The genomic architecture of ecological speciation

Speciation reshapes the ways genetic diversity is distributed in the genome — it’s been said that the establishment of reproductive isolation is essentially the evolution of genome-wide linkage disequilibrium. The “genomic islands of speciation” model of ecological isolation imagines genome-wide differentiation spreading outward from individual genes that experience selection for different variants in different environmental conditions. But the ways in which genes under differential selection are arranged in the genome, and how variation at those genes is assorted, also alters the opportunity for isolation to evolve.
A recent Molecular Ecology paper digs into this latter scenario, using linkage mapping and association genetics in a classic case of ecological isolation, the apple maggot fly Rhagoletis pomonella. Rhagoletis pomonella lays its eggs, and its larvae feed, inside the fruits of hawthorn. When European colonists arrived in North America and started planting domestic apple trees, some hawthorn flies discovered that apples were tasty, too, and they occasionally laid eggs on those. These apple-eating flies multiplied, and by about the middle of the nineteenth century they were numerous enough to attract attention as a pest in the orchards of the Hudson River valley in New York, and they’ve been spreading westward ever since.

Apples produce fruit earlier than hawthorn, and this necessitates a shift in the timing of the life cycle for flies using apples. Since apple flies are mating and laying eggs earlier than hawthorn flies, this difference in phenology creates reduced opportunity for flies using the different host plants to mate — and it’s brought them to the brink of speciation.
The phenological differences between the apple and hawthorn “host races” of R. pomonella are mediated by differences in the winter phase of the flies’ lifecycle. The flies overwinter as pupae, putting their development on hold to withstand freezing temperatures, in a state of diapause. Apple flies enter a more intense diapause — they need longer exposure to warm temperature to “wake up” — than hawthorn flies, and then after winter they exit diapause earlier to take advantage of their earlier-fruiting host. The degree to which those two diapause traits, initial intensity and termination timing, are free to evolve independently places a potential limit on how much R. pomonella host races can diverge from each other, and what new hosts they can colonize.
Diapause intensity and timing might be non-independent, or genetically correlated, if they’re controlled by the same genes, or if they’re controlled by different genes that are so close to each other in the genome that recombination cannot easily separate variation in intensity genes from variation in timing genes. This second scenario can be addressed with a linkage map, which describes the frequency of recombination between genetic markers distributed across the genome.
In the new paper, Gregory Ragland and coauthors present and analyze a linkage map for R. pomonella, and examine the degree to which SNP markers showing associations to diapause timing and intensity are closely linked to each other, or whether the two traits are “modular” — free to evolve independently. The diapause intensity SNPs were candidates from a prior experimental evolution study of the trait; to identify diapause timing SNPs, Ragland et al. conducted genome-wide association using GEMMA, which can control for linkage relationships among tested markers as well as population structure.
The linkage mapping assigned about 4,200 SNPs to five different linkage groups, probably corresponding to five of the six R. pomonella chromosomes. (The sixth is a “heterochromatic dot” that apparently doesn’t contain much sequence variation.) Four of the five linkage groups have core clusters of “high LD” SNPs, which all show reduced recombination relative to each other — the authors say they suspect these clusters represent major structural polymorphisms of the chromosomes, though the data presented in this paper can’t really determine that directly.
Diapause timing candidates are among the SNPs assigned to three of the linkage groups, and included in “high LD” clusters; but SNPs that showed association to diapause intensity were distributed on all five linkage groups, and in none of the clusters. That’s consistent with two rather different genetic architectures for these two phenology traits, and it would, indeed, make them “modular.” Selection to match two different host plants is pulling the two traits in coordinated directions, though — at a single field site where populations of apple- and hawthorn-associated flies coexist, Ragland et al. found that SNPs with stronger association to either trait showed greater allele frequency differences between flies from the different hosts. Independent genes shaping each component of winter phenology may make ecological isolation easier to establish, and it means that a larger portion of the genome is under divergent selection arising from using the different host fruits.
References
Bush GL. 1969. Sympatric host race formation and speciation in frugivorous flies of the genus Rhagoletis (Diptera, Tephritidae). Evolution, 23: 237-251. doi: 10.1111/j.1558-5646.1969.tb03508.x
Egan SP, GJ Ragland, L Assour, TH Powell, GR Hood, S Emrich, P Nosil, and JL Feder, 2015. Experimental evidence of genome‐wide impact of ecological selection during early stages of speciation‐with‐gene‐flow. Ecology Letters, 18(8), pp.817-825. doi: 10.1111/ele.12460
Ragland GJ, MM Doellman, P Meyers, GR Hood, SP Egan, TH Powell, DA Hahn, P Nosil, JL and Feder. 2017. A test of genomic modularity among life history adaptations promoting speciation‐with‐gene‐flow. Molecular Ecology, 26: 3926–3942. doi: 10.1111/mec.14178