Overview of data sets and performed phylogenetic analyses

The phylogenomic data sets mintax4 and mintax8 comprise a broad range of ortholog protein sequences detected using HaMStR and selected using MARE (keeping only genes that are available for at least four or eight taxa, respectively). The most purposive subset (MPS) data set comprises a fraction of the ortholog proteins in the mintax4 data set, that were at least partially covered by at least one representative per syndermatan subgroup plus one non-syndermatan species. Three additional data sets were compiled by excluding ribosomal proteins from the former mentioned protein selections (mintax4_noRPs, mintax8_noRPs, and MPS_noRPs). To account for potential long-branch attraction (LBA) errors, a data set comprising rather slowly evolving genes out of the mintax4 data set was compiled (mintax4_slow). Data set mintax4 was also modified by keeping only one species per syndermatan subgroup (*_4Synd) and subsequently diminished by deleting “singletons” and “dingletons” (*-DS). For each data set, the numbers of protein sequences (“# proteins”) and amino acid positions (“# aa”) are indicated. Performed analyses are encoded by numbers for the used programs (1 = RAxML, 2 = TreeFinder, 3 = MrBayes, 4 = PhyloBayes) and letters for the used substitution models (A = 8 partitions, B = 7 partitions, C = CAT, D = LG+I+G+F, E = rtREV+I+G+F). For details see text and Supporting Information.