Models of coding sequence evolution

Probabilistic models of sequence evolution are in widespreaduse in phylogenetics and molecular sequence evolution. Thesemodels have become increasingly sophisticated and combined withstatistical model comparison techniques have helped to shedlight on how genes and proteins evolve. Models of codon evolutionhave been particularly useful, because, in addition to providinga significant improvement in model realism for protein-codingsequences, codon models can also be designed to test hypothesesabout the selective pressures that shape the evolution of thesequences. Such models typically assume a phylogeny and canbe used to identify sites or lineages that have evolved adaptively.Recently some of the key assumptions that underlie phylogenetictests of selection have been questioned, such as the assumptionthat the rate of synonymous changes is constant across sitesor that a single phylogenetic tree can be assumed at all sitesfor recombining sequences. While some of these issues have beenaddressed through the development of novel methods, others remainas caveats that need to be considered on a case-by-case basis.Here, we outline the theory of codon models and their applicationto the detection of positive selection. We review some of themore recent developments that have improved their power andutility, laying a foundation for further advances in the modelingof coding sequence evolution.