Phylogenomic analysis of carangimorph fishes reveals flatfish asymmetry arose in a blink of the evolutionary eye

Richard C. Harrington, Brant C. Faircloth, Ron I. Eytan, W. Leo Smith, Thomas J. Near, Michael E. Alfaro, Matt Friedman

Flatfish cranial asymmetry represents one of the most remarkable morphological innovations among vertebrates, and has fueled vigorous debate on the manner and rate at which strikingly divergent phenotypes evolve. A surprising result of many recent molecular phylogenetic studies is the lack of support for flatfish monophyly, where increasingly larger DNA datasets of up to 23 loci have either yielded a weakly supported flatfish clade or indicated the group is polyphyletic. Lack of resolution for flatfish relationships has been attributed to analytical limitations for dealing with processes such as nucleotide non-stationarity and incomplete lineage sorting (ILS). We tackle this phylogenetic problem using a sequence dataset comprising more than 1,000 ultraconserved DNA element (UCE) loci covering 45 carangimorphs, the broader clade containing flatfishes and several other specialized lineages such as remoras, billfishes, and archerfishes. We present a phylogeny based on UCE loci that unequivocally supports flatfish monophyly and a single origin of asymmetry. We document similar levels of discordance among UCE loci as in previous, smaller molecular datasets. However, relationships among flatfishes and carangimorphs recovered from multilocus concatenated and species tree analyses of our data are robust to the analytical framework applied and size of data matrix used. By integrating the UCE data with a rich fossil record, we find that the most distinctive carangimorph bodyplans arose rapidly during the Paleogene (66.0-23.03 Ma). Flatfish asymmetry, for example, likely evolved over an interval of no more than 2.97 million years. The longstanding uncertainty in phylogenetic hypotheses for flatfishes and their carangimorph relatives highlights the limitations of smaller molecular datasets when applied to successive, rapid divergences. Here, we recovered significant support for flatfish monophyly and relationships among carangimorphs through analysis of over 1,000 UCE loci. The resulting time-calibrated phylogeny points to phenotypic divergence early within carangimorph history that broadly matches with the predictions of adaptive models of lineage diversification.