Data from: The pipid root

The estimation of phylogenetic relationships is an essential component of
understanding evolution. Accurate phylogenetic estimation is difficult,
however, when internodes are short and old, when genealogical discordance
is common due to large ancestral effective population sizes or ancestral
population structure, and when homoplasy is prevalent. Inference of
divergence times is also hampered by unknown and uneven rates of
evolution, the incomplete fossil record, uncertainty in relationships
between fossil and extant lineages, and uncertainty in the age of fossils.
Ideally these challenges can be overcome by developing large
“phylogenomic” datasets, and by analyzing them with methods that
accommodate features of the evolutionary process such as genealogical
discordance, recurrent substitution, recombination, ancestral population
structure, gene flow after speciation among sampled and unsampled taxa,
and variation in evolutionary rates. In some phylogenetic problems it is
possible to use information that is independent of fossils, such as the
geological record, to identify putative triggers for diversification whose
associated estimated divergence times can then be compared a posteriori to
estimated relationships and ages of fossils. The history of
diversification of pipid frog genera Pipa, Hymenochirus, Silurana, and
Xenopus, for instance, is characterized by many of these evolutionary and
analytical challenges. These frogs diversified dozens of millions of years
ago, they have a relatively rich fossil record, their distributions span
continental plates with a well characterized geological record of ancient
connectivity, and there is considerable disagreement across studies in
estimated evolutionary relationships. We used high throughput sequencing
and public databases to generate a large phylogenomic dataset with which
we estimated evolutionary relationships using multilocus coalescence
methods. We collected sequence data from Pipa, Hymenochirus, Silurana, and
Xenopus and the outgroup taxon Rhinophrynus dorsalis from coding sequence
of 113 autosomal regions, averaging ~300 base pairs in length (range: 102
– 1695 base pairs), and also a portion of the mitochondrial genome.
Analysis of these data using multiple approaches recovers strong support
for the ((Xenopus, Silurana)(Pipa, Hymenochirus)) topology, and
geologically calibrated divergence time estimates that are consistent with
estimated ages and phylogenetic affinities of many fossils. These results
provide new insights into the biogeography and chronology of pipid
diversification during the breakup of Gondwanaland, and illustrate how
phylogenomic data may be necessary to tackle tough problems in molecular
systematics.