selection of those assemblies providing the largest amount of potentially phylogenetically informative sites, and quality-driven compilation of data sets resulted in three different supermatrices: one with a full coverage of genes per species (446 single-copy protein-coding genes) and two with a less stringent coverage (667 genes with 98.9% partition coverage and 1,767 genes with 86% partition coverage, respectively). We used these supermatrices to infer statistically robust maximum-likelihood trees. All analyses, irrespective of the data set, indicate maximum statistical support for all major splits and phylogenetic relationships on family level. The only discordance between the inferred trees is the position of Embletonia pulchra. Extensive testing using Four-cluster Likelihood Mapping, Approximately Unbiased tests, and Quartet Scores revealed that its position is not due to any informative phylogenetic signal, but caused by confounding signal. Conclusions: Our data matrices and the inferred trees inferred can serve as a solid foundation for future work on the taxonomy and evolutionary history of Cladobranchia. The correct placement of E. pulchra, however, proves challenging, even with large data sets. Moreover, quartet mapping shows that confounding signal present in the data is sufficient to explain the inferred position of E. pulchra, again leaving its phylogenetic position as an enigma.