Population genetic structure of wolves in the northwestern Dinaric-Balkan region

Abstract

The Balkan Peninsula and the Dinaric Mountains possess extraordinary biodiversity and support one of the largest and most diverse wolf (Canis lupus) populations in Europe. Results obtained with diverse genetic markers show west-east substructure, also seen in various other species, despite the absence of obvious barriers to movement. However, the spatial extent of the genetic clusters remains unresolved, and our aim was to combine fine-scale sampling with population and spatial genetic analyses to improve resolution of wolf genetic clusters. We analyzed 16 autosomal microsatellites from 255 wolves sampled in Slovenia, Croatia, Bosnia and Herzegovina (BIH), and Serbia and documented three genetic clusters. These comprised (1) Slovenia and the regions of Gorski kotar and Lika in Croatia, (2) the region of Dalmatia in southern Croatia and BIH, and (3) Serbia. When we mapped the clusters geographically, we observed west-east genetic structure across the study area, together with some specific structure in BIH-Dalmatia. We observed that cluster 1 had a smaller effective population size, consistent with earlier reports of population recovery since the 1980s. Our results provide foundation for future genomic studies that would further resolve the observed west-east population structure and its evolutionary history in wolves and other taxa in the region and identify focal areas for habitat conservation. They also have immediate importance for conservation planning for the wolves in one of the most important parts of the species' European range.

Keywords: Balkan Peninsula; Canis lupus; Dinaric Mountains; effective population size; microsatellites; population structure.

FIGURE 1

Map showing sampling localities of wolves from Slovenia, Croatia, Bosnia and Herzegovina, and Serbia, including the three regions of Croatia (Dal, Dalmatia; GK, Gorski kotar; Lika)

FIGURE 2

(a) The most likely number of clusters (K = 4) detected for wolves from Slovenia, Croatia, Bosnia and Herzegovina, and Serbia in STRUCTURE, based on log‐likelihoods [ln Pr(X|K)] and the ΔK method. (b) Bar plot from the STRUCTURE analyses showing four genetic clusters (K = 4). Each color corresponds to one cluster; each line represents one individual, showing probability of assignment (range 0–1) per cluster. (c) Bar plot from the STRUCTURE analyses showing three genetic clusters (K = 3)

FIGURE 3

(a) The geographical distribution and observed west‐east gradient of wolf genetic clusters (K = 3) based on individual q‐values in STRUCTURE. Pie chart colors designate the individual membership assignment to each cluster. Abbreviations used are BIH, Bosnia and Herzegovina; HRV‐Dal, Croatia‐Dalmatia; HRV‐GK, Croatia‐Gorski kotar; HRV‐Lika, Croatia‐Lika; SLO, Slovenia; SRB, Serbia. (b) HWDS analysis. Distance coincides with the distance on the traveling window (TW) axis in (a). The values at the bottom indicate the number of loci deviating from Hardy–Weinberg expectations at p < .05. Some of the values were removed where the points are very dense (particularly between 0 and 100 km) to improve legibility

FIGURE 4

Spatial principal component analysis (sPCA) showing genetic population structure in the study area. C1, C2: Global sPCA components 1 and 2, the size and colors of squares indicate differences in value (between individuals) for the respective component; C1 + C2 – colorplot of sPCA components 1 and 2, each mapped to a RGB color (component 1 = red, component 2 = blue). Scree plot: left—barplot of the eigenvalues; right—spatial and variance components of the eigenvalues