Statistical methods for detecting molecular adaptation

Abstract

The past few years have seen the development of powerful statistical methods for detecting adaptive molecular evolution. These methods compare synonymous and nonsynonymous substitution rates in protein-coding genes, and regard a nonsynonymous rate elevated above the synonymous rate as evidence for darwinian selection. Numerous cases of molecular adaptation are being identified in various systems from viruses to humans. Although previous analyses averaging rates over sites and time have little power, recent methods designed to detect positive selection at individual sites and lineages have been successful. Here, we summarize recent statistical methods for detecting molecular adaptation, and discuss their limitations and possible improvements.

Fig. 1

The identification of sites under positive selection from the sperm lysin genes of 25 abalone species. (a) Posterior probabilities for site classes with different ω ratios along the sequence. The lysin sequence of the red abalone (Haliotis rufescens) is shown below the x-axis. ML estimates under Model M3 (discrete) suggest three site classes with the ω ratios at ω₀=0.085 (grey), ω₁=0.911 (green) and ω₂=3.065 (red), and with proportions p₀=0.329, ρ₁=0.402 and ρ₂=0.269. These proportions are the prior probabilities (Box 1) that any site belongs to the three classes. The data (codon configurations in different species) at a site alter the prior probabilities dramatically, and thus the posterior probabilities might be different from the prior probabilities. For example, the posterior probabilities for Site 1 are 0.944, 0.056 and 0.000, and thus this site is likely to be under strong purifying selection. The posterior probabilities for Site 4 are 0.000, 0.000 and 1.000, and thus this site is almost certainly under diversifying selection. (b) Lysin crystal structure from the red abalone (Protein Data Bank file 1LIS), with sites coloured according to their most likely class inferred in (a). The structure starts from amino acid four (His) at the N-terminus, because the first three amino acids are unresolved. The five α-helices are indicated: α1 from amino acids 13 to 38, α2 from 44 to 74, α3 from 82 to 95, α4 from 99 to 107 and α5 from 116 to 123. Note that sites potentially under positive selection (red) are scattered all over the primary sequence but tend to cluster around the top and bottom of the crystal structure. Reproduced, with permission, from.