Uncovering the genomic basis of an extraordinary plant invasion

Abstract

Invasive species are a key driver of the global biodiversity crisis, but the drivers of invasiveness, including the role of pathogens, remain debated. We investigated the genomic basis of invasiveness in Ambrosia artemisiifolia (common ragweed), introduced to Europe in the late 19th century, by resequencing 655 ragweed genomes, including 308 herbarium specimens collected up to 190 years ago. In invasive European populations, we found selection signatures in defense genes and lower prevalence of disease-inducing plant pathogens. Together with temporal changes in population structure associated with introgression from closely related Ambrosia species, escape from specific microbial enemies likely favored the plant's remarkable success as an invasive species.

Fig. 1.. Population structure in A. artemisiifolia.

(A) PCA of A. artemisiifolia samples. NA populations are defined on the basis of genetic clustering and geography. Dark pink, modern West NA; light pink, historical West NA; dark orange, modern Mideast NA; light orange, historical Mideast NA; dark turquoise, modern South NA; light turquoise, historical South NA; dark blue, modern East NA; light blue, historical East NA; dark gray, modern Europe; light gray, historical Europe. Circles, historical herbarium samples; triangles, contemporary samples. (B) Genetic structure estimated by pairwise F_ST (weighted) between native-range populations and Europe. Populations are split by time period (historical and modern). Shading of the boxes corresponds to the F_ST value, with yellow boxes indicating low F_ST and purple boxes indicating high F_ST.

Fig. 2.. Admixture proportions for A. artemisiifolia populations.

The NGSadmix run of a joint analysis including all samples with the highest likelihood for K = 9 was used for plotting. The same color scheme was used across all panels. (A, B, E, and F) Admixture maps. Samples within 100 km were grouped together, and the average ancestry across those groups was plotted. If samples were grouped together, then ancestry values were plotted at the centroid of the group. (C, D, G, and H) Admixture barplots. Each bar represents one individual. Samples are grouped on the basis of their geography: W, West; M, Mideast; E, East; S, South; O, other. (A) Historical North America. (B) Historical Europe. (C) Historical North America. (D) Historical Europe. (E) Modern North America. (F) Modern Europe. (G) Modern North America. (H) Modern Europe.

Fig. 3.. Genetic structure of spatial groups.

(A) MDS plot of pairwise F_ST between spatial groups (stress = 0.17). Gray circles, historical Europe; orange triangles, modern Europe; blue squares, historical North America; green crosses, modern North America. (B) Location of spatial groups in North America. (C) Location of spatial groups in Europe. (D) Admixture barplots for K = 9 of European spatial groups. (E) Admixture barplots for K = 9 of NA spatial groups. If the spatial group was split into an older (collected before 1900) and younger (collected after 1900), then the top row shows the older historical time period, and the middle row shows the younger historical time period. Otherwise, the middle row shows the historical time period. Bottom row: Modern time period. For the admixture barplots, the same color scheme as in Fig. 2 was used. Admixture barplots are from a joint analysis run including all samples and were split into different panels.

Fig. 4.. Heterozygosity.

(A) Heterozygosity of individuals within NA and European populations. NA populations are defined on the basis of genetic clustering and geography. Color scheme is the same as in the PCA (Fig. 1). Samples with nuclear genome coverage below 0.5× after MAPQ 25 filtering were removed as downsampling experiments showed that the heterozygosity estimate at such low coverage diverges from the true heterozygosity (see the Supplementary Materials). (B) P values of Mann-Whitney U test between heterozygosity levels of different populations. Samples with coverage below 0.5× are removed. P values below 0.05 are highlighted in yellow, and P values below 0.01 are highlighted in green.

Fig. 5.. Selection scan.

(A and B) Significantly enriched GO terms in the F_ST outlier windows. The size of the circles represents the number of significant genes annotated for the respective GO term. The color represents the P value for the enrichment, with yellow representing high and purple low P values. The gene ratio is the number of genes with significantly enriched GO terms in F_ST outlier windows divided by the total number of genes in F_ST outlier windows. (A) Enriched GO terms for historical versus modern Europe. (B) Enriched GO terms for modern Europe versus modern North America. (C and D) Fay and Wu’s H in F_ST outlier (yellow) and nonoutlier (green) windows. The boxplots show the median and first and third quantiles. (C) Outlier windows for F_ST between historical and modern Europe. (D) Outlier windows for F_ST between modern Europe and modern North America.

Fig. 6.. Pathogen presence.

(A and B) Prevalence of Xanthomonas spp. in contemporary samples. Samples within 100 km were grouped together. The pie chart indicates the fraction of samples in which Xanthomonas spp. are present, with black indicating no Xanthomonas species identified. The color indicates how many different Xanthomonas species were identified at a location. (A) Modern North America. (B) Modern Europe. (C) Venn diagram of pathogens identified in modern European (green), modern NA (yellow), historical European (blue), and historical NA (orange) samples. (D) Venn diagram of pathogens identified in modern European (green) and modern NA (yellow) samples. (E) Venn diagram of pathogens identified in historical European (blue) and historical NA (orange) samples.