The Evolution of Recombination

In a recent publication, Lesecque et al (2014). provide key evidence that fills in some of the blanks to an age old question – how do recombination hotspots evolve? Their analyses of major PRDM9 (a polymorphic zinc finger protein with a DNA binding domain at recombination hotspots) target sites on genomes along hominidae indicate several interesting patterns: (a) human hotspots are relatively young, (b) they have enhanced levels of GC-biased gene conversion, (c) Denisovans have different hotspots than humans, despite relatively recent divergence, and shared PRDM9 target motifs, and evidently, (d) recombination hotspots have a fast turnover rate, indicating strong support for what’s come to be known as the ‘Red Queen Theory’ of the evolution of recombination hotspots.

In light of these neat findings, I thought it would be interesting to revisit some classical theoretical considerations on the bigger picture – the evolution of recombination itself.

 Rapid turnover of recombination hotspots: the comparative analysis of modern and archaic human genomes supports a model of Red Queen evolution.

Image Credit: Pauline Sémon doi:10.1371/image.pgen.v10.i11.g001

The primary evolutionary advantage of recombination has to do with the propagation and eventual fixation of new and/or recurrent favorable mutations (and maintaining diversity at the population level) at multiple loci – something that is slow and less likely in a non-recombinant population (where favorable mutations can be fixed in a population only if they occur in offspring of a mutant). Thus a sexually recombining population should predictably evolve faster – calling for selection for the recombination machinery (Fisher (1932), Muller (1932)). On the flip side, non-recombining populations also accumulate deleterious mutations faster, in a phenomenon commonly called ‘Muller’s rachet’ (1964).

Consequently, the rate of evolutionary change prompted by recombination depends on (a) initial frequency of mutant alleles – Crow and Kimura (1965), (b) linkage between loci – Maynard-Smith (1968), (c) effective size of the population (and drift) – Otto and Barton (2007), and (d) strength of selection due to environmental change – explained in some very interesting theory by Hill and Robertson in what’s come to be known eponymously as the ‘Hill-Robertson Effect’ (1966) – see Felsenstein (1974) for review and references.

Theory also predicts that recombination should occur more `often’ in regions of the genome closer to genes than farther away from genes (Barton (1995), Barton and Charlesworth (1998)), also called hot and cold spots respectively – see Hey (2004) for review and references. Concurrently, recombination rate determines (and perhaps is determined) by mutations (that disrupt the recombination machinery) that are in strong linkage disequilibrium with other loci under selection. This determines the efficacy of selection at linked loci.

We have also known for a while, that hotspots evolve quickly – for instance, humans and chimpanzees don’t share hotspots (eg. Ptak et al.(2004)). Hotspots are expected to be constantly under (fertility) `selection’ pressure due to biased gene conversion – wherein DNA repair uses the unbroken homologous chromosome as template during a double strand break, resulting in transmission of non-recombinant alleles (as we have discussed above).

In summary, this persistence of hotspots (and recombination) is paradoxical, involving – (a) selection at linked/epistatic loci (and mutant alleles in these loci) flanking hotspots, favoring recombinants, (b) DNA repair mechanisms and biased gene conversion constantly acting to remove hotspots, (c) an ‘arms race’ between (a) and (b), leading into Lesecque et al’s theory of interest, the “Red Queen Theory” of the maintenance of recombination hotspots (Ubeda and Wilkins (2010)).

Lesecque et al. (2014) predict that at the current rate of gene conversion, current hotspots around major PRDM9 target motifs in humans should be lost within the next 3 MYR, calling for the maintenance of recombination by conflicting selection on potentially changing its target(s).

Putting this into the context of what we’ve discussed above, it brings us to two interesting hypotheses regarding the evolution of recombination hotspots in hominids.

  1. The emergence of `new’ binding motifs for PRDM9 in humans will be expected to quickly be driven to higher frequencies due to selection for maintenance of recombination.
  2. Loss of recombinant activity around existing (ancestral) HM motifs in Denisovans (and other non-African human populations that do not contain the ‘A’ allele) was likely due to random genetic drift in small populations, rather than owing to increased biased gene conversion at ancestral motifs.

What with PRDM9 double knockout mice being sterile, one could argue very strong support for the ‘Red Queen’ theory – potentially causing relatively rapid turnover of recombination hotspots across the human genome. Testing the above hypotheses with simulation studies should yield some insights. Perhaps revisiting some of the classic theory of the evolutionary advantage of recombination (see Felsenstein (1974) for an excellent review, and more recently, Coop and Przeworski (2007)) would also help understand the complex mechanism behind the `Red Queen’ dynamics of genomic conflict.


Barton, N. H. “A general model for the evolution of recombination.” Genetical research 65.02 (1995): 123-144.

Barton, Nicholas H., and Brian Charlesworth. “Why sex and recombination?.”Science 281.5385 (1998): 1986-1990.

Coop, Graham, and Molly Przeworski. “An evolutionary view of human recombination.” Nature Reviews Genetics 8.1 (2006): 23-34.

Felsenstein, Joseph. “The evolutionary advantage of recombination.” Genetics78.2 (1974): 737-756. –

Hey, Jody. “What’s so hot about recombination hotspots?.” PLoS biology 2.6 (2004): e190.

Lesecque, Yann, et al. “The Red Queen Model of Recombination Hotspots Evolution in the Light of Archaic and Modern Human Genomes.” PLoS genetics10.11 (2014): e1004790.

Ptak, Susan E., et al. “Absence of the TAP2 human recombination hotspot in chimpanzees.” PLoS biology 2.6 (2004): e155.

Ubeda, F., and J. F. Wilkins. “The Red Queen theory of recombination hotspots.” Journal of evolutionary biology 24.3 (2011): 541-553.


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 adaptation, genomics, mutation, population genetics, theory and tagged , . Bookmark the permalink.