dN(eutralist) > dS(electionist)? Part 1

In a new series of posts, I will now proffer neutralist and selectionist reviews of recent publications. I point readers to an excellent review of the debate by Masatoshi Nei (2005). Besides being a fun exercise in PoV’s, I hope that these posts will prompt some healthy discussions on the state-of-the-science.

First blood – efficacy of purifying selection, prompted by publications I reported on from this post, and another recent publication by Do et al. (2015). I will leave the debate over ‘efficiency’ versus ‘efficacy’ of selection for another day (eg, see this thread).

What is purifying selection?

Simply put, the purging of deleterious (harmful) allelic variants from a population is termed purifying selection. It is also termed negative selection (to signify the sign of the selection coefficient at such alleles).

How is the efficacy of purifying selection measured?

The efficacy of purifying selection depends on at least two major contributors – the selection coefficient (s) of the mutant allele (a measure of its fitness and viability in the population), and the effective population size (Ne – a measure of the amount of drift in the population). A large portion of the neutralist-selectionist debate focused on the contributions of these two factors to molecular evolution as against mutation itself. Consequently, alleles under purifying selection would be expected to have greater levels of synonymous polymorphisms, than non-synonymous polymorphisms (dS > dN), and be maintained in populations (particularly in genes involved in reproductive fitness). Studies that have examined the differences in polymorphisms across modern humans in Africa and Europeans have determined (a) smaller Ne in human populations outside of Africa (bottlenecks), and subsequently (b) lower efficacy of purifying selection in removing weakly deleterious mutant non-synonymous substitutions.

What did Do et al. (2015) do?

They define a statistic called Rx|y which is a ratio of the number of derived mutations that are seen in one genome X, and not in genome Y, to the vice versa. This measures the relative number of mutations accumulated (given that mutation rate is the same across X and Y), compared to an outgroup (here Pan troglodytes), and should be an indicator of the efficacy of purifying selection. If selection has been equally effective, Rx|y should be = 1. Looking across a variety of datasets (exomes, and whole genomes) from West Africans versus Europeans (and through simulations), Do et al. find that Rx|y is “indistinguishable” from 1.

Thus, our data provide no evidence that purging of weakly deleterious mutations has been less effective in Europeans than in West Africans

Supporting the nearly neutral theory, since most of the genome is comprised of alleles segregating neutrally (or nearly neutrally, meaning under very weak balancing selection), the fate of a new mutant allele (and the rate at which selection affects the Rx|y statistic) is determined by drift on two different classes of alleles – non-synonymous and synonymous sites during the course of, and after a population bottleneck (which also supports previous observations of dS > dN in Europeans versus West Africans).

Simulations showing variation of the R statistic versus selection coefficients in modern humans. Figure 2 from Do et al. (2015) - courtesy: http://www.nature.com/ng/journal/v47/n2/full/ng.3186.html?WT.ec_id=NG-201502

Simulations showing variation of the R statistic versus selection coefficients in modern humans. Figure 2 from Do et al. (2015) – courtesy: http://www.nature.com/ng/journal/v47/n2/full/ng.3186.html?WT.ec_id=NG-201502

Analyses of deep sequenced genomes of a Denisovan and Neandertal to estimate the same Rx|y statistic, compared to modern humans showed (a) smaller population sizes in both archaic populations, (b) correspondingly lower levels of genetic diversity in both, but (c) faster accumulation of deleterious mutations in Denisovans than in Neandertals since their divergence, apparently contradicting the observation in modern humans (bottlenecked ancestral populations undergoing an overall depletion of non-synonymous sites, which are affected to a greater degree by purifying selection).

From a selectionist PoV, this implies that a good demographic model (which parametrizes the length of bottlenecks, and a joint distribution of selection and dominance coefficients) is required before denouncing differences in genomic load between populations.

Interested in this debate? Leave your comments below!


Do, Ron, et al. “No evidence that selection has been less effective at removing deleterious mutations in Europeans than in Africans.” Nature genetics (2015). http://dx.doi.org/10.1038/ng.3186

Nei, Masatoshi. “Selectionism and neutralism in molecular evolution.” Molecular biology and evolution 22.12 (2005): 2318-2342. http://dx.doi.org/10.1093/molbev/msi242

Dobzhansky, Theodosius. Genetics of the evolutionary process. Vol. 139. New York: Columbia University Press, 1970.


About Arun Sethuraman

I am a computational biologist, and I build statistical models and tools for population genetics. I am particularly interested in studying the dynamics of structured populations, genetic admixture, and ancestral demography.
This entry was posted in evolution, genomics, mutation, population genetics, theory and tagged , , . Bookmark the permalink.