The beetle superfamily Dytiscoidea, placed within the suborder Adephaga, comprises six families. The phylogenetic relationships of these families, whose species are aquatic, remain highly contentious. In particular the monophyly of the geographically disjunct Aspidytidae (China and South Africa) remains unclear. Here we use a phylogenomic approach to demonstrate that Aspidytidae are indeed monophyletic, as we inferred this phylogenetic relationship from analyzing nucleotide sequence data filtered for compositional heterogeneity and from analyzing amino-acid sequence data. Our analyses suggest that Aspidytidae are the sister group of Amphizoidae, although the support for this relationship is not unequivocal. A sister group relationship of Hygrobiidae to a clade comprising Amphizoidae, Aspidytidae, and Dytiscidae is supported by analyses in which model assumptions are violated the least. In general, we find that both concatenation and the applied coalescent method are sensitive to the effect of among-species compositional heterogeneity.Four-cluster likelihood-mapping suggests that despite the substantial size of the dataset and the use of advanced analytical methods, statistical support is weak for the inferred phylogenetic placement of Hygrobiidae. These results indicate that other kinds of data (e.g. genomic meta-characters) are possibly required to resolve the above-specified persisting phylogenetic uncertainties. Our study illustrates various data-driven confounding effects in phylogenetic reconstructions and highlights the need for careful monitoring of model violations prior to phylogenomic analysis. to coordination of taxon sampling and transcriptome sequencing. BM, DRM, MB, ON, RGB, and XZ, contributed to funding acquisition. DRM, DTB, FJ, HEE, KM, LH, MB, RSP, YA, and XZ collected samples and/or contributed to the data processing of the sequenced transcriptomes. AD, AV, JMP, LP, and SL performed the de novo transcriptome assembly and cross-contamination checks. AD, AV, and JMP performed the NCBI sequence submissions. AV, ON, and RMW performed the orthology inference and orthology assignment analyses. AV performed the phylogenetic analyses with contributions, suggestions and comments from BM, KM, and CM. AV, BM, ON, MB and RGB wrote the first draft of the manuscript, with AV taking the lead. All authors contributed with comments and suggestions on later versions of the manuscript.
Declarations of interest: none
Appendix A. Supplementary materialSupplementary data associated with this article can be found, in the online version, at (doi link upon acceptance). The filtered and unfiltered COGs as well as all inferred matrices and their partition files are available at the MENDELEY DATA repository (XXXXX).