Rewinding the molecular clock in the genus Carabus (Coleoptera: Carabidae) in light of fossil evidence and the Gondwana split: a re-analyses.

Background: Molecular clocks have become powerful tools given increasing sequencing and fossil resources. However, outcome of calibration analyses depend on choosing priors. Here we revisit a seminal dating study of the genus Carabus by Andujar et al. proposing that their prior choices need re-evaluation with the hypothesis that reflecting fossil evidence and the Gondwanan split properly rewinds the molecular clock significantly. We used the same dataset including five mitochondrial and four nuclear DNA fragments with 7888 nt total length. We set the root age based on the fossil evidence of Harpalinae ground beetles in the Upper Cretaceous and introduce the Paleogene divergence of the outgroup taxa Ceroglossus (endemic to South-America) and Pamborus + Maoripamborus (Australia, New Zealand) as a new prior based on current paleontological and geological literature. Results: The ultrametric time-calibrated tree of the extended nd5 dataset resulted in a median TMRCA Carabus age of 58.48 Ma (HPD95% 46.61-72.04), roughly 35 Ma older than in the Andujar study. The splits between C. rugosus and C. morbillosus (A), between C. riffensis from the European Mesocarabus (B), and between Eurycarabus and Nesaeocarabus (C) were dated to 19.19 (13.54-25.87), 25.95 (18.8-34.62), and 23.98 (17.28-31.47) Ma and were thus decidedly older than previously reported (7.48, 10.93, and 9.51 Ma). These changes were driven solely by constraining the Carabidae time tree root with Harpalinae amber fossils at ~99 Ma. Utilizing the nd5 dating results of three well supported Carabus clades as secondary calibration points for the complete MIT-NUC data set lead to a TMRCA of Carabus of 53.56 (41.25-67.05) Ma compared to 25.16 (18.41-33.04) in Andujar9s study. Conclusion: Taking into account the Gondwanan split as a new prior, together with the fossil evidence of the outgroup taxon Harpalini in the Late Cretaceous, our new approach supports an origin of the genus Carabus in the Paleocene-Early Eocene. Our results are preliminary due to the heavy reliance on the nd5 gene and thus will have to be tested with sufficient set of nuclear markers. In addition, uncertainties arise from dating the root age of the tree based on a single fossil and outgroup taxon which has a major effect on the results. Improvement of the fossil data base particularly in the supertribe Carabitae is thus strongly needed to reduce the currently large uncertainties in dating Carabus phylogeny.