The Build-Up of Population Genetic Divergence along the Speciation Continuum during a Recent Adaptive Radiation of Rhagoletis Flies

Abstract

New species form through the evolution of genetic barriers to gene flow between previously interbreeding populations. The understanding of how speciation proceeds is hampered by our inability to follow cases of incipient speciation through time. Comparative approaches examining different diverging taxa may offer limited inferences, unless they fulfill criteria that make the comparisons relevant. Here, we test for those criteria in a recent adaptive radiation of the Rhagoletis pomonella species group (RPSG) hypothesized to have diverged in sympatry via adaptation to different host fruits. We use a large-scale population genetic survey of 1568 flies across 33 populations to: (1) detect on-going hybridization, (2) determine whether the RPSG is derived from the same proximate ancestor, and (3) examine patterns of clustering and differentiation among sympatric populations. We find that divergence of each in-group RPSG taxon is occurring under current gene flow, that the derived members are nested within the large pool of genetic variation present in hawthorn-infesting populations of R. pomonella, and that sympatric population pairs differ markedly in their degree of genotypic clustering and differentiation across loci. We conclude that the RPSG provides a particularly robust opportunity to make direct comparisons to test hypotheses about how ecological speciation proceeds despite on-going gene flow.

Keywords: Tephritidae; clines; ecological speciation; gene flow; hybridization; sympatric speciation.

Figure 5

Summary of genetic and phenotypic differentiation in the R. pomonella species complex, including mean Nei’s D between each population of that taxon and each hawthorn-race R. pomonella within the respective region, the presence of range-wide clustering in population genetic distance networks, estimated migration from previous studies, the range of frequencies for alleles private to each taxon relative to R. pomonella, the presence of evidence for chemosensory behavioral differentiation from previous studies, and the relationship of host plant phenology to that of co-occurring R. pomonella. Subscripts denote the reference from which data was extracted: 1 = [50]; 2 = [22]; 3 = [87]; 4 = [88]_; 5 = [117]; 6 = [118]; 7 = [33]; 8= [119]; 9 = Linn CE, personal communication; 10 = [29]; 11 = [120].

Figure 1

Mean estimated allelic richness (±SE) across 19 microsatellite loci for R. pomonella species complex members, based on fitting rarefaction curves to a asymptotic model. Lowercase letters reflect significant differences based on Tukey’s post hoc test. R. cornivora, to the right of the dashed lines, lacks error bars and post hoc test significance designations because only a single population of this taxon was analyzed.

Figure 2

Neighbor-joining network based on Nei’s D (1972) estimated from microsatellite allele frequencies for 19 loci for 32 populations of in-group RPSG taxa. Open squares = hawthorn-infesting R. pomonella; closed squares = apple-infesting R. pomonella; open circles = snowberry-infesting R. zephyria; closed circles = blueberry-infesting R. mendax; pentagons = flowering dogwood fly. R. cornivora is omitted from this network due to poorly supported positioning and a branch length that obscures the primary features of the in-group network. The full network, including R. cornivora, is presented in Figure S4. Bootstrap values come from 10,000 replicates across loci. Numbers within nodes refer to sites listed in Table 1.

Figure 3

Patterns of genetic differentiation at paired sympatric comparisons between R. pomonella and taxa in the R. pomonella species group: (A) barplots of K = 2 structure analyses for each set of paired populations and (B) plots of the distribution locus-specific genetic differentiation for each paired population, measured as D_EST.

Figure 4

Distribution of estimated selection coefficients needed to maintain observed frequency differences between microsatellite allele classes under gene flow for each pairwise sympatric comparison. (A) Estimated selection coefficients against ancestral (hawthorn) allele in populations of the derived taxa and (B) estimated selection coefficients against allele favored in the derived population in the sympatric hawthorn population.