RAD sequencing sheds new light on the genetic structure and local adaptation of European scallops and resolves their demographic histories

Abstract

Recent developments in genomics are advancing our understanding of the processes shaping population structure in wild organisms. In particular, reduced representation sequencing has facilitated the generation of dense genetic marker datasets that provide greater power for resolving population structure, investigating the role of selection and reconstructing demographic histories. We therefore used RAD sequencing to study the great scallop Pecten maximus and its sister species P. jacobeus along a latitudinal cline in Europe. Analysis of 219 samples genotyped at 82,439 single nucleotide polymorphisms clearly resolved an Atlantic and a Norwegian group within P. maximus as well as P. jacobeus, in support of previous studies. Fine-scale structure was also detected, including pronounced differences involving Mulroy Bay in Ireland, where scallops are commercially cultured. Furthermore, we identified a suite of 279 environmentally associated loci that resolved a contrasting phylogenetic pattern to the remaining neutral loci, consistent with ecologically mediated divergence. Finally, demographic inference provided support for the two P. maximus groups having diverged during the last glacial maximum and subsequently expanded, whereas P. jacobeus diverged around 95,000 generations ago and experienced less pronounced expansion. Our results provide an integrative perspective on the factors shaping genome-wide differentiation in a commercially important marine invertebrate.

Figure 1

Map showing scallop sampling locations against a backdrop of variation in mean annual sea surface temperature. Blue circles represent populations from the Norwegian P. maximus group defined by Morvezen et al.. Populations from the Atlantic group defined by Morvezen et al. are subdivided for ease of interpretation into a southern Atlantic group (color coded in red) and a northern Atlantic group (color coded in green) as described in the Results. Purple circles represent P. jacobeus populations from the Mediterranean Sea. The oceanic raster visualizes a composite of annual mean sea-surface temperatures (°C) measured from the Aqua Modis satellite between 2009 and 2013. The map was generated using the program ArcGIS 10.1 (ESRI, Redlands, CA, USA; https://www.esri.com/en-us/arcgis) using public domain vector and raster spatial data from naturalearthdata.com.

Figure 2

Scatterplot showing individual variation in principal component (PC) scores derived from principal component analysis (PCA) of the genomic data. Panels (a) and (b) show results including and excluding the Mediterranean populations respectively. The amounts of variation explained by each PC are given as percentages. Samples are color coded as described in the legend of Fig. 1 and scallops from Mulroy Bay are indicated by diamonds.

Figure 3

Output of the fineRADstructure analysis of the genomic data. In the cladogram, branches are color coded according to sampling origin as shown in Fig. 1 and nodes with bootstrap support greater than 90% are marked by black points. The heat map depicts variation in pairwise coancestry among individuals according to the scale shown on the right.

Figure 4

Parameter estimates obtained from the best supported demographic model of P. maximus and P. jacobeus (see Materials and Methods and Results for details). Estimates from the model are shown as vertical lines together with their associated 95% confidence intervals (horizontal bars) and distributions of parametric bootstrap estimates (shape files). Panel (a) shows the estimated divergence time of P. maximus and P. jacobeus (in generations ago) and panel (b) shows the estimated divergence time of the Atlantic and Norwegian P. maximus groups (in generations ago). Panel (c) shows historical effective population size estimates for all three scallop lineages and panel (d) shows the corresponding contemporary effective population size estimates. In panels (c) and (d), the Atlantic and Norwegian P. maximus groups are color coded in red and blue respectively, while P. jacobeus is shown in purple. The additional grey colored lineage in panel (c) represents P. maximus prior to the divergence of the Atlantic and Norwegian groups.

Figure 5

Graphical summary of our analysis of environmentally associated population structure (see Materials and Methods for details). Panel (a) shows the importance of each environmental variable together with latitude and longitude in explaining genetic variation across populations as obtained from the gradient forest analysis. Panel (b) shows the number of loci identified in LFMM analyses of mean annual dissolved oxygen concentration (DOC, light blue) and mean annual sea surface temperature (SST, coral) and their overlap (purple). Panel (c) shows the number of loci identified in BayPass analyses of mean annual dissolved oxygen concentration (DOC, light blue) and mean annual sea surface temperature (SST, coral) and their overlap (purple). Panel (d) shows the total number of candidate loci identified by LFMM (light blue), BayPass (coral) and by both methods (purple). Panels (e) and (f) represent phylogenetic trees constructed from the neutral loci and environmentally associated loci respectively. Tree edges represent individuals and are color coded according to their population of origin as shown in Fig. 1. Nodes with bootstrap support greater than 90% are marked by black points.