Understanding factors influencing patterns of genetic diversity and the population genetic structure of species is of particular importance in the current era of global climate change and habitat loss. These factors include the evolutionary history of a species as well as heterogeneity in the environment it occupies, which in turn can change across time. Most studies investigating spatio-temporal genetic patterns have focused on patterns across wide geographical areas rather than local variation, but the latter can nevertheless be important particularly in topographically complex areas. Here we consider these issues in the Sooty Copper butterfly (Lycaena tityrus) from the European Alps, using genome-wide SNPs identified through RADseq. We found strong genetic differentiation within the Alps with four genetic clusters, indicating western, central, and eastern refuges, and a strong reduction of genetic diversity from west to east. This reduction in diversity may suggest that the southwestern refuge was the largest one in comparison to other refuges. Also, the high genetic diversity in the West may result from (1) admixture of different western refuges, (2) more recent demographic changes, or (3) introgression of lowland L. tityrus populations. At small spatial scales, populations were structured by several landscape features and especially by high mountain ridges and large river valleys. We detected 36 outlier loci likely under altitudinal selection, including several loci related to membranes and cellular processes. We suggest that efforts to preserve alpine L. tityrus should focus on the genetically diverse populations in the western Alps, and that the dolomite populations should be treated as genetically distinct management units, since they appear to be currently more threatened than others. This study demonstrates the usefulness of SNP-based approaches for understanding patterns of genetic diversity, gene flow and selection in a region that is expected to be particularly vulnerable to climate change.

From each male, we used head and thorax for the extraction of genomic DNA with the E.Z.N.A. ® Tissue DNA Kit (Omega bio-tek, Germany). We followed the manufacturers’ instructions but included an additional step of RNase A treatment. Afterwards, we applied double digest restriction-site associated DNA sequencing (ddRADseq) following Trense et al. (2021) and used the restriction enzymes NlaIII and MluCI. We pooled individuals into four libraries, each containing DNA fragments from individuals with different adapter pairs. Libraries were cleaned with 1.5x volume of Sera-Mag beads. We selected 250-400 base pair (bp) fragments using a 2% gel cassette and Pippin-Prep software 4.3 (Sage Science, USA). This was followed by a polymerase chain reaction (PCR) enrichment in a 10 µL reaction with 1 µL of size selected DNA, 2 µL 5x Phusion ® HF Reaction Buffer, 0.2 µL dNTPs (10 mM), 0.1 µL (100 units) Phusion ® HF DNA Polymerase (New England BioLabs Inc, USA), nuclease free water, and 2 µL (10 µM) each of Illumina P1 and P2 primers (Peterson, Weber, Kay, Fisher, & Hoekstra, 2012). The profile of thermal cycling consisted of denaturation at 98°C for 30 sec, followed by 12 cycles with 10 sec at 98°C, 30 sec at 65°C, and 70 sec at 72°C, and an extension for 5 min at 72°C. For the final library, we used seven PCR reactions. The four libraries were sequenced on an Illumina Novaseq 6000 platform generating 150 bp paired-end reads.

Possibly the coordinates of the collection points could be used.