Supergenes are genomic regions containing sets of tightly linked loci that control multi-trait phenotypic polymorphisms under balancing selection. Recent advances in genomics have uncovered significant variation in both the genomic architecture as well as the mode of origin of supergenes across diverse organismal systems. Although the role of genomic architecture for the origin of supergenes has been much discussed, differences in the genomic architecture also subsequently affect the evolutionary trajectory of supergenes and the rate of degeneration of supergene haplotypes. In this review, we synthesize recent genomic work and historical models of supergene evolution, highlighting how the genomic architecture of supergenes affects their evolutionary fate. We discuss how recent findings on classic supergenes involved in governing ant colony social form, mimicry in butterflies, and heterostyly in flowering plants relate to theoretical expectations. Furthermore, we use forward simulations to demonstrate that differences in genomic architecture affect the degeneration of supergenes. Finally, we discuss implications of the evolution of supergene haplotypes for the long-term fate of balanced polymorphisms governed by supergenes.

No abstract

Although high levels of self-fertilization (>85%) are not uncommon in nature, organisms reproducing entirely through selfing are extremely rare. Predominant selfers are expected to have low genetic diversity because genetic variation is distributed among rather than within lineages and is readily lost through genetic drift. We examined genetic diversity at 22 microsatellite loci in 105 individuals from a population of the semelparous herb Lobelia inflata L. and found (i) no evidence of heterozygosity through outcrossing, yet (ii) high rates of genetic polymorphism (2-4 alleles per locus). Furthermore, this genetic variation among lineages was associated with phenotypic traits (e.g. flower colour, size at first flower). Coupled with previous work characterizing the fitness consequences of reproductive timing, our results suggest that temporal genotype-by-environment interaction may maintain genetic variation and, because genetic variation occurs only among lineages, this simple system offers a unique opportunity for future tests of this mechanism.