Re-evaluating and dating myriapod diversification with phylotranscriptomics under a regime of dense taxon sampling.

ABSTRACT

Recent transcriptomic studies of myriapod phylogeny have been based on relatively small datasets with <40 myriapod terminals and variably supported or contradicted the traditional morphological groupings of Progoneata and Dignatha. Here we amassed a large dataset of 104 myriapod terminals, including multiple species for each of the four myriapod classes. Across the tree, most nodes are stable and well supported. Most analyses across a range of gene occupancy levels provide moderate to strong support for a deep split of Myriapoda into Symphyla + Pauropoda (=Edafopoda) and an uncontradicted grouping of Chilopoda + Diplopoda (=Pectinopoda nov.), as in other recent transcriptome-based analyses; no analysis recovers Progoneata or Dignatha as clades. As in all recent multi-locus and phylogenomic studies, chilopod interrelationships resolve with Craterostigmus excluded from Amalpighiata rather than uniting with other centipedes with maternal brood care in Phylactometria. Diplopod ordinal interrelationships are largely congruent with morphology-based classifications. Chilognathan clades that are not invariably advocated by morphologists include Glomerida + Glomeridesmida, such that the volvation-related characters of pill millipedes may be convergent, and Stemmiulida + Polydesmida more closely allied to Juliformia than to Callipodida + Chordeumatida. The latter relationship implies homoplasy in spinnerets and contradicts Nematophora. A time-tree with nodes calibrated by 25 myriapod and six outgroup fossil terminals recovers Cambrian-Ordovician divergences for the deepest splits in Myriapoda, Edafopoda and Pectinopoda, predating the terrestrial fossil record of myriapods as in other published chronograms, whereas age estimates within Chilopoda and Diplopoda overlap with or do not appreciably predate the calibration fossils. The grouping of Chilopoda and Diplopoda is recovered in all our analyses and is formalized as Pectinopoda nov., named for the shared presence of mandibular comb lamellae. New taxonomic proposals for Chilopoda based on uncontradicted clades are Tykhepoda nov. for the three blind families of Scolopendromorpha that share a "sieve-type" gizzard, and Taktikospina nov. for Scolopendromorpha to the exclusion of Mimopidae.