Data from: Retracing the response of Rangifer to postglacial climate change in Arctic islands
Abstract
We compiled published datasets of mitochondrial DNA sequences which informed Quaternary population history scenarios. These scenarios were evaluated in a coalescent-based approximate Bayesian computation (ABC) modelling framework (using DIYABC RF) to test hypotheses of postglacial island (re)colonisation and to estimate timings of population divergence and admixture. Population events were compared with modelled paleo-sea ice cover and published ice sheet chronologies.
Rangifer tarandus L. play a key role in Arctic ecosystems as the most numerous and widespread large herbivore. Sea ice is vital for maintaining genetic connectivity in Arctic islands, yet the historical role of sea ice in shaping R. tarandus biogeography is unknown. We studied the role of sea ice changes and ice sheet retreat since the last glacial period in the timing of island dispersal. We compiled published datasets of mitochondrial control region sequences that informed population history scenarios, which were evaluated in a coalescent-based approximate Bayesian computation (ABC) modelling framework to test hypotheses of island (re)colonisation and to estimate divergence and admixture. Population events were compared with modelled and proxy-based paleo-sea ice cover and published ice sheet chronologies. Our analysis supports Holocene dispersal onto deglaciated Arctic islands, rather than High Arctic glacial refugia. The degree of population admixture and the effect of sea ice was dependent on regional geography and climate history. North American initial island population divergence occurred as sea ice cover was declining. A lack of strong genetic structure and the occurrence of late Holocene admixture suggest that Canadian Arctic Archipelago populations were somewhat connected by sea ice during the Holocene. The Svalbard, Franz Josef land, and West Greenland colonisations arose through long-distance dispersal. Here, divergence times occurred post-deglaciation but broadly align with subfossil-based colonisation estimates, suggesting dispersal limitation due to sea ice conditions, potentially requiring appropriate ocean currents and sea ice drift directionality and speeds. Our study sheds light on the Late Quaternary (~60 ka - present) history of Arctic island Rangifer and suggests that ice sheet retreat, sea ice, and ocean currents were important in shaping present-day genetic patterns. Regional differences in postglacial dynamics suggest that dispersal during contemporary climate change may vary regionally and depend upon diminishing connectivity provided by sea ice.
https://doi.org/10.5061/dryad.j9kd51cjf
This dataset contains data and scripts used in the analyses for the paper "Retracing the response of Rangifer to postglacial climate change in Arctic islands". Supplementary information from the paper is also provided.
We compiled published datasets of mitochondrial DNA sequences which informed Quaternary population history scenarios. These scenarios were evaluated in a coalescent-based approximate Bayesian computation (ABC) modelling framework (using DIYABC RF) to test hypotheses of postglacial island (re)colonisation and to estimate timings of population divergence and admixture. Population events were compared with modelled paleo-sea ice cover and published ice sheet chronologies.
Description of the data and file structure
The data (data.zip) include:
- "genetic data" which includes mitochondrial sequence data Fasta files (from published sources obtained from GENBANK). General naming convention: species_author_data_sequence_.FASTA. Fasta sequences are generally individual-level ("indiv"_ but files with "pseudoindiv" were haplotype-level sequences that were converted to individual-level sequences with a corresponding population identification using study metadata in script #02. Note that the mitogenome data from Hold et al. (2024) is not present but the multiple sequence alignments from their paper have been deposited on the Dryad repository (https://doi.org/10.5061/dryad.tx95x6b53).
- Subfolders:
- "metadata" contains .csv files of metadata associated with the genetic data. Includes two files of geographic centroids that were calculated in scripts 01b_extract_centroids_caribou_herds.rmd.
- "raw" contains the unprocessed metadata, as it was downloaded from the relevant publication or scraped from genbank.
- "alignment" contains aligned fasta sequences and associated metadata for the alignment. R_tarandus_seqs.fasta are the unaligned sequences, mitogenome_alignment.fasta are the sequences aligned to the Hold (2024) mitogenome sequences, with sequences unwanted for the particular analysis removed in mitogenome_alignment_subset.fasta. R_tarandus_align_BSI_trim.fas and R_tarandus_align_NAAI_trim.fas are the trimmed aligned sequences for each region. The metadata r_tarandus_whole_alignment_metadata_2025_02.csv contains metadata for the aligned sequences, and r_tarandus_mitogenome_and_CR_metadata_ORDERED.csv is ordered according to the sequence order in mitogenome_alignment_subset.fas.
- "summary_stats" contains genetic summary statistics calculated for sample populations and DIYABC populations (specified in the file name "sample"/"diyabc") for the North American Arctic islands ("naai") and Barents Sea islands ("bsi"). Statistics include Nei's GST ("[name]_nei_gst.csv") and in "[name]genetic_diversity_table.csv" number of individuals ("num_inds"), number of haplotypes ("num_haps"), gene diversity ("gene.diversity") and allelic richness ("allele.rich").
- "metadata" contains .csv files of metadata associated with the genetic data. Includes two files of geographic centroids that were calculated in scripts 01b_extract_centroids_caribou_herds.rmd.
- Subfolders:
- "diyabc_input" - DIYABC RF input files, known as header files. The .mss files are the input files for DIYABC RF software, the .txt and the .gen (genepop) files are interim data used in the construction of the .mss files (see script #05 and #06 for details). Naming convention: [location_model number_date]. We have found .mss files are most easily read by the software if you use an existing .mss file as a template and clear the contents below the first line before pasting in your new sequences from the .txt file that is generated in R (see script #05 and #06 for details). Note that the first line of each file is just a placeholder and not reflective of the actual content of the file. Note that DIYABC output data is only provided for the final set of models due to the large size of the model output (reftableRF.bin), although input data are provided for all model runs.
- "diyabc_output" - DIYABC RF output for the final preferred scenarios. The .mss file contains the sequence data. "header.rf" contains the header file with all the input and parameter specifications for the model runs. reftableRF.bin files contain the simulated data under the different scenarios. modelchoice_out.confusion shows the confusion matrix for RF model classification and modelchoice_out.predictions contains the chosen model. statobsRF.txt are the summary statistics. Each parameter that was estimated from the best scenario has various files output. See the DIYABC RF manual on how to interpret these output files. Note that DIYABC output data is only provided for the final set of models due to the large size of the model output (reftableRF.bin), although input data and scripts are provided for all model runs.
- Subfolders:
- "Final_NAAI_model" - final model set for the North American Arctic islands
- "Final_BSI_model" - final model set for the Barents Sea islands
- Subfolders:
- "interim_data" for processed sea ice concentration time series. Naming convention is [model number_optional location_demographic event code_ice fraction_radius.csv]. Monthly sea ice concentration values (min, mean, max, standard deviation) in units of fraction of 100%, within a given buffer (radius) of a location ("population centroid"). The corresponding year is n thousand years before present: the data come from 26 simulations covering the last 28,000 years run at a time resolution of 4,000-1,000 years. Population centroid locations were calcualted in scripts #10 and #12, and sea ice concentration time series were extracted from buffers in scripts #11 and #13. Paleoclimate model data from which the sea ice data were derived, including year codes, can be accessed from the repository at: www.bridge.bris.ac.uk/resources/simulations.
- Subfolders:
- "centroids" - geographical centroids for the model regions calculated in scripts #10 and #12, in the form of .rdata objects that are loaded in scripts #11 and #13.
- Subfolders:
Sharing/Access information
Paleoclimate model data can be accessed from the repository at:
Land polygon shapefiles from:
Migratory caribou range shapefiles from:
- Alaska Center for Conservation Science CC BY-SA 4.0
- NWT Species and Habitat Viewer, Environment and Climate Change, Government of the Northwest Territories, Yellowknife, NT.
Ice sheet shapefiles and maps can be accessed from:
- Batchelor, C. L., Margold, M., Krapp, M., Murton, D. K., Dalton, A. S., Gibbard, P. L., Stokes, C. R., Murton, J. B., & Manica, A. (2019). The configuration of Northern Hemisphere ice sheets through the Quaternary. Nature Communications, 10(1), Article 1. https://doi.org/10.1038/s41467-019-11601-2
- Dalton, A. S., Margold, M., Stokes, C. R., Tarasov, L., Dyke, A. S., Adams, R. S., Allard, S., Arends, H. E., Atkinson, N., Attig, J. W., Barnett, P. J., Barnett, R. L., Batterson, M., Bernatchez, P., Borns, H. W., Breckenridge, A., Briner, J. P., Brouard, E., Campbell, J. E., … Wright, H. E. (2020). An updated radiocarbon-based ice margin chronology for the last deglaciation of the North American Ice Sheet Complex. Quaternary Science Reviews, 234, 106223. https://doi.org/10.1016/j.quascirev.2020.106223
- Dalton, A. S., Stokes, C. R., & Batchelor, C. L. (2022). Evolution of the Laurentide and Innuitian ice sheets prior to the Last Glacial Maximum (115 ka to 25 ka). Earth-Science Reviews, 224, 103875. https://doi.org/10.1016/j.earscirev.2021.103875
- Hughes, A. L. C., Gyllencreutz, R., Lohne, Ø. S., Mangerud, J., & Svendsen, J. I. (2016). The last Eurasian ice sheets – a chronological database and time-slice reconstruction, DATED-1. Boreas, 45(1), 1–45. https://doi.org/10.1111/bor.12142
Genetic data was derived from the following sources:
| Publication | Publication title | Publication DOI | Number of populations | Number of individuals | Spatial coverage |
|---|---|---|---|---|---|
| Gravlund et al., 1998 | Polyphyletic Origin of the Small-Bodied, High-Arctic Subspecies of Tundra Reindeer (Rangifer tarandus) | 10.1006/mpev.1998.0525 | 8 | 98 | Circumpolar |
| Flagstad & Røed, 2003 | Refugial Origins of Reindeer (Rangifer tarandus L.) Inferred from Mitochondrial Dna Sequences | 10.1111/j.0014-3820.2003.tb01557.x | 7 | 89 | Circumpolar |
| Røed et al., 2008 | Genetic analyses reveal independent domestication origins of Eurasian reindeer | 10.1098/rspb.2008.0332 | 4 | 125 | Palearctic |
| Eger et al., 2009 | Genetic diversity and history of Peary caribou (Rangifer tarandus) in North America | https://www.academia.edu/ | 9 | 90 | American Arctic |
| Kuhn et al., 2010 | Modern and ancient DNA reveal recent partial replacement of caribou in the southwest Yukon | 10.1111/j.1365-294X.2010.04565.x | 4 | 17 | American Arctic |
| Petersen et al., 2010 | Bottlenecks, isolation, and life at the northern range limit: Peary caribou on Ellesmere Island, Canada | 10.1644/09-MAMM-A-231.1 | 1 | 121 | CAA |
| Letts et al., 2012 | Ancient DNA Reveals Genetic Continuity in Mountain Woodland Caribou of the Mackenzie and Selwyn Mountains, Northwest Territories, Canada | 10.14430/arctic4186 | 5 | 38 | American Arctic |
| Weckworth et al., 2012 | Reconstruction of caribou evolutionary history in Western North America and its implications for conservation | 10.1111/j.1365-294X.2012.05621.x | 36 | 711 | American Arctic |
| Kvie et al., 2016 | Colonizing the High Arctic: Mitochondrial DNA Reveals Common Origin of Eurasian Archipelagic Reindeer (Rangifer tarandus) | 10.1371/journal.pone.0165237 | 6 | 92 | Barents region / Western Siberia |
| Røed et al., 2020 | Temporal and structural genetic variation in reindeer (Rangifer tarandus) associated with the pastoral transition in northwest Siberia | 10.1002/ece3.6314 | 1 | 36 | Western Siberia |
| Hold et al., 2024* | Ancient reindeer mitogenomes reveal island-hopping colonisation of the Arctic archipelagos | 10.1038/s41598-024-54296-2 | 26 regions, 38 locations. | 174 | Canada, Russia, Greenland, Svalbard, Franz Josef Land, Novaya Zemlya |
* Includes published sequences from: Burnett et al., 2023; Dussex et al., 2023; Kellner et al., 2024; Taylor et al., 2020.
Code/Software
Code for the analyses are provided in the form of R markdown scripts which are numbered in the order they should be run. These can be found on Zenodo at https://doi.org/10.5281/zenodo.12948971. All analyses were run using R v4.0.3 (R Development Core Team, 2020) in R studio v3.1 (Posit team, 2020). For details of packages, see individual scripts.
An exception is that code for the DIYABC-RF analyses is provided as bash shell scripts for linux command line, to be run after R markdown scripts 05/06. To run these analyses the command-line software DIYABC RF v1.1.27 and abcranger v1.16.30 for linux were used. Releases can be viewed here: