Abstract
Motivation Knowledge of the models of evolutionary rate variation in a phylogeny is of fundamental importance in molecular phylogenetics and systematics, not only to inform about the relationship among molecular, biological, and life history traits, but also for reliable estimation of divergence times among species and genes. Correlated and independent branch rates have emerged as two major competing models. The independent branch rate (IBR) model posits that evolutionary rates vary randomly throughout a phylogeny, in contrast to the alternative that these rates are correlated (CBR). However, currently available statistical tests lack sufficient power to reject the IBR model, which has caused many controversies because very different biological inferences are produced by the use of these models.
Results We have developed a new method (CorrTest) to accurately detect the correlation of branch rates in large phylogenies. CorrTest is computationally efficient, and it performs better than the available state-of-the-art methods. CorrTest's application to multigene and genome-scale sequence alignments from mammals, birds, insects, metazoans, plants, fungi, and prokaryotes, suggests that DNA and amino acid sequence evolutionary rates are correlated throughout the tree of life. These findings suggest concordance between molecular and non-molecular evolutionary patterns and will foster unbiased and precise dating of the tree of life.
Availability and Implementation The R source code of CorrTest is freely available for download at https://github.com/cathyqqtao/CorrTest.
Contact s.kumar@temple.edu
Supplementary information All empirical datasets, results, and source code for generating each figure are available at https://github.com/cathyqqtao/CorrTest. All simulated datasets are available on request.