RES EARCH

DOG GENOMICS Port au Choix dog from Maritime Archaic cul-

tural context ~4000 YBP (3). Unsupervised clus-
Arctic-adapted dogs emerged at the tering analyses with NGSadmix software (4)
(fig. S6) grouped modern domestic dogs into
Pleistocene–Holocene transition four clusters: African, European, Asian, and
sled dogs including Zhokhov. These relation-
Mikkel-Holger S. Sinding1,2,3,4,5*†, Shyam Gopalakrishnan1*, Jazmín Ramos-Madrigal1*, ships were confirmed by an admixture graph
Marc de Manuel6*, Vladimir V. Pitulko7*, Lukas Kuderna6, Tatiana R. Feuerborn1,3,8,9, in which Yana was more closely related to a
Laurent A. F. Frantz10,11, Filipe G. Vieira1, Jonas Niemann1,12, Jose A. Samaniego Castruita1, Pleistocene wolf from Taimyr Peninsula than
Christian Carøe1, Emilie U. Andersen-Ranberg3,13, Peter D. Jordan14, Elena Y. Pavlova15, to modern wolves, whereas Zhokhov represents
Pavel A. Nikolskiy16, Aleksei K. Kasparov7, Varvara V. Ivanova17, Eske Willerslev1,18,19,20, a lineage that diverged from the ancestor of
Pontus Skoglund21,22, Merete Fredholm23, Sanne Eline Wennerberg24, Mads Peter Heide-Jørgensen4, present-day sled dogs (Fig. 1C and figs. S8 and
Rune Dietz25, Christian Sonne3,25,26, Morten Meldgaard1,3, Love Dalén8,27, Greger Larson10, S9). This suggests genetic continuity in Arctic
Bent Petersen1,28, Thomas Sicheritz-Pontén1,28, Lutz Bachmann2, Øystein Wiig2, dog breeds for at least the past ~9500 years,
Tomas Marques-Bonet6,29,30,31†‡, Anders J. Hansen1,3†‡, M. Thomas P. Gilbert1,32†‡ setting a lower bound on the origin of the sled
dog lineage.
Although sled dogs are one of the most specialized groups of dogs, their origin and evolution has Next, D statistics indicated an excess of allele
received much less attention than many other dog groups. We applied a genomic approach to investigate sharing between Yana-Taimyr wolves and PCDs-
their spatiotemporal emergence by sequencing the genomes of 10 modern Greenland sled dogs, an Zhokhov-sled dogs (Fig. 1D and fig. S14), corrob-
~9500-year-old Siberian dog associated with archaeological evidence for sled technology, and an orating previous reports (3, 5). This suggests
~33,000-year-old Siberian wolf. We found noteworthy genetic similarity between the ancient dog and that the admixture occurred between Pleistocene

Downloaded from http://science.sciencemag.org/ on June 26, 2020

modern sled dogs. We detected gene flow from Pleistocene Siberian wolves, but not modern American wolves and the ancestors of PCDs, Zhokhov, and
wolves, to present-day sled dogs. The results indicate that the major ancestry of modern sled dogs sled dogs.
traces back to Siberia, where sled dog–specific haplotypes of genes that potentially relate to Arctic Previous studies have demonstrated an asso-
adaptation were established by 9500 years ago. ciation between canine transmissible venereal
tumors (CTVTs) and sled dogs, especially PCDs

D
(3). Here, we evaluated the relationship among
espite decades of studies, consensus has (“Yana,” 4.7× coverage), dated to 33,019.5 YBP Zhokhov, two CTVT genomes (table S1), and
yet to be reached on when and where (Fig. 1A and fig. S3). In addition, we sequenced dogs and wolves using f3 statistics and phylo-
dogs were first domesticated and when 10 modern Greenland sled dog genomes, a genetic analysis. Recent analyses of exome data
they were first deliberately used in many dog best described as an indigenous land- suggested that CTVT expanded across Eurasia
of the roles they exhibit today. In Siberia, race breed used for hunting and sledging by ~6000 years ago (6), thus reducing the likeli-
late Upper Paleolithic artifacts of carved bone, Inuit. Samples consisted of two individuals hood that this transmissible cancer originated
antler, and ivory similar to tools used by modern from each of five geographically diverse lo- in the Americas. In our study, both the phylo-
Inuit for securing dog harness straps suggest calities (Fig. 1A), thus providing a broad rep- genetic analysis (fig. S9) and f3 statistics (fig. S10)
ancient origins of dog sledding (1). Furthermore, resentation of the indigenous dog diversity. placed the CTVT genomes closer to PCDs than
archeological findings from Zhokhov Island pro- We analyzed our data alongside genomes to sled dogs or Zhokhov. These results suggest
vide evidence of sled technology and dogs by from 114 geographically and genetically diverse that the basal dog lineage that led to PCDs (3)
the Sumnagin Mesolithic culture ~9000 to canids (table S1) using whole-genome pairwise occurred in Eurasia ~6000 years ago and/or
8000 years ago (1–3) (fig. S1), offering an op- distances, principal component analysis, TreeMix there were multiple introductions of PCD-like
portunity to use genomics to further our un- (4) admixture graphs, and D statistics (Fig. 1). dogs to the Americas.
derstanding of early dog domestication and Yana appeared alongside wolves (Fig. 1, B and C), We used NGSadmix, admixture analyses, and
the origin of sled dogs. whereas Zhokhov was found to be most closely D statistics (figs. S6 to S8 and S11 to S15) to
We generated nuclear genomes from a related to dogs. Specifically, Zhokhov was most evaluate gene flow and shared ancestry be-
dog mandible present at this site (“Zhokhov,” similar to modern sled dogs (Greenland sled tween Zhokhov and modern dogs and wolves.
9.6× coverage), dated to 9524 calendar years dogs, Alaskan malamutes, and Alaskan and We found no significant gene flow between
before present (YBP) (Fig. 1A and fig. S2), Siberian huskies) and American pre-European- any sled dog (including Zhokhov) and modern
and a Siberian Pleistocene wolf mandible contact dogs (PCDs), best illustrated by the ~2× American–Arctic wolf populations compared

1
The GLOBE Institute, University of Copenhagen, Copenhagen, Denmark. 2Natural History Museum, University of Oslo, Oslo, Norway. 3The Qimmeq Project, University of Greenland, Nuussuaq,
Greenland. 4Greenland Institute of Natural Resources, Nuuk, Greenland. 5Smurfit Institute of Genetics, Trinity College Dublin, Dublin, Ireland. 6Institute of Evolutionary Biology (UPF-CSIC),
Barcelona, Spain. 7Institute for the History of Material Culture, Russian Academy of Sciences, St. Petersburg, Russia. 8Department of Bioinformatics and Genetics, Swedish Museum of Natural
History, Stockholm, Sweden. 9Department of Archaeology and Classical Studies, Stockholm University, Stockholm, Sweden. 10The Palaeogenomics and Bio-Archaeology Research Network,
Research Laboratory for Archaeology and History of Art, University of Oxford, Oxford, UK. 11School of Biological and Chemical Sciences, Queen Mary University of London, London, UK. 12BioArch,
Department of Archaeology, University of York, York, UK. 13Department of Clinical Veterinary Sciences, University of Copenhagen, Frederiksberg C, Denmark. 14Arctic Centre and Groningen
Institute of Archaeology, University of Groningen, Netherlands. 15Arctic and Antarctic Research Institute, St. Petersburg, Russia. 16Geological Institute, Russian Academy of Sciences, Moscow,
Russia. 17VNIIOkeangeologia Research Institute (The All-Russian Research Institute of Geology and Mineral Resources of the World Ocean), St. Petersburg, Russia. 18Danish Institute for Advanced
Study (D-IAS), University of Southern Denmark, Odense, Denmark. 19Department of Zoology, University of Cambridge, Cambridge, UK. 20Wellcome Trust Sanger Institute, University of
Cambridge, Cambridge, UK. 21Department of Genetics, Harvard Medical School, Boston, MA, USA. 22Francis Crick Institute, London, UK. 23Department of Veterinary and Animal Sciences,
University of Copenhagen, Frederiksberg C, Denmark. 24Ministry of Fisheries, Hunting and Agriculture, Government of Greenland, Nuuk, Greenland. 25Department of Bioscience, Arctic Research
Centre, Aarhus University, Roskilde, Denmark. 26Henan Province Engineering Research Center for Biomass Value-added Products, School of Forestry, Henan Agricultural University, Zhengzhou,
Henan, China. 27Centre for Palaeogenetics, Stockholm, Sweden. 28Centre of Excellence for Omics-Driven Computational Biodiscovery (COMBio), Faculty of Applied Sciences, AIMST University,
Kedah, Malaysia. 29Catalan Institution of Research and Advanced Studies, Barcelona, Spain. 30CNAG-CRG, Centre for Genomic Regulation (CRG), Barcelona Institute of Science and Technology,
Barcelona, Spain. 31Institut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de Barcelona, Barcelona, Spain. 32University Museum, Norwegian University of Science and
Technology, Trondheim, Norway.
*These authors contributed equally to this work.
†Corresponding author. Email: [email protected] (M.-H.S.S.); [email protected] (T.M.-B.); [email protected] (A.J.H.); [email protected] (M.T.P.G.)
‡These authors cosupervised this work.

Sinding et al., Science 368, 1495–1499 (2020) 26 June 2020 1 of 5

RES EARCH | R E P O R T

with the Eurasian wolf (fig. S15), suggesting (Fig. 2A). Although pairs of Greenland sled dogs showed that these dogs had a relatively stable
that gene flow from modern wolves has not are symmetrically related to Zhokhov (D~0), population size until a severe bottleneck
contributed to the sled dog gene pool within indicating a lack of admixture, comparisons ~850 years ago. The timing of the bottleneck
the past 9500 years. This result was surprising involving non-Greenland sled dogs were not is consistent with the colonization of Green-
given genetic evidence for postdomestication always consistent with the null hypothesis land by Inuit (9), suggesting isolation in Green-
admixture between other wolves and dog breeds of no admixture. D-statistics and admixture land ever since.
(5, 7). Furthermore, ethnographic evidence from analyses (Fig. 2B and fig. S13) indicated that Numerous generations of sled dogs living in
Greenland indicates that, at least historically, non-Greenland sled dogs carry ancestry from the Arctic environment and being used as
dog-wolf matings were not uncommon (8). If non-sled dogs and that Greenland sled dogs draft animals may have provided a unique
true, then the lack of gene flow from modern are the least admixed. These results imply that selection pressure to these dogs. To detect pu-
American-Arctic wolves into sled dogs implies Greenland sled dogs have largely been kept tative signals of positive selection, we used
selection against hybrids. isolated from contact with other dog breeds, population branch statistics (PBS) (10) to scan
The clustering and admixture results show and that their lineage traces more genomics for genomic regions highly differentiated in
gene flow between some sled dogs and other ancestry to Zhokhov-like dogs relative to other modern sled dogs relative to non-sled dogs
modern dog breeds (Fig. 1C and figs. S6 to S8). dog breeds. Isolation of Greenland sled dogs (hereafter referred to as “other dogs”) and
We further explored this by comparing pairs was supported by inference of their histori- wolves. We computed these statistics on mod-
of sled dogs with Zhokhov using D statistics cal effective population size (fig. S16), which ern genomes of 17 sled dogs, 61 other dogs,

Ilulissat 1,2
A Zhokhov site
Siberian Husky 1,2,3 Qaanaaq 1,2 Greenland sled dog
Swedish Lapphund
Alaskan Husky 1,2 Unknown 1

Downloaded from http://science.sciencemag.org/ on June 26, 2020

Finnish Lapphund
Samoyed Yana site Aasiaat 1,2
Jämthund
Gray Norwegian Elkhound Lapponian Herder East Siberian Laika
Uyak Sisimiut 1,2 Tasiilaq 1,2
Cherry Tree Cave
Newgrange Herxheim Alaskan Malamute
Gansu 1,2 Shaanxi 1,2 Port au Choix
Belgium Malinois Boxer Xinjiang
Galgo Español German Shepherd Hebei
Chinese indigenous dog 1,3 Dalian Weyanoke Old Town 1,2
Lebanon 1,2,3 Shanxi 1,2
Sloughi Afghan Anhui
Egypt 1,2 Tibetan Mastiff Chihuahua
Guizhou
India 1-6 Guangdong Mexican naked
Qatar 1,2 Yunnan 1,2 China/Vietnam border 1,2
0.046 Ondo Chinese indigenous dog 2
Ibadan Jalingo City
0.044
0.043 Basenji Borneo 1,2,3
Uyo
0.041
0.039 Papua New Guinea 1,2,3
Peruvian naked
0.037
0.036
0.034 Dingo
0.032
0.03
pairwise−distance
Ancient dog genome
Present day dog genome

B C D
Grey wolf Eurasia Other dog
0.06

Greenland sled dog Port au Choix

D(H1, Boxer dog; Yana, Andean fox)

Other sled dog Other sled dog

Asia
Other dog Greenland sled dog
Zhokhov
0.2

Ancient dog Port au Choix Ancient dog

0.04

Grey wolf
Pleistocene wolf Alaskan Malamute
Migration
PCA2 (4.08%)

weight
Zhokhov
0.5 Greenland Sled Dog
0.1

Zhokhov dog
0.02

Weyanoke Old Town 1 Huskies

Port au Choix Africa
Yana 0 Europe
0.00

Taimyr wolf 10 s.e.

0.0

Pleistocene wolf
Herxheim
Grey wolf America (Arctic)
−0.02

Cherry Tree Cave Grey wolf America (Alaska)

Grey wolf America (Mexico)
−0.1

Newgrange
Coyote

−0.10 −0.05 0.00 0.05 0.10 0.15 0.00 0.05 0.10 0.15 0.20 −0.06 −0.04 −0.02 0.00 0.02 0.04 0.06

PCA1 (7.91%) Drift parameter D(H1, Boxer dog; Taimyr, Andean fox)

Fig. 1. Geographic location of the samples and overall genetic affinities. profile (table S1 and fig. S6). Colors indicate main groups as in (B). Arrows
(A) Identity by state pairwise distances between Zhokhov and present-day show inferred admixture edges colored by migration weight. (D) D statistic
dogs (table S1) of geographic affiliation of dogs and archaeological sites. of the form D(H1, boxer dog; Taimyr or Yana, Andean fox) testing for
Color scale indicates genetic distance between Zhokhov and each sample. Pleistocene wolf gene flow in ancient and modern dogs and whether samples
Circles and triangles represent modern and ancient dogs, respectively. share more alleles with Taimyr (x-axis) or Yana (y-axis) wolves when compared
Stars show Zhokhov and Yana sites. (B) Principal component analysis with the boxer dog. Color indicates the type of sample in H1. Points show
(PCA) using whole-genome data (2,200,623 transversion sites) on all the D statistic, and horizontal and vertical lines show 3 SEs for the test with the
samples. (C) TreeMix admixture graph built using whole-genome data Taimyr (x-axis) and Yana (y-axis), respectively. The results obtained from
(766,082 transversion sites) on a dataset consisting of 66 canids merged both ancient wolves fall along the diagonal, suggesting that they are
into 15 groups according to their geographic location and admixture symmetrically related to all dogs.

Sinding et al., Science 368, 1495–1499 (2020) 26 June 2020 2 of 5

RES EARCH | R E P O R T

and 30 wolves (table S1). A sliding window TRPC4 is highly differentiated in sled dogs, functions beneficial to physical activity in the
analysis revealed several genomic regions and the putatively selected haplotype bears a Arctic. If so, given that the differentiated hap-
with high PBS values, hinting at selection marked similarity to Zhokhov (Fig. 3, A and B). lotypes are also found in Zhokhov (Fig. 3, A
in sled dogs (Fig. 3A). We took an outlier TRPC4 is a transient receptor potential (TRP) and B, and fig. S19A), any advantages that they
approach and focused on the most extreme channel protein that plays an important role confer would have been important to dogs in
values of the empirical distribution (above in vasorelaxation and lung microvascular per- the Arctic ~9500 YBP.
the 99.95th percentile). For each of these meability (11). It is also involved in a temper- Most domestic dogs are adapted to starch-
outlier regions (table S4), we identified over- ature sensitivity pathway (12, 13), where it rich diets through marked increases in
lapping genes and compared haplotypes interacts with TRPV2, which is also highly AMY2B copy numbers and strong positive
across samples. differentiated in sled dogs (99.8th PBS per- selection for a dog-specific MGAM haplo-
Enrichment analysis (4) on genomic regions centile; table S4 and fig. S19A) and codes for type (19). Consistent with previous findings
with high PBS values (above the 99.95th per- temperature and potentially pain receptors (20), we observed that sled dogs carry sub-
centile) identified three gene ontology (GO) (14). Several related thermo-TRP sensors in stantially fewer AMY2B copies than other
terms that were overrepresented (table S6): the same pathway, calcium ion transmembrane dog breeds (fig. S20). We also found that
g-aminobutyric acid secretion (GO: 0014051, transport, have been previously reported to be MGAM and AMY2B are the regions of the
p = 0.119), calcium ion import (GO: 0070509, under selection in cold-adapted woolly mam- genome with the lowest PBS, suggesting high
p = 0.119), and calcium ion transmembrane moths (15), which suggests convergent evolu- differentiation of other dogs relative to sled
transport (GO: 0070588, p = 0.382). To inves- tion in Arctic adaptation. dogs and wolves (Fig. 3A). Because negative
tigate further, we focused on eight genomic Another highly differentiated gene in sled PBS can arise under different demographic
regions that are highly differentiated in sled dogs is CACNA1A (Fig. 3, A and C), a calcium scenarios, we confirmed these observations
dogs and three regions where other dogs dif- channel subunit that plays an essential role by computing PBS with other dogs as the

Downloaded from http://science.sciencemag.org/ on June 26, 2020

fer from sled dogs and wolves (Fig. 3A and fig. in skeletal muscle contraction (16). Further, focal population (fig. S18). Indeed, modern
S18), and validated the autosomal regions with CACNA1A has been reported to be under posi- sled dogs and Zhokhov are among the only
a cross-population composite likelihood ratio tive selection in humans, specifically the Bajau dogs in our dataset that carry the ancestral
statistic (5) (fig. S21). In the differentiated re- sea nomads (17), where it is involved in hy- MGAM haplotype found at high frequency in
gions, we focused on two sets of genes: those poxia adaptation (18), indicating a possible wolves (Fig. 3C and fig. S18). Therefore, our
in which Zhokhov carries the same haplotype role in managing exercise-induced hypoxia observations suggest that sled dogs do not
as modern sled dogs and those involved in in sled dogs. We hypothesize that the TRPC4, carry the genetic adaptations to starch-rich
adaptation to different diets. TRPV2, and CACNA1A genes are involved in D(Greenland Village Dog Aasiaat 2, H2; Zhokhov, Andean fox) diets seen in other dog breeds.

A Greenland sled dog Qaanaaq 1 H0 B

Greenland sled dog Qaanaaq 2
Greenland sled dog Aasiaat 1
0.10
Andean fox
Dog 2
Dog 1

Zhokhov

Greenland sled dog Aasiaat 2

Greenland sled dog Ilulissat 1
Greenland sled dog Ilulissat 2
Dog1 Zhokhov
Deviation from H0

Greenland sled dog Sisimiut 1 −5.1

0.0 5

Greenland sled dog Sisimiut 2

Greenland sled dog Tasiilaq 1
Greenland sled dog Tasiilaq 2
Greenland sled dog Unknown 1
17.9
Alaskan Malamute 1 Dog2 Zhokhov
0.0 0

Dog 1 Alaskan Husky 1

Alaskan Husky 2
Siberian Husky 1
Siberian Husky 2 Other sled dogs
Siberian Husky 3 Greenland sled dogs

−0.06 −0.04 −0.0 2 0.0 0 0.02

G nla d sl do aan q 1

re an le g si 2
re n le g si 1
G la sle do luli at 2
G la sle o uli at 1
re la sl o isi t 2
nd sle dog isim t 1
ed o s t 2
as g U si 1

Al M now 2
Al kan mu 1

Si ria us 1
Si ria us 2
ria u 1
H y2

3
Si ka us 1
G enl d s do Aa aq
G nla d s do Aa at

Al do Ta ilaq

n k q
as ala n

be n H ky
be n H ky
be n H ky

ky
as H te
G een nd d d S ssa
S iu
sl d d Ta iu

n sk
a

ka n ila
a
a
re n d g ss

us
en n ed g m

D(Greenland Village Dog Aasiaat 2, H2; German Shepherd, Andean fox)

re n ed Q na
re n ed g a

g i
en d do Qaa

Eurasian dog gene flow

l
I
I
g
G nla sled og
G la sl g

d
d

d
en d d
G lan sled

d
en d
re n
en d
re d
re lan

la
e

e
en

Dog 2
r
re
G
G

Fig. 2. Relationships between Zhokhov and present-day sled dogs. (A) D Greenland sled dog Aasiaat 2 [y-axis: D(Greenland sled dog Aasiaat 2, H2;
statistics testing the relationships between pairs of sled dogs and Zhokhov. Zhokhov, Andean fox)] also show evidence of significant gene flow from other
Cell colors indicate the Z scores obtained from the test D(dog1, dog2; dogs [x-axis: D(Greenland sled dog Aasiaat 2, H2; German shepherd dog,
Zhokhov, Andean fox), where dog1 and dog2 are all possible pairs of sled Andean fox)]. Points indicate the D statistic, and horizontal and vertical
dogs. Comparisons involving pairs of Greenland sled dogs and non-Greenland lines indicate 3 SEs for the x-axis and y-axis, respectively. We consider the
sled dogs resulted in significant deviations from H0 (|Z| > 3). (B) D statistics test to be significant for gene flow when these lines do not overlap with
showing that sled dogs that are significantly further from Zhokhov compared with the dotted line (|Z| > 3).

Sinding et al., Science 368, 1495–1499 (2020) 26 June 2020 3 of 5

RES EARCH | R E P O R T

Fig. 3. Adaptation. A ADAMTS2 CACNA1A

(A) Manhattan plot of APOO
SLC25A40 TRPC4 PTPN20
NUP214 EFCAB8
the PBS values (y-axis)
in windows of 100 kg
base pairs (kb) using a
20-kb slide across
chromosomes (x-axis).
Data points between
the 20th and 80th
percentile of the empir-
ical distribution are
not plotted and dashed
red lines show the
99.95th and 0.05th per- AMY2B
MGAM
centiles. Names of RPS6KA3
genes within the highest
peaks are shown, with
B C D
asterisks representing
no overlap with genes.
We note that other Sled Greenland SDs (11)
dogs Other SDs (6)

Downloaded from http://science.sciencemag.org/ on June 26, 2020

genes not displayed in Zhokhov (1)
Newgrange (1)
the figure can overlap Herxheim (1)
Cherry Tree Cave (1)
the outlier regions; a full
list can be found in
tables S4 and 5.
(B to D) Haplotype Other dogs (61)
structures for TRPC4
(B), CACNA1A (C), and
MGAM (D). Rows Yana (1)
represent individuals,
and columns represent Wolves (30)
polymorphic positions in
the dog genome. Cells
are colored by genotype:
Dark gray indicates that the alternative allele is homozygous, light gray that it is heterozygous, and white that the reference allele is homozygous. The row height for
ancient individuals was increased to facilitate visualization. Zhokhov is highlighted with a red asterisk. SDs, sled dogs.

By contrast, sled dogs harbor specific hap- Bone composition of polar bears and rein- 3. M. Ní Leathlobhair et al., Science 361, 81–85 (2018).
lotypes of genes involved in coping with a high deer consumed at the Zhokhov site indicate an 4. See the supplementary materials.
5. P. Skoglund, E. Ersmark, E. Palkopoulou, L. Dalén, Curr. Biol.
intake of fatty acids. SLC25A40, a mitochondrial extensive hunting range and transport of large 25, 1515–1519 (2015).
carrier protein involved in clearing triglycerides body parts back to camp (26). Further, abun- 6. A. Baez-Ortega et al., Science 365, eaau9923 (2019).
from the blood (21), and APOO, an apolipo- dant obsidian tools found at the site reveal 7. Z. Fan et al., Genome Res. 26, 163–173 (2016).
protein gene involved in regulating high levels movement of obsidian from ~1500 km away 8. B. Muus, F. Salomonsen, C. Vibe, Grønlands Fauna [in Danish]
(Gyldendal, 1981).
of fat and fatty acid metabolism (22), are both (3). Together, these findings indicate substan- 9. M. Raghavan et al., Science 345, 1255832 (2014).
highly differentiated in sled dogs (Figs. 3A). tial long-distance travel and transportation 10. X. Yi et al., Science 329, 75–78 (2010).
The derived haplotypes of both genes are ab- of resources, in which dog sledding would 11. C. Tiruppathi et al., Circ. Res. 91, 70–76 (2002).
sent in Zhokhov, indicating that the haplo- have been highly advantageous—if not neces- 12. T. Hofmann, M. Schaefer, G. Schultz, T. Gudermann, Proc. Natl.
Acad. Sci. U.S.A. 99, 7461–7466 (2002).
types are specific to modern sled dogs and sary. Putative sled remains and our genomic 13. K. Zimmermann et al., Proc. Natl. Acad. Sci. U.S.A. 108,
postdate their common ancestors with Zhokhov analyses of a 9500-year-old dog from the 18114–18119 (2011).
(fig. S19, B and E). As another example of con- Zhokhov site indicate that the traditions and 14. N. Qin et al., J. Neurosci. 28, 6231–6238 (2008).
vergent evolution, another gene of the apolipo- key genomic variations that define modern 15. V. J. Lynch et al., Cell Rep. 12, 217–228 (2015).
16. S. Kaja et al., Eur. J. Neurosci. 25, 2009–2020 (2007).
protein family, APOB, is reported to be under sled dogs were established in the northeast 17. M. A. Ilardo et al., Cell 173, 569–580.e15 (2018).
selection in polar bears, possibly as a result of Asian Arctic >9500 years ago. Our results 18. V. Wang, D. A. Davis, M. Haque, L. E. Huang, R. Yarchoan,
adaptation to fat-rich diets and clearance imply that the combination of these dogs with Cancer Res. 65, 3299–3306 (2005).
19. E. Axelsson et al., Nature 495, 360–364 (2013).
of cholesterol from the blood (23). Overall, the innovation of sled technology facilitated 20. M. Arendt, K. M. Cairns, J. W. O. Ballard, P. Savolainen,
similar adaptations to high intake of fatty human subsistence since the earliest Holocene E. Axelsson, Heredity 117, 301–306 (2016).
acids have been described in Inuit and other in the Arctic. 21. E. A. Rosenthal et al., Am. J. Hum. Genet. 93, 1035–1045
(2013).
Arctic human populations (24, 25), so our ob- 22. A. Turkieh et al., J. Clin. Invest. 124, 2277–2286 (2014).
servations suggest that sled dogs adapted to a RE FERENCES AND NOTES
23. S. Liu et al., Cell 157, 785–794 (2014).
fat-rich and starch-poor diet, echoing the dietary 1. V. V. Pitulko, A. K. Kasparov, Arctic Anthropol. 33, 1–36 24. A. Cardona et al., PLOS ONE 9, e98076 (2014).
(1996). 25. M. Fumagalli et al., Science 349, 1343–1347 (2015).
adaptations of the Arctic human cultures with 2. V. V. Pitulko, A. K. Kasparov, J. Archaeol. Sci. Rep. 13, 491–515 26. V. V. Pitulko, V. V. Ivanova, A. K. Kasparov, E. Y. Pavlova,
whom they coexisted. (2017). Environ. Archaeol. 20, 120–157 (2015).

Sinding et al., Science 368, 1495–1499 (2020) 26 June 2020 4 of 5

RES EARCH | R E P O R T

ACKN OW LEDG MEN TS 10265-RNF. T.M.B. was supported by BFU2017-86471-P T.M.-B., A.J.H., and M.T.P.G. interpreted results with considerable
We thank J. A. Leonard and B. von Holdt for input and (MINECO/FEDER, UE), Howard Hughes International Early input from B.P., T.S.-P., V.V.P., T.R.F., E.U.A.-R., P.D.J., M.M., L.D.,
comments in the conceptualization of this study, the Danish Career, Obra Social “La Caixa” and Secretaria d’Universitats i G.L., L.B., and Ø.W. M.-H.S.S., S.G., J.R.-M., M.d.M.M., and
National High-Throughput Sequencing Centre and BGI-Europe Recerca and CERCA Programme del Departament d’Economia i M.T.P.G. wrote the paper with input from all other authors.
for assistance in Illumina data generation, and the Danish Coneixement de la Generalitat de Catalunya (GRC 2017 SGR Competing interests: The authors declare no competing interests.
National Supercomputer for Life Sciences – Computerome 880). M.T.P.G. was supported by a European Research Council Data and materials availability: Raw sequencing data can
(https://computerome.dtu.dk) for the computational resources grant (ERC-2015-CoG-681396–Extinction Genomics). G.L. and be accessed at the NCBI Short Read Archive under project
to perform the sequence analyses. Funding: This work is L.A.F. were supported by the ERC (Grant ERC-2013-StG-337574- number PRJNA608847.
embedded in “The Qimmeq Project,” funded by the Velux UNDEAD) and the Natural Environmental Research Council
Foundations and Aage og Johanne Louis-Hansens Fond, and (Grants NE/K005243/1 and NE/K003259/1). P.S. was
supported by ArchSci2020, funded by the European Union’s EU supported by the Francis Crick Institute (FC001595). SUPPLEMENTARY MATERIALS
Framework Programme for Research and Innovation Horizon Author contributions: M.-H.S.S., S.G., J.R.-M., M.d.M.M.,
science.sciencemag.org/content/368/6498/1495/suppl/DC1
2020 under Marie Curie Actions grant no. 676154. We thank the and M.T.P.G. conceived of the project and designed the
Materials and Methods
Rock Foundation of New York for funding excavations at the research. V.V.P., E.Y.P., P.A.N., A.K.K., V.V.I., and E.W. provided
Figs. S1 to S21
Zhokhov and Yana sites in a 15-year-long effort starting in 2000. archaeological work, logistics, and/or ancient collected
Tables S1 to S6
M.-H.S.S. was supported by the Independent Research Fund samples. M.-H.S.S., M.F., S.E.W., M.P.H.-J., R.D., and C.S.
References (27–70)
Denmark (8028-00005B) and NHM Oslo. S.G. was supported by coordinated logistics of and/or provided modern samples.
MDAR Reproducibility Checklist
Marie Skłodowska-Curie Actions (H2020 655732 - WhereWolf) C.C. and M.-H.S.S. conducted the laboratory work. S.G.,
and Carlsberg (CF14 - 0995). M.d.M.M. was supported by a J.R.-M., M.d.M.M., L.K., L.A.F.F., F.G.V., J.N., and J.A.S.C. View/request a protocol for this paper from Bio-protocol.
Formació de Personal Investigador fellowship from Generalitat conducted the analyses of data with considerable input from
de Catalunya (FI_B01111). V.V.P., E.Y.P., and P.A.N. were M.-H.S.S., B.P., T.S.-P., T.M.-B., A.J.H., and M.T.P.G. S.G., J.R.-M., 15 October 2019; accepted 6 May 2020
supported by the Russian Science Foundation project no. 16-18- M.d.M.M., L.K., L.A.F.F., F.G.V., J.N., J.A.S.C., P.S., M.-H.S.S., 10.1126/science.aaz8599

Downloaded from http://science.sciencemag.org/ on June 26, 2020

Sinding et al., Science 368, 1495–1499 (2020) 26 June 2020 5 of 5

Arctic-adapted dogs emerged at the Pleistocene−Holocene transition
Mikkel-Holger S. Sinding, Shyam Gopalakrishnan, Jazmín Ramos-Madrigal, Marc de Manuel, Vladimir V. Pitulko, Lukas
Kuderna, Tatiana R. Feuerborn, Laurent A. F. Frantz, Filipe G. Vieira, Jonas Niemann, Jose A. Samaniego Castruita, Christian
Carøe, Emilie U. Andersen-Ranberg, Peter D. Jordan, Elena Y. Pavlova, Pavel A. Nikolskiy, Aleksei K. Kasparov, Varvara V.
Ivanova, Eske Willerslev, Pontus Skoglund, Merete Fredholm, Sanne Eline Wennerberg, Mads Peter Heide-Jørgensen, Rune
Dietz, Christian Sonne, Morten Meldgaard, Love Dalén, Greger Larson, Bent Petersen, Thomas Sicheritz-Pontén, Lutz
Bachmann, Øystein Wiig, Tomas Marques-Bonet, Anders J. Hansen and M. Thomas P. Gilbert

Science 368 (6498), 1495-1499.

DOI: 10.1126/science.aaz8599

Sled dog arctic adaptations go far back

Downloaded from http://science.sciencemag.org/ on June 26, 2020

Dogs have been used for sledding in the Arctic as far back as ∼9500 years ago. However, the relationships
among the earliest sled dogs, other dog populations, and wolves are unknown. Sinding et al. sequenced an ancient sled
dog, 10 modern sled dogs, and an ancient wolf and analyzed their genetic relationships with other modern dogs. This
analysis indicates that sled dogs represent an ancient lineage going back at least 9500 years and that wolves bred with
the ancestors of sled dogs and precontact American dogs. However, gene flow between sled dogs and wolves likely
stopped before ∼9500 years ago.
Science, this issue p. 1495

ARTICLE TOOLS http://science.sciencemag.org/content/368/6498/1495

SUPPLEMENTARY http://science.sciencemag.org/content/suppl/2020/06/24/368.6498.1495.DC1
MATERIALS

RELATED file:/content
CONTENT

REFERENCES This article cites 69 articles, 25 of which you can access for free
http://science.sciencemag.org/content/368/6498/1495#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. The title Science is a registered trademark of AAAS.
Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of
Science. No claim to original U.S. Government Works