Using whole genome shotgun sequences from 92 white-tailed eagles (Haliaeetus albicilla) sampled from Greenland, Iceland, Norway, Denmark, Estonia, and Turkey between 1885–1950 and after 1990, we investigate the genomic variation within countries over time, and between countries. Clear signatures of ancient biogeographic substructure across Europe and the North‐East Atlantic are observed. The greatest genomic differentiation was observed between island (Greenland and Iceland) and mainland (Denmark, Norway and Estonia) populations. The two island populations share a common ancestry from a single mainland population, distinct from the other sampled mainland populations, and despite the potential for high connectivity between Iceland and Greenland they are well separated from each other and are characterized by inbreeding and little variation. Temporal differences also highlight a pattern of regional populations persisting despite the potential for admixture. All sampled populations generally showed a decline in effective population size over time, which may have been shaped by four historical events: I) isolation of refugia during the last glacial period 110‐115,000 years ago, II) population divergence following the colonization of the deglaciated areas ~10,000 years ago, III) human population expansion, which led to the settlement in Iceland ~1,100 years ago, and IV) human persecution and exposure to toxic pollutants during the last two centuries.

Tissue was obtained from 92 specimens: 63 contemporary and 29 historic, from six different countries. These included 12 contemporary and eight historic individuals from Greenland, 25 contemporary and two historic individuals from Iceland, 12 contemporary and 13 historic individuals from Norway, 11 contemporary and five historic individuals from Denmark, three contemporary individuals from Estonia, and one historic individual from Turkey (Figure 1). The historic specimens were sampled between 1885 and 1950 (all but the two Icelandic individuals were sampled prior to 1937), while all contemporary individuals were sampled post-1990 (full individual information is presented in Table 1).

Muscle tissue and whole blood from contemporary samples from Estonia, Denmark, and Greenland (Table 1) were stored at -20 °C until DNA extraction and were provided by the Department of Ecoscience, Arctic Research Centre, AU, Roskilde, Denmark (Estonian, Danish, and Greenland samples), Natural History Museum of Denmark, University of Copenhagen, Copenhagen, Denmark (Danish samples) and the Greenland Institute of Natural Resources, Nuuk, Greenland (Greenland samples). Whole blood samples from contemporary samples from Iceland were collected in an ongoing monitoring project of the white-tailed eagle in Iceland (led by the Icelandic Institute of Natural History) and stored in EDTA at -20 °C until DNA extraction. Whole genome shotgun DNA sequences from twelve Norwegian individuals were provided by the Department of Natural History, University Museum, Norwegian University of Science and Technology (NTNU), Trondheim, Norway. DNA extraction, library building, and sequencing of all contemporary samples are described in Hansen et al (2021, in review).

Historic samples consisting of toepad clippings, taken with disposable sterile scalpel blades, from museum samples provided by The Natural History Museum of Denmark, University of Copenhagen, Denmark; Icelandic Institute of Natural History, Reykjavik, Iceland, and Department of Natural History, NTNU University Museum, Norwegian University of Science and Technology (NTNU), Trondheim, Norway.

Historic samples from Greenland, Iceland, Denmark, Turkey, and five of the thirteen Norwegian specimens were processed at the clean laboratory facilities at the Globe Institute at the University of Copenhagen. Firstly, to prevent cross-contamination from other museum specimens, the samples were cleaned with a dilute bleach solution (ca. 5% commercial strength), then rinsed with 70% ethanol followed by molecular biology grade water performed using a proteinase-based lysis-buffer according to Gilbert et al. (2008). Each sample was added 300 µL lysing buffer including 20 µL proteinase K and incubated for 3 hours. The supernatant was purified by combining 720 µL binding buffer modified as in Allentoft et al. (2015), with 80 µL sample lysate, vortexed and centrifuged through a Monarch® DNA Cleanup Column (New England Biolabs Inc., Beverly, Massachusetts, USA). The binding step was repeated 3 times after which the column was washed with 800 µL PE buffer, from where the DNA eluded into 21.5 µL EBT buffer. Throughout the entire process, only LoBind Eppendorf tubes were used.

The remaining eight Norwegian historic specimens were processed at the Norwegian University of Science and Technology (NTNU) University Museum’s dedicated palaeo-genomics laboratory. For these, the genomic DNA extractions were performed with a Qiagen DNeasy Blood & Tissue kit. The manufacturer’s protocol was used except that the amount of proteinase K was doubled, and the lysis step incubation at 56°C was extended to 15 hours. The DNA solutions were incubated at 37°C for 10 minutes prior to elution.

For all historic samples, blunt-end Illumina shotgun sequencing libraries were prepared using the BEST protocol (Carøe et al. 2018). In both of the aDNA laboratories, extraction and library blanks were also included to monitor for contamination.

Indexed libraries from historic samples from Greenland, Iceland, Denmark, Turkey and five Norwegian specimens processed at the University of Copenhagen were paired-end sequenced on four flow cells with 2x150 bp read length at deCODE Genetics in Iceland using an Illumina NovaSeq 6000.

The purified and indexed libraries for the eight Norwegian specimens processed at NTNU were pooled and paired-end sequenced over two runs on the Illumina HiSeq 4000 platform at the NTNU Genomics Core Facility, and over one run on an Illumina NovaSeq 6000 at the University of Oslo Norwegian National Sequencing Centre.

Fastq file quality of all samples was checked using FastQC (Babraham Bioinformatics 2010), then run through AdapterRemoval v2 using standard-setting, but providing adapter sequences for samples, and using the arguments --collapse and –trimns (Schubert et al. 2016). The fastq files were mapped to the golden eagle (Aquila chrysaetos) genome (GCA_900496995.3) using bwa aln, samse, and sampe, with the flags -q 15 and -k 1 (Li and Durbin 2009). Although a white-tailed eagle genome is available, the golden eagle was deliberately chosen as the reference to minimize the potential of mapping biases derived from the fact that the available white-tailed eagle genome is not equally related to all populations studied here (the published white-tailed eagle genomes come from Greenland, UK, and Germany), thus might introduce errors in the analyses (Gopalakrishnan et al. 2017). A further benefit of aligning to the golden eagle genome is that it has been assembled to chromosome level completeness and annotated, thus enabling us to both identify and exclude sex chromosomes as needed in some of the downstream analyses, and identify the genes present in regions under selection. Picard (Broad Institute 2020) was used to remove duplicate reads. To identify likely damaged bases the base quality score was rescaled with mapDamage 2.0 (Jónsson et al. 2013). Genotypes were called using GraphTyper2 (Eggertsson et al. 2019) with standard settings. The VCF file for the 92 individuals was filtered using VCFtools, BCFtools, and VCF-annotate; SNPs had to have a minor allele count of one.