Global invasion history of the agricultural pest butterfly Pieris rapae revealed with genomics and citizen science

Abstract

The small cabbage white butterfly, Pieris rapae, is a major agricultural pest of cruciferous crops and has been introduced to every continent except South America and Antarctica as a result of human activities. In an effort to reconstruct the near-global invasion history of P. rapae, we developed a citizen science project, the "Pieris Project," and successfully amassed thousands of specimens from 32 countries worldwide. We then generated and analyzed nuclear (double-digest restriction site-associated DNA fragment procedure [ddRAD]) and mitochondrial DNA sequence data for these samples to reconstruct and compare different global invasion history scenarios. Our results bolster historical accounts of the global spread and timing of P. rapae introductions. We provide molecular evidence supporting the hypothesis that the ongoing divergence of the European and Asian subspecies of P. rapae (∼1,200 y B.P.) coincides with the diversification of brassicaceous crops and the development of human trade routes such as the Silk Route (Silk Road). The further spread of P. rapae over the last ∼160 y was facilitated by human movement and trade, resulting in an almost linear series of at least 4 founding events, with each introduced population going through a severe bottleneck and serving as the source for the next introduction. Management efforts of this agricultural pest may need to consider the current existence of multiple genetically distinct populations. Finally, the international success of the Pieris Project demonstrates the power of the public to aid scientists in collections-based research addressing important questions in invasion biology, and in ecology and evolutionary biology more broadly.

Keywords: agricultural pest; approximate Bayesian computation; citizen science; genomics; invasive.

Fig. 1.

Sample size by location and dataset. (A) ddRADseq (n = 559). (B) mtDNA (n = 1,002). The size of the points corresponds to the sample size. Explore these data further through interactive data visualizations (http://www.pierisproject.org/ResultsInvasionHistory.html).

Fig. 2.

Global invasion history and patterns of genetic structure and diversity of P. rapae. (A) Genetic ancestry assignments based on the program ADMIXTURE. (B) Rooted neighbor-joining tree based on Nei’s genetic distance. (C) Among population genetic differentiation based on Weir and Cockerham’s F_ST (64), New Zealand and Australia are treated separately. (D) Graphical illustration of divergence scenario chosen in ABC-RF analysis (Table 1). (E) Geographic representation of divergence scenario with the highest likelihood based on ABC-RF analysis; points are colored based on their population assignment using ADMIXTURE as in A, and dates represent median estimates from ABC-RF analysis. All analyses are based on 558 individuals genotyped for 17,917 ddRADseq SNPs. Explore these data further through interactive data visualizations (http://www.pierisproject.org/ResultsInvasionHistory.html).

Fig. 3.

Patterns of autosomal genetic diversity—observed heterozygosity, pairwise nucleotide diversity, and Tajima’s D—by population.

Fig. 4.

Global patterns of mitochondrial haplotype diversity. (A) Geographic distribution of all 88 mtDNA haplotypes discovered (unique color for each haplotype; explore individual haplotypes further through interactive data visualizations [http://www.pierisproject.org/ResultsInvasionHistory.html]). Note points jittered to avoid overlapping (hidden) points; thus, coordinates are approximate, and the colors used for haplotypes are unrelated to those used in other panels. (B) Haplotype network inferred using median-joining algorithm and colored by population. Hash marks between haplotypes represent base changes (mutations). (C) Number of unique mtDNA haplotypes by population as well as subpopulation estimated using a rarefaction approach (Methods) and plotted by geographic location. (D) Pairwise nucleotide diversity by population.