Genetic Diversity and Thermal Performance in Invasive and Native Populations of African Fig Flies

Abstract

During biological invasions, invasive populations can suffer losses of genetic diversity that are predicted to negatively impact their fitness/performance. Despite examples of invasive populations harboring lower diversity than conspecific populations in their native range, few studies have linked this lower diversity to a decrease in fitness. Using genome sequences, we show that invasive populations of the African fig fly, Zaprionus indianus, have less genetic diversity than conspecific populations in their native range and that diversity is proportionally lower in regions of the genome experiencing low recombination rates. This result suggests that selection may have played a role in lowering diversity in the invasive populations. We next use interspecific comparisons to show that genetic diversity remains relatively high in invasive populations of Z. indianus when compared with other closely related species. By comparing genetic diversity in orthologous gene regions, we also show that the genome-wide landscape of genetic diversity differs between invasive and native populations of Z. indianus indicating that invasion not only affects amounts of genetic diversity but also how that diversity is distributed across the genome. Finally, we use parameter estimates from thermal performance curves for 13 species of Zaprionus to show that Z. indianus has the broadest thermal niche of measured species, and that performance does not differ between invasive and native populations. These results illustrate how aspects of genetic diversity in invasive species can be decoupled from measures of fitness, and that a broad thermal niche may have helped facilitate Z. indianus's range expansion.

Keywords: Zaprionus; climate adaptation; genetic diversity; genomics; invasive species; thermal performance.

Fig. 1.

Collection locations across sub-Saharan Africa, eastern North America, and Hawaii (a) and genetic differentiation among Zaprionus indianus samples (b and c). (a) All collection locations are highlighted with bold type. Zaprionus indianus was sampled from all locations, Z. tuberculatus was sampled from São Tomé and Senegal (forest site), Z. africanus from São Tomé and Kenya, and Z. inermis, Z. tsacasi, Z. taronus, and Z. nigranus from São Tomé. Invasive populations of Z. indianus are differentiated from populations in their native range (b and c), with the strongest genetic differentiation between invasive and native populations (c; PC1) and among African populations (c; PC2). Invasive populations in the eastern United States and Hawaii also show weak differentiation (c; PC3). The inset in the top right of (c) shows the genome-wide distribution of differentiation (F_ST; 5-kb genomic windows) between invasive Z. indianus and each of the four African populations (Senegal sample locations grouped as one population), and also between eastern United States and Hawaiian samples (black line).

Fig. 2.

Estimates of genetic diversity summarized across 5-kb genomic windows for each population included in this study. Colored backgrounds group populations as invasive Zaprionus indianus (three leftmost violins), native Z. indianus (five central violins), and other species (eight rightmost violins). See supplementary figure S2, Supplementary Material online, for estimates for windows overlapping with BUSCO annotations and supplementary table S4, Supplementary Material online, for estimates in all subsamples from populations where we sampled more than four individuals. Superscript 1 indicates population of Z. africanus or Z. tuberculatus sampled from São Tomé and superscript 2 indicates population of Z. africanus from Kenya or Z. tuberculatus from Senegal (forest site).

Fig. 3.

Correlations between genetic diversity (the number of segregating sites: S) and recombination rate across populations of Zaprionus indianus (a). Correlations did not systematically differ between populations in the invasive and native regions of the species’ range. However, the mean difference in diversity for a given genomic window was positively correlated with recombination rate (b). Populations with two points in (a) represent populations where we sampled more than four individuals and estimated S using two independent random subsamples of those individuals. In (b), the difference in S (mean in invasive populations−mean in native populations) was scaled by mean levels of diversity for a given recombination rate quantile.

Fig. 4.

Genetic diversity (S) is correlated across regions of the genome containing annotated single-copy orthologs in interspecific comparisons (a). The closely related species Zaprionus africanus from São Tomé and Z. indianus from Zambia had the strongest interspecific correlation in S (b; see fig. 4a for phylogeny) and the distantly related species Z. africanus from Kenya and Z. inermis from São Tomé had the weakest correlation in S (c). Genetic diversity is strongly correlated in all pairwise comparisons between populations of Z. indianus (d). (e and f) Data from the strongest and weakest between-population correlations for Z. indianus, and (g) shows the strongest correlation between invasive populations of Z. indianus. Red rectangles in (e) through (g) highlight genomic windows that have low diversity in one population (fewer than ten segregating SNPs), but higher diversity in the other (more than ten segregating SNPs).

Fig. 5.

The correlation of genetic diversity (S) across the genomes of different species (a) decreases with increasing genetic distance (b and c). Maximum likelihood phylogeny in (a) was estimated with RAxML run on an alignment of 1,709 BUSCO genes annotated across all seven species’ genomes. Correlation coefficients decreased with increasing genetic distance (Mantel r = −0.816; b), with this pattern being particularly evident for within-clade comparisons (highlighted in c).

Fig. 6.

The correlation in genetic diversity (S) between invasive populations of Zaprionus indianus and other species of Zaprionus (dark red points) is weaker than the correlation in S between native populations of Z. indianus and other species of Zaprionus (black points).

Fig. 7.

Populations and species with a larger thermal niche breadth (B₅₀) also harbor more genetic diversity (median number of segregating sites; S). Results for all populations (a) and for only the seven species sampled on the island of São Tomé (b) are shown. Species names that are not labeled with a location in (a) are from São Tomé. Results from linear models testing the relationship between thermal niche breadth and genetic diversity are reported in the bottom right of each panel.

Fig. 8.

Median and 95% credible intervals for parameters estimated by jointly fitting performance curves to thermal performance data collected at mean temperatures of 13.9–28.9 °C. Populations of Zaprionus indianus are shown to the left of the dashed line and all other species of Zaprionus to the right. For populations of Z. indianus, parameter estimates that had a probability >0.95 of being different from one another in pairwise comparisons (analogous to P < 0.05) are denoted by letters (“a” > “c” and “d,” and “b” > “d”). Red bars indicate significant differences between the population of Z. indianus from Zambia and other populations of Z. indianus.