Parental-effect gene-drive elements under partial selfing, or why doCaenorhabditisgenomes have hyperdivergent regions?

Abstract

ABSTRACTSelf-fertileCaenorhabditisnematodes carry a surprising number ofMedeaelements, alleles that act in heterozygous mothers and cause death or developmental delay in offspring that don’t inherit them. At some loci, both alleles in a cross operate as independentMedeas, affecting all the homozygous progeny of a selfing heterozygote. The genomic coincidence ofMedeaelements and ancient, deeply coalescing haplotypes, which pepper the otherwise homogeneous genomes of these animals, raises questions about how these apparent gene-drive elements persist for long periods of time. Here I investigate how mating system affects the evolution ofMedeas, and their paternal-effect counterparts,peels. Despite an intuition that antagonistic alleles should induce balancing selection by killing homozygotes, models show that, under partial selfing, antagonistic elements experience positive frequency dependence: the common allele drives the rare one extinct, even if the rare one is more penetrant. Analytical results for the threshold frequency required for one allele to invade a population show that a very weakly penetrant allele, one whose effects would escape laboratory detection, could nevertheless prevent a much more penetrant allele from invading under high rates of selfing. Ubiquitous weak antagonisticMedeasandpeelscould then act as localized barriers to gene flow between populations, generating genomic islands of deep coalescence. Analysis of gene expression data, however, suggest that this cannot be the whole story. A complementary explanation is that ordinary ecological balancing selection generates ancient haplotypes on whichMedeascan evolve, while high homozygosity in these selfers minimizes the role of gene drive in their evolution.