3.1 | Population structure
Strong genetic structure was uncovered among the R. flavipesindividuals from fastSTRUCTURE, with K = 4 best explaining the structure in the data (Fig. 1). At this value of K, most individuals in the dataset (57.2%) were clearly assigned to one of the four clusters (assignment probability higher than 99%) (73.3% of individuals were assigned to a unique cluster probability higher than 80%). However, the strong genetic structure uncovered among individuals in the native range was inconsistent with their geographic origin, as neighboring samples often exhibited completely different assignment profiles (Fig. 1). In the French introduced range, most samples could be assigned to the same cluster, although some samples from the Paris region had a mixed assignment; a similar mixed assignment was found for the lone German sample. A comparable pattern was observed in the Chilean introduced range, with most samples displaying fixed assignments and only a few with mixed assignments. Although most individuals were assigned to a unique genetic group within each introduced population (France, Chile and Canada), the three introduced populations were separately assigned to three different genetic groups and did not segregate into a single ‘introduced’ cluster. Because the genetic clustering of the native range did not consistently align with geographic origin, inferring a source population for each introduced population becomes difficult. For example, most samples from Chile were assigned to the same cluster as samples from New York, Wisconsin and Texas. Similarly, although the introduced population in France shared its strongest tie to the native range with Arkansas, France also had ties with Louisiana, Missouri and even one sample in South Carolina. The origin of the samples in Canada was even more complicated, as the genetic cluster present in this population was spread across most native localities.
Similar results to that of fastSTRUCTURE were uncovered using the PCA and DAPC approach (Figure 2). The PCA indicated strong differentiation across R. flavipes samples, as they broadly segregated along the two axes. For most localities, genetic clustering was not correlated with geography, as samples from a given locality did not always cluster together. Likewise, no genetic similarity was observed between geographically neighboring localities. Interestingly, such a pattern was also found to a lesser extent in the introduced populations, especially between France and Chile (only a single sample was available from Germany and Uruguay, and just two from the Bahamas). In France, most of the samples segregated together, except for six individuals clustering separately from the rest of the main population. These samples correspond to the individuals from the Paris region exhibiting a mixed fastSTRUCTURE assignment. A similar pattern was observed for the samples from Chile, with three samples clustering apart from the main Chilean population. The find.clusters algorithm found the best support for four genetic clusters in the dataset (Figure 2). Notably, the introduced localities of R. flavipes did not cluster together; instead, the different introduced populations were spread across the four different DAPC clusters, with some even split between two clusters (Chile and France). Remarkably, a similar pattern was observed from localities within the native range, with samples from a given locality clustering into two (e.g., Texas, Mississippi, Wisconsin) or even three (Louisiana) distinct DAPC clusters.
The co-ancestry matrix highlighted similar patterns when clustering individuals based on their level of relatedness (Figure 3). Using fineRADstructure, all samples from a given locality were not more related to one another than they were to samples from another locality (Figure 3). This result is indicative of an absence of geographic structure in the native range, as most localities were disjunct on the co-ancestry matrix. Notably, the same pattern was observed for the introduced populations, with clustering observed in two (Canada) or three (France and Chile) distinct co-ancestry groups.