Multi-locus models of deleterious epimutation-selection balance and the evolution of recombination modifiers.

Epigenetic inheritance has the potential to act as an additional source of phenotypic variation in natural populations of organisms. Previous theoretical works have investigated the deleterious potential of epigenetic variation at single loci. Nevertheless, epialleles are often associated with a genetic background, including the presence or absence of nearby transposons. The presence of transposons may indirectly generate epialleles at nearby loci as a byproduct of transposon regulation through methylation. Like genetic mutation, the epialleles at loci near transposons are expected to be deleterious, on average. Transposons are also known to be potentially deleterious upon insertion near genes, such that transposon insertion and subsequent epigenetic modification of nearby loci have been termed a "double edged" sword. In order to gain a deeper biological understanding of this process, this paper investigates theoretically deleterious epimutation-transposition balance. In contrast to previous theoretical works, this model is multi-locus, which gives rise to additional fundamental biological processes in comparison to single-locus models. In particular, recombination rates are a major element of multi-locus models. Our analytical and numerical results suggest the evolution of reduced recombination as a result of the balance of transposition and epimutation; however, we also find conditions where recombination may evolve to increase, such as due to a high background epiallele reversion rate, as found in mammals. Additionally, since selfing reduces the effective recombination rate, there may be a potentially critical interaction between mating systems and transposition-epimutation deleterious effects.