Phylogenetic Relationships and Divergence Times in Rodents Based on Both Genes and Fossils

Molecular and paleontological approaches have produced extremely different estimates for divergence times among orders of placental mammals and within rodents with molecular studies suggesting a much older date than fossils. We evaluated the conflict between the fossil record and molecular data and find a significant correlation between dates estimated by fossils and relative branch lengths, suggesting that molecular data agree with the fossil record regarding divergence times in rodents. Our approach includes a correction for tree hierarchy involving simulating the random appearance of fossils. We also present a ghost lineage approach that attempts to incorporate the potential for the discovery of older fossils into a Bayesian analysis of divergence dates. Applying this approach to a set of Eocene rodent fossils, we estimated the earliest divergence in rodents appears to have occurred at approximately the K/T boundary, but interordinal splits were estimated to have taken place late in the Cretaceous. We propose that some molecular clock studies may overestimate divergence times due to periods of accelerated molecular evolution across multiple lineages or due to saturation of data that is not adequately corrected by the evolutionary model. We have sequenced the complete mitochondrial genomes of three rodent species, Anomalurus beecrofti, Castor canadensis, and Dipodomys ordii, and attempt to resolve phylogenetic relationships within rodents using the mitochondrial genome, a nuclear dataset of comparable size, and a combined analysis containing 26 kbp of sequence data. The combined analysis recovered a Sciuromorpha – Hystricomorpha clade with strong support. Our data suggest that increased character sampling improves resolution at these early nodes while better taxon sampling of mitochondrial genomes has led to better supported clades that converge on conclusions obtained from nuclear datasets. Several molecular studies have concluded that the zokors, genus Myospalax, evolved from within the rodent subfamily Cricetinae. We tested this conclusion using mitochondrial data and determined that Myospalax is sister to a clade containing the subfamilies Spalacinae and Rhizomyinae, and all three of these lineages appear to be basal to the superfamily Muroidea. Based on the position of these three lineages, we suggested that they be placed in a distinct family, the Spalacidae. The murine genera Mus and Rattus are thought to have diverged about 12 million years ago (Ma) based on a series of fossils from the Siwaliks of Pakistan, but assumptions of murid relationships that led to this conclusion have been shown to be false by molecular data. Equally parsimonious hypotheses can be proposed which place the 12 million year old Progonomys fossil at the base of the family Muridae, basal to the subfamily Murinae, or at the Mus - Rattus divergence. We here test the dates of evolutionary divergences in murids. Our results indicate that the family Muridae probably diverged earlier than the Siwalik fossils, but Mus and Rattus diverged at the same time or prior to the 12 Ma fossil date. We also cannot reject the hypothesis that the 12 Ma date represents the oldest split in the Murinae instead of the more derived Mus – Rattus date. We also recovered phylogenetic results suggesting that Taterillus is related to the tribe Gerbillini and not to other genera that are treated as Taterillini and that Gerbillurus evolved from within Gerbilliscus