Global diversity of dipteran families (Insecta Diptera) in freshwater (excluding Simulidae, Culicidae, Chironomidae, Tipulidae and Tabanidae)

We sequenced nearly the entire carbomoylphosphate synthase (CPS) domain of CAD, or rudimentary, (ca. 4 kb) from 29 species of flies representing all major clades within Eremoneura, or higher flies, and several orthorrhaphous brachyceran outgroups. We compared these sequences with orthologs from Anopheles gambiae and Drosophila melanogaster to assess structure, compositional bias, and phylogenetic utility. CAD is large (6.6+ kb), complex (comprised of three major and myriad minor functional domains) and relatively free of introns, extreme nucleotide bias (except third codon positions), and large hypervariable regions. The CPS domain possesses moderate levels of nonsynonymous divergence among taxa of intermediate evolutionary age and conveys considerable phylogenetic signal. Phylogenetic analysis of CPS sequences under varying methods and assumptions resulted in well-resolved, strongly supported trees concordant with many traditional ideas about higher dipteran phylogeny and with prior inferences from 28S rDNA. The most robustly supported major eremoneuran clades were Cyclorrhapha, Platypezoidea, Eumuscomorpha, Empidoidea, Atelestidae, Empidoidea exclusive of Atelestidae, Hybotidae s.l., Microphoridae+Dolichopodidae, and Empididae s. str. Because CAD is ubiquitous, apparently single copy (at least within holometabolous insects), readily obtained from several insect orders using primers described herein, and exhibits considerable phylogenetic utility, it should have wide applicability in insect molecular systematics.

The insect order Diptera, the true flies, contains one of the four largest Mesozoic insect radiations within its suborder Brachycera. Estimates of phylogenetic relationships and divergence dates among the major brachyceran lineages have been problematic or vague because of a lack of consistent evidence and the rarity of well-preserved fossils. Here, we combine new evidence from nucleotide sequence data, morphological reinterpretations, and fossils to improve estimates of brachyceran evolutionary relationships and ages. The 28S ribosomal DNA (rDNA) gene was sequenced for a broad diversity of taxa, and the data were combined with recently published morphological scorings for a parsimony-based phylogenetic analysis. The phylogenetic topology inferred from the combined 28S rDNA and morphology data set supports brachyceran monophyly and the monophyly of the four major brachyceran infraorders and suggests relationships largely consistent with previous classifications. Weak support was found for a basal brachyceran clade comprising the infraorders Stratiomyomorpha (soldier flies and relatives), Xylophagomorpha (xylophagid flies), and Tabanomorpha (horse flies, snipe flies, and relatives). This topology and similar alternative arrangements were used to obtain Bayesian estimates of divergence times, both with and without the assumption of a constant evolutionary rate. The estimated times were relatively robust to the choice of prior distributions. Divergence times based on the 28S rDNA and several fossil constraints indicate that the Brachycera originated in the late Triassic or earliest Mesozoic and that all major lower brachyceran fly lineages had near contemporaneous origins in the mid-Jurassic prior to the origin of flowering plants (angiosperms). This study provides increased resolution of brachyceran phylogeny, and our revised estimates of fly ages should improve the temporal context of evolutionary inferences and genomic comparisons between fly model organisms.