Adaptive Molecular Evolution

This chapter reviews statistical methods for detecting adaptive molecular evolution by comparing synonymous and nonsynonymous substitution rates in protein-coding DNA sequences. A Markov process model of codon substitution, which forms the basis for all later discussions in this chapter, is introduced first. The case of comparing two sequences to estimate the numbers of synonymous (dS) and nonsynonymous (dN ) substitutions per site is then considered. The maximum likelihood (ML) method and a number of ad hoc counting methods are evaluated. The rest of the chapter deals with joint analyses of multiple sequences on a phylogeny. Review is done on Markov models of codon substitution that allow the nonsynonymous/synonymous rate ratio (ω = dN /dS) to vary among branches in a phylogeny or among amino acid sites in a protein. Those models can be used to construct likelihood ratio tests (LRTs) to identify evolutionary lineages under episodic Darwinian selection or to infer critical amino acids in a protein under diversifying selection. Real-data examples are used to demonstrate the application of the methods. The chapter finishes with a discussion of the limitations of current methods.