Origin and higher-level diversification of acariform mites – evidence from nuclear ribosomal genes, extensive taxon sampling, and secondary structure alignment

BACKGROUND: Acariformes is the most species-rich and morphologically diverse radiation of chelicerate arthropods, known from the oldest terrestrial ecosystems. It is also a key lineage in understanding the evolution of this group, with the most vexing question whether mites, or Acari (Parasitiformes and Acariformes) is monophyletic. Previous molecular studies recovered Acari either as monophyletic or non-monophyletic, albeit with a limited taxon sampling. Similarly, relationships between basal acariform groups (include little-known, deep-soil 'endeostigmatan' mites) and major lineages of Acariformes (Sarcoptiformes, Prostigmata) are virtually unknown. We infer phylogeny of chelicerate arthropods, using a large and representative dataset, comprising all main in- and outgroups (228 taxa). Basal diversity of Acariformes is particularly well sampled. With this dataset, we conduct a series of phylogenetically explicit tests of chelicerate and acariform relationships and present a phylogenetic framework for internal relationships of acariform mites. RESULTS: Our molecular data strongly support a diphyletic Acari, with Acariformes as the sister group to Solifugae (PP =1.0; BP = 100), the so called Poecilophysidea. Among Acariformes, some representatives of the basal group Endeostigmata (mainly deep-soil mites) were recovered as sister-groups to the remaining Acariformes (i. e., Trombidiformes + and most of Sarcoptiformes). Desmonomatan oribatid mites (soil and litter mites) were recovered as the monophyletic sister group of Astigmata (e. g., stored product mites, house dust mites, mange mites, feather and fur mites). Trombidiformes (Sphaerolichida + Prostigmata) is strongly supported (PP =1.0; BP = 98–100). Labidostommatina was inferred as the basal lineage of Prostigmata. Eleutherengona (e. g., spider mites) and Parasitengona (e. g., chiggers, fresh water mites) were recovered as monophyletic. By contrast, Eupodina (e. g., snout mites and relatives) was not. Marine mites (Halacaridae) were traditionally regarded as the sister-group to Bdelloidea (Eupodina), but our analyses show their close relationships to Parasitengona. CONCLUSIONS: Non-trivial relationships recovered by our analyses with high support (i.e., basal arrangement of endeostigmatid lineages, the position of marine mites, polyphyly of Eupodina) had been proposed by previous underappreciated morphological studies. Thus, we update currently the accepted taxonomic classification to reflect these results: the superfamily Halacaroidea Murray, 1877 is moved from the infraorder Eupodina Krantz, 1978 to Anystina van der Hammen, 1972; and the subfamily Erythracarinae Oudemans, 1936 (formerly in Anystidae Oudemans, 1902) is elevated to family rank, Erythracaridae stat. ressur., leaving Anystidae only with the nominal subfamily. Our study also shows that a clade comprising early derivative Endeostigmata (Alycidae, Nanorchestidae, Nematalycidae, and maybe Alicorhagiidae) should be treated as a taxon with the same rank as Sarcoptiformes and Trombidiformes, and the scope of the superfamily Bdelloidea should be changed. Before turning those findings into nomenclatural changes, however, we consider that our study calls for (i) finding shared apomorphies of the early derivative Endeostigmata clade and the clade including the remaining Acariformes; (ii) a well-supported hypothesis for Alicorhagiidae placement; (iii) sampling the families Proterorhagiidae, Proteonematalycidae and Grandjeanicidae not yet included in molecular analyses; (iv) undertake a denser sampling of clades traditionally placed in Eupodina, Anystina (Trombidiformes) and Palaeosomata (Sarcoptiformes), since consensus networks and Internode certainty (IC) and IC All (ICA) indices indicate high levels of conflict in these tree regions. Our study shows that regions of ambiguous alignment may provide useful phylogenetic signal when secondary structure information is used to guide the alignment procedure and provides an R implementation to the Bayesian Relative Rates test. ELECTRONIC SUPPLEMENTARY MATERIAL: The online version of this article (doi:10.1186/s12862-015-0458-2) contains supplementary material, which is available to authorized users.