A Mutation–Selection Model of Protein Evolution under Persistent Positive Selection

Abstract

We use first principles of population genetics to model the evolution of proteins under persistent positive selection (PPS). PPS may occur when organisms are subjected to persistent environmental change, during adaptive radiations, or in host-pathogen interactions. Our mutation-selection model indicates protein evolution under PPS is an irreversible Markov process, and thus proteins under PPS show a strongly asymmetrical distribution of selection coefficients among amino acid substitutions. Our model shows the criteria ω > 1 (where ω is the ratio of non-synonymous over synonymous codon substitution rates) to detect positive selection is conservative and indeed arbitrary, because in real proteins many mutations are highly deleterious and are removed by selection even at positively-selected sites. We use a penalized-likelihood implementation of the PPS model to successfully detect PPS in plant RuBisCO and influenza HA proteins. By directly estimating selection coefficients at protein sites, our inference procedure bypasses the need for using ω as a surrogate measure of selection and improves our ability to detect molecular adaptation in proteins.

Evaluation of methods to detect shifts in directional selection at the genome scale

“…In the second setting, we simulate under a strong selection regime, with N e = 10 (i.e., differences in amino acid fitnesses between −10 and 10), and P P S = 10. This second setting resembles parameter values observed on the sites showing the strongest positive selection in [Tamuri and dos Reis, 2021], and is also similar to their own simulation settings. HA sites were simulated with different profiles for background and foreground branches, and H0 sites were simulated with PPS running both on background and foreground branches.…”

“…However, the use of penalized likelihoods would prevent us from relying on likelihood ratio tests to compute pvalues and detect positive sites. Instead, [Tamuri and dos Reis, 2021] relied on simulations to compute p-values, which is more ressource intensive and would compromise Pelican's scalability. More work is needed to investigate the benefits of using penalization in Pelican, and, if any, come up with a fast method to compute p-values or scores.…”

“…In both cases, there is a transient increase in the value of ω , followed by a decrease towards ω < 1, as represented on the right panel in 1b. Blue sub-tree is an example of persistent positive selection [Tamuri and dos Reis, 2021], where the fitness profile rapidly changes along the branch, at intervals marked with red bars. In this case, the value of d N /d S remains greater than 1 while positive selection continues.…”

“…Performance measurements are done on simulated datasets, allowing us to characterize the behaviour of the methods on a range of tree shapes, branch lengths, and number of transitions along the phylogeny. We also investigate whether the detection methods are sensitive to confounding signal generated by non-adaptive processes of molecular evolution [Ratnakumar et al, 2010, Bolívar et al, 2019, Meunier et al, 2005], or by persistent positive selection [Tamuri and dos Reis, 2021].…”

Lifestyle Transitions in Fusarioid Fungi are Frequent and Lack Clear Genomic Signatures

“…Subsequently dN/dS methods have been shown to overestimate the frequency of positive selection and underestimate the strength of negative selection in bacteria, even when selection on synonymous sites is weak ( Rahman et al 2021 ). Furthermore, using dN/dS > 1 as a signifier of positive selection has been declared arbitrary ( Tamuri and dos Reis 2021 ). As flexible dN/dS methods accounting for selection on synonymous substitutions have yet to be integrated into the widely used tools for detecting positive selection, this remains a caveat of our dN/dS analyses.…”