Time-series genomics reveals how an invasive fruit fly recovered and adapted at Europe’s western edge

The spotted wing drosophila, Drosophila suzukii, is one of the most economically damaging invasive insects of the past two decades. Native to Southeast Asia, this tiny fly lays its eggs in soft-skinned, thin-skinned fruit such as raspberries, strawberries, and cherries, and its larvae can render entire harvests unmarketable. First detected outside its native range in Hawaii in 1980, it invaded Europe and North America almost simultaneously in 2008 and has since spread across both continents. A new study, published in the journal NeoBiota, has now tracked the fly’s genomic evolution in Portugal, the westernmost edge of its European expansion, by sampling the same populations two to four years apart. The results offer a rare time-series view of an invasion in progress, revealing a dramatic founder event, a rapid genomic recovery, and a set of candidate genes that may underpin local adaptation.

Portugal is a particularly informative place to study this invasion. The fly was first recorded there in 2012, when adults were found in a raspberry greenhouse on the Alentejo coast in the south of the country. From that foothold, it spread rapidly northward across the mainland and eventually reached the Azores and Madeira archipelagos. A previous whole-genome sequencing survey of three Portuguese populations had suggested that the invasion arrived along a southern European route, with Portuguese flies genetically closer to populations from Spain, Italy, and southern France than to those from northern Europe. Yet one population, sampled from Vieira do Minho in northern Portugal in 2019 and designated PT-VM19, showed genetic patterns that did not fully fit this scenario, displaying extremely low genetic diversity that hinted at complexity in the invasion dynamics.

To resolve this puzzle, the research team returned to the same three northern Portuguese locations—Vieira do Minho, Oliveira de Azeméis, and Castelo de Paiva—in 2023, collecting 240 adult flies identified morphologically using the European Plant Protection Organization dichotomous key to prevent contamination with other Drosophila species. The specimens were surface-sterilized and stored at minus 80 degrees Celsius until DNA extraction. Because D. suzukii is so small, the team extracted DNA from pooled sets of thoraxes, combining five sets of 16 individuals (8 females and 8 males) per population to generate a single pooled DNA sample per location. This pooling strategy, combined with coverage-aware filtering in downstream analyses, was designed to reduce biases arising from unequal individual contribution or variation in sequencing depth among samples.

The pooled libraries were sequenced on an Illumina NovaSeq platform with paired-end 150-base-pair reads, yielding average median coverages of 83-fold for autosomes and 63-fold for the X chromosome in the new samples. The reads were aligned to the Dsuz_RU_1.0 reference genome, a highly complete assembly with a scaffold N50 of 26.7 megabases and a BUSCO completeness score of 98.8 percent. Mapping quality was consistently high across all populations, with at least 99 percent of reads mapped and more than 95 percent proper pairs. The new data were combined with previously sequenced Portuguese samples from 2019 and 2021, as well as a worldwide dataset of 22 populations retrieved from an earlier global study, creating a genomic time series spanning multiple years at the invasion’s range edge.

The diversity estimates confirmed that PT-VM19 was a unique case. Its markedly lower genetic diversity, measured both as Watterson’s theta and as nucleotide diversity, was statistically significant for both autosomes and the X chromosome and was not observed in any other Portuguese population. Crucially, when the same location was resampled four years later, the resulting population PT-VM23 showed diversity levels comparable to those of the other two Portuguese populations. This suggests that the low diversity of the older sample was due to a founder effect—the genetic bottleneck that occurs when a new population is established by a small number of colonizers—and that the 2019 sample was likely taken shortly after the establishment of the population. A high genome-wide Tajima’s D value for PT-VM on the autosomes, 1.67 plus or minus 1.72, further supports the interpretation that a founder effect was occurring at the time of the earlier sampling.

The rapid recovery of diversity in PT-VM23 likely reflects key biological characteristics of the species. D. suzukii has a short generation time of roughly 12 to 15 days, large effective population sizes estimated between 2.5 times 10 to the fifth and 2.5 times 10 to the sixth individuals, and a high dispersal capacity, all of which can homogenize diversity levels quickly. Gene flow from neighboring populations probably contributed as well. The TreeMix analysis, which models population relationships from allele frequency covariance, detected an admixture signal connecting the PT-CP23 population and the node leading to PT-VM19. Although this seems counterintuitive given that PT-VM19 was sampled four years earlier, the authors note that TreeMix does not account for sampling time and that this signal may instead represent shared ancestry, gene flow from unsampled intermediate populations, or historical admixture affecting both populations.

The population structure analyses confirmed the Mediterranean invasion route. In the worldwide dataset, the scaled covariance matrix of allele frequencies grouped the 28 populations into two distinct clusters, one comprising the Asian and American populations and the other the European populations, including all Portuguese samples. Within the European cluster, the Portuguese populations clustered with those from Italy, Spain, and southern France, consistent with an introduction along the Mediterranean coast. Notably, PT-VM19 formed a separate branch within the European cluster, while PT-VM23 clustered with the remaining Portuguese populations, indicating that PT-VM19 does not represent a different colonization history from the other Portuguese populations. Interestingly, when the analysis was restricted to Portuguese populations, the samples clustered by sampling year rather than by geographic location, suggesting that these populations may have evolved along a common trajectory—though an alternative explanation is that some local populations may have gone extinct and been replaced by new migrants.

The study also re-examined signatures of selection across the genome, with important implications for how such signals should be interpreted. Using a hidden Markov model approach implemented in Pool-hmm, the team identified genomic regions classified under the hidden state of selection in each population. The PT-VM populations showed by far the highest number of windows under selection, with 584 windows across sampling years, but 462 of these were driven by PT-VM19 alone. This overestimation is likely due to the population’s notably low diversity, which represents a significant deviation from demographic equilibrium and can masquerade as selection. Indeed, when the team compared the genes detected in their analyses with candidate genes previously reported for the Portuguese invasion and the global invasion, most of the overlap was attributable to PT-VM19, suggesting that many earlier signals were false positives associated with its low diversity. Even a previously suggested candidate for insecticide resistance, the nAChRalpha7 gene, was not confirmed once the second time point was added; polymorphic sites within that gene classified as under selection were detected only in PT-VM19.

However, nine genes with known functions showed consistent signatures of selection over time across all focal Portuguese populations: Sema1a and nrm, involved in neural development; hkb, involved in embryogenesis; Gpdh1, Acer, and beta-Man, involved in metabolism; H2.0, involved in chromatin organization; Uba4, which plays a critical role in protein quality control by tagging proteins for degradation; and l(2)k09848. Among these, Uba4 stands out for having functions that closely reflect the terms identified by the team’s gene ontology enrichment analysis, which highlighted protein and heparin binding, immune system processes, and molecular function regulator activity. Another gene of interest, Muc68Ca, was consistently detected across all other populations when PT-VM19 was excluded. In Drosophila melanogaster, mucin genes have been shown to influence adult size, starvation tolerance, and cold recovery, and Muc68Ca is expressed in the larval and embryonic midgut and salivary gland, where it may contribute to gut barrier integrity, nutrient absorption, or protection from environmental stress. Selection on this gene in Portuguese populations may thus reflect adaptation to thermal extremes or to biotic factors such as diet and pathogens that differ in invaded habitats, though functional studies will be required to test this hypothesis.

To translate these findings into practical tools, the team developed and validated a set of SNP markers targeting the genes with the strongest and most consistent signatures of selection: Gpdh1, Sema1a, Acer, H2.0, Uba4, and Muc68Ca. All primer pairs successfully amplified the expected regions, and sequencing confirmed the presence of the target SNPs, providing a robust set of markers for future monitoring efforts with the potential to detect temporal and spatial changes in adaptive allele frequencies. The study underscores how temporal genomic data from recently invaded regions can uncover demographic recovery and ongoing adaptation, distinguishing genuine selective signals from demographic artifacts that plague single-time-point analyses. As invasive species continue to reshape ecosystems and agriculture worldwide, approaches like this one—tracking the same populations across multiple years at the invasion’s leading edge—offer a powerful window into evolution happening on contemporary timescales, and the validated marker set gives pest managers a genomic instrument for monitoring how this destructive fly continues to adapt at Europe’s western frontier.

Subject of Research: Temporal population genomics of the invasive spotted wing drosophila at the western edge of its European range

Article Title: Temporal genomics unravels demographic and selective invasion dynamics of Drosophila suzukii at a European range edge

Article References: Carvalho, J., Marques, J. P., Sario, S., Farelo, L., Santos, C., & Melo-Ferreira, J. (2026). Temporal genomics unravels demographic and selective invasion dynamics of Drosophila suzukii at a European range edge. NeoBiota, 107, 131-153. https://doi.org/10.3897/neobiota.107.176004

Image Credits: AI Generated

DOI: 10.3897/neobiota.107.176004

Keywords: Drosophila suzukii, invasive species, population genomics, founder effect, local adaptation, selective sweeps, Portugal, spotted wing drosophila, Temporal, genomics, unravels, demographic