University of Groningen

The paleogenetics of cat dispersal in the ancient world

Ottoni, Claudio; Van Neer, Wim; De Cupere, Bea; Daligault, Julien; Guimaraes, Silvia;
Peters, Joris; Spassov, Nikolai; Prendergast, Mary E.; Boivin, Nicole; Morales-Muniz, Arturo
Published in:
Nature Ecology & Evolution

DOI:
10.1038/s41559-017-0139

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.

Document Version
Publisher's PDF, also known as Version of record

Publication date:
2017

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Ottoni, C., Van Neer, W., De Cupere, B., Daligault, J., Guimaraes, S., Peters, J., Spassov, N., Prendergast,
M. E., Boivin, N., Morales-Muniz, A., Balasescu, A., Becker, C., Benecke, N., Boroneant, A., Buitenhuis, H.,
Chahoud, J., Crowther, A., Llorente, L., Manaseryan, N., ... Geigl, E.-M. (2017). The paleogenetics of cat
dispersal in the ancient world. Nature Ecology & Evolution, 1(7), Article 0139.
https://doi.org/10.1038/s41559-017-0139

Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.

Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.

Download date: 16-10-2024

 ARTICLES
PUBLISHED: 19 JUNE 2017 | VOLUME: 1 | ARTICLE NUMBER: 0139

The palaeogenetics of cat dispersal in the

ancient world
Claudio Ottoni1,2†*, Wim Van Neer3,4, Bea De Cupere3, Julien Daligault2, Silvia Guimaraes2,
Joris Peters5,6, Nikolai Spassov7, Mary E. Prendergast8, Nicole Boivin9, Arturo Morales-Muñiz10,
Adrian Bălăşescu11, Cornelia Becker12, Norbert Benecke13, Adina Boroneant14, Hijlke Buitenhuis15,
Jwana Chahoud16,17, Alison Crowther18, Laura Llorente10†, Nina Manaseryan19, Hervé Monchot20,
Vedat Onar21, Marta Osypińska22, Olivier Putelat23, Eréndira M. Quintana Morales24,
Jacqueline Studer25, Ursula Wierer26, Ronny Decorte1, Thierry Grange2‡* and Eva-Maria Geigl2‡*

The cat has long been important to human societies as a pest-control agent, object of symbolic value and companion animal, but
little is known about its domestication process and early anthropogenic dispersal. Here we show, using ancient DNA analysis
of geographically and temporally widespread archaeological cat remains, that both the Near Eastern and Egyptian populations
of Felis silvestris lybica contributed to the gene pool of the domestic cat at different historical times. While the cat’s worldwide
conquest began during the Neolithic period in the Near East, its dispersal gained momentum during the Classical period, when
the Egyptian cat successfully spread throughout the Old World. The expansion patterns and ranges suggest dispersal along
human maritime and terrestrial routes of trade and connectivity. A coat-colour variant was found at high frequency only after
the Middle Ages, suggesting that directed breeding of cats occurred later than with most other domesticated animals.

T
he domestic cat is present on all continents except Antarctica, subspecies: Felis silvestris silvestris, Felis silvestris lybica, Felis
and in the most remote regions of the world, and its evolu- silvestris ornata, Felis silvestris cafra and Felis silvestris bieti4. Modern
tionary success is unquestioned. While it is nowadays one of genetic data analyses of nuclear short tandem repeats (STR) and
the most cherished companion animals in the Western world, for 16% of the mitochondrial DNA (mtDNA) genome in extant wild
ancient societies barn cats, village cats and ships’ cats provided criti- and domestic cats revealed that only one of them, the north African/
cal protection against vermin, especially rodent pests responsible southwest Asian F. s. lybica, was ultimately domesticated5.
for economic loss and disease1. Owing to a paucity of cat remains Wildcats are solitary, territorial hunters and lack a hierarchi-
in the archaeological record, current hypotheses about early cat cal social structure6,7, features that make them poor candidates
domestication rely on only a few zooarchaeological case studies. for domestication8. Indeed, zooarchaeological evidence points to
These studies suggest that ancient societies in both the Near East a commensal relationship between cats and humans lasting thou-
and Egypt could have played key roles in cat domestication2,3. sands of years before humans exerted substantial influence on their
Wildcats (Felis silvestris) are distributed all over the Old World. breeding2,3,9. Throughout this period of commensal interaction,
Current taxonomy distinguishes five wild, geographically partitioned tamed and domestic cats became feral and/or intermixed with wild

1
KU Leuven—University of Leuven, Department of Imaging and Pathology, Center for Archaeological Sciences; University Hospitals Leuven, Laboratory
of Forensic Genetics and Molecular Archaeology, B-3000 Leuven, Belgium. 2Institut Jacques Monod, UMR 7592, CNRS and University Paris Diderot,
F-75013 Paris, France. 3Royal Belgian Institute of Natural Sciences, B-1000 Brussels, Belgium. 4KU Leuven—University of Leuven, Department of Biology,
Laboratory of Biodiversity and Evolutionary Genomics, Center of Archaeological Sciences, B-3000 Leuven, Belgium. 5Institute of Palaeoanatomy,
Domestication Research and the History of Veterinary Medicine, Ludwig-Maximilian University, D-80539 Munich, Germany. 6Bavarian Natural History
Collections, Bavarian State Collection of Anthropology and Palaeoanatomy, D-80333 Munich, Germany. 7National Museum of Natural History at the
Bulgarian Academy of Sciences, BG-1000 Sofia, Bulgaria. 8Radcliffe Institute for Advanced Study, Harvard University, Cambridge, 02138 Massachusetts,
USA. 9Max Planck Institute for the Science of Human History, Jena D-07743, Germany. 10Laboratorio de Arquezoología, Universidad Autónoma de
Madrid, E-28049 Madrid, Spain. 11National History Museum of Romania, RO-030026 Bucharest, Romania. 12Institute of Prehistoric Archaeology,
Free University Berlin, D-14195 Berlin, Germany. 13German Archaeological Institute, D-14195 Berlin, Germany. 14‘Vasile Pârvan’ Institute of Archaeology
of the Romanian Academy, RO-010667 Bucharest, Romania. 15Groningen Institute of Archaeology, University of Groningen, NL-9712 ER Groningen,
the Netherlands. 16Archéorient; CNRS/UMR 5133, Université Lumière Lyon II, F-69007 Lyon, France. 17Natural History Museum, Lebanese University,
LB-1107-2020 Beirut, Lebanon. 18School of Social Science, The University of Queensland, AU-4072 Brisbane, Queensland 4072, Australia. 19Institute of
Zoology, National Academy of Sciences of Armenia, AM-0019 Yerevan, Armenia. 20Labex Resmed, Université Paris IV la Sorbonne, F-75005 Paris, France.
21
Istanbul University, Osteoarchaeology Practice and Research Center, Faculty of Veterinary Medicine, TR-34320 Avcilar-Istanbul, Turkey. 22Institute of
Archaeology and Ethnology, Polish Academy of Sciences, PL-61-712 Poznań, Poland. 23Archéologie Alsace, F-67600 Sélestat, and UMR 7041, ArScan,
Nanterre, France. 24Department of Anthropology, Rice University, Houston, Texas 77005, USA. 25Natural History Museum of Geneva, CH-1208 Genève,
Switzerland. 26Soprintendenza Archeologia della Toscana, I-50121 Firenze, Italy. †Present addresses: Centre for Ecological and Evolutionary Synthesis
(CEES), Department of Biosciences, University of Oslo, NO-0316 Oslo, Norway (C.O.); BioArch, Department of Archaeology, University of York, York YO10
5NG, UK (L.L). ‡These authors contributed equally to this work. *e-mail: claudio.ottoni@ibv.uio.no; eva-maria.geigl@ijm.fr; thierry.grange@ijm.fr

NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol 1

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
ARTICLES NATURE ECOLOGY & EVOLUTION

F. s. lybica or other wild subspecies as is common today10. These Anatolian cats from the Neolithic period to the Bronze age.
regular genetic exchanges may have contributed to the low level of A mitotype belonging to subclade IV-A (hereafter IV-A*, see
differentiation observed between modern wild and domestic cat Methods section) was predominant (12 out of 14) from around
genome sequences11. Accordingly, the domestication process seem- 8000 to 800 BC in Anatolia (10-13) (Fig. 1a–c). Its range may have
ingly has not profoundly altered the morphological, physiological, also extended to Lebanon (15). The frequencies of IV-A1 and IV-A*
behavioural and ecological features of cats9, in contrast to what has found in southeast Europe and Anatolia, respectively, are signifi-
been observed, for example, for dogs12. cantly different (Fisher’s exact test; P < 0.001), suggesting a phy-
To address questions related to the contribution of the two pur- logeographic structure that mirrors the original distribution of
ported centres of cat domestication, the Near East and Egypt, and the genetically distinct wildcat populations carrying F. s. lybica mtDNA.
history of human-mediated cat dispersal, we analysed ancient and The earliest occurrence of IV-A* outside the Anatolian range in our
modern cats from Europe, north and east Africa, and southwest Asia dataset was detected in two directly radiocarbon-dated specimens
(SWA), spanning around 9,000 years, from the Mesolithic period from southeast Europe, in Bulgaria (4400 BC) and Romania (3200 BC),
to the twentieth century AD. We analysed ancient DNA (aDNA) to clearly postdating the introduction of Neolithic farming prac-
explore whether a fine phylogeographic structure of maternal lin- tices, and in two Late Bronze/Iron age cats (around 1200 BC) from
eages existed prior to the domestication of F. s. lybica and whether, Greece. The range expansion of this mitotype suggests human-
when and how it was reconfigured over time in response to human mediated translocation.
intervention, thereby documenting the domestication process of the
cat. We also studied a genetically defined coat-colour marker, the Ancient Levant and Africa. Owing to very poor DNA preservation,
blotched tabby marking13, to monitor a phenotypic change reflect- we could not explore the phylogeographic structure of F. s. libyca in
ing human-driven selection along the domestication pathway. this area prior to the Bronze Age. Therefore, we inferred the original
distribution of the other subclades (IV-B/E) by taking into account
Results their temporal appearance in our dataset. We found IV-B in three
Strategy for data acquisition. The mtDNA phylogeny recon- ancient cat remains dated to the 1st millennium BC from southeast
structed from extant wild and domestic cats5 identified five geo- Anatolia and Jordan (13, 16, Fig. 1a–c), the 6th century BC in Syria and
graphically distinct clades (I–V, Supplementary Fig. 1), representing later in Jordan (15, 16). This clade is still found in modern wildcats from
the five F. silvestris subspecies. The modern domestic cat mtDNA Israel5,15. These data suggest that this subclade was mainly restricted to
pool was traced back to five deeply divergent subclades (IV-A to a Levantine range, throughout history. Outside of this range, IV-B was
IV-E) of the F. s. lybica clade, representing multiple wildcat lineages found only in Medieval Iran (17) at very low frequencies (7%).
incorporated over time and space6. These subclades lack a phylo- In Africa, two lineages of IV-C (named IV-C1 and IV-C*) were
geographic structure, which may reflect either poor sampling of the detected in five out of seven cats (including three mummies) from
truly wild modern F. s. lybica, particularly in its African range, or Egypt with dates ranging from the 7th century BC to the 4th century
multiple domestication events and/or extensive gene flow between AD (20, 21, Fig. 1a–c). The original range of IV-C may have extended
wild and domestic populations following the dispersal of domestic from Egypt along the Nile River as far south as Congo and Burundi
cats. In order to screen and analyse a large number of ancient sam- (27, 28), where we detected a novel sub-lineage of IV-C (IV-C2,
ples in parallel, many of which were expected to be poorly preserved Fig. 1) in modern wildcats that had not yet been described in the
owing to higher-temperature burial environments, we applied an mtDNA pool of present-day domestic cats.
ultrasensitive high-throughput approach14 to target informative Subclades IV-D and IV-E were found at low frequencies solely
single-nucleotide polymorphisms (SNPs) on the mtDNA that reca- in recent temporal bins of our ancient dataset (1, 9–11), most likely
pitulate the most salient features of the previously obtained phy- as a result of human-mediated dispersal. Their basal position in
logeny (Supplementary Fig. 1). Although mtDNA alone cannot clade IV, shared with lineages found in ancient African cats (light
assess possible hybridization between different populations at the pink symbols in Fig. 1b, c; 20, 25, 28, 29) and not detected so far in
individual level, the absence of recombination and the high copy domestic cats, may suggest an African origin.
number make it a useful genetic marker for ancient population
analyses involving a large number of poorly preserved samples. The The dispersal of Egyptian cats. Outside Africa, from the 8th cen-
mtDNA phylogeny (Fig. 1b) reconstructed from 286 bp sequenced tury BC to the 5th century AD, we found IV-C1 in five Classical
in our ancient samples alongside modern data from the literature5 Antiquity period cats from Bulgaria, Jordan and Turkey (8, 11 and
clearly separates the five clades of F. silvestris (posterior probabili- 16, respectively, Fig. 1a–c). This range expansion is more evident
ties >0.88, Supplementary Fig. 3, Supplementary Methods) and between the 5th and 13th centuries AD, when the two IV-C lineages
the five subclades of F. s. lybica (posterior probabilities >0.77). found in ancient Egyptian cats became substantially more frequent
We examined the phylogeographic pattern and its changes across both in Europe (78%; 7 out of 9) and in SWA (46%; 32 out of 70). By
time by grouping the mtDNA haplotypes from our study into nine contrast, none of the 41 European and 18 southwest Asian cats from
time bins (Fig. 1c). archaeological contexts predating the 8th century BC possessed IV-C
haplotypes (Fisher’s exact test; P < 0.001 in both cases).
Ancient European wildcats. We found the mtDNA clade I, represen- The territorial behaviour of cats and the rapid reconfiguration
tative of European wildcats (F. s. silvestris), exclusively in Europe. From of the phylogeographic pattern observed in Europe and SWA suggest
the Mesolithic period to the 8th century bc in Western Europe (geo- that cats carrying IV-C haplotypes were spread by humans through-
graphic locations 1-5 in Fig. 1a, c), all cats analysed (9 out of 9) carried out the eastern Mediterranean region in Classical antiquity. Further
clade I mtDNA, whereas in southeast Europe (6-8) we observed similar expansion occurred during the Medieval period, whereby the IV-C1
frequencies of clade I (n = 13, 42%) and clade IV (n = 18, 58%), repre- haplotype was found at the Viking trading port of Ralswiek on the
sentative of F. s. lybica. The latter was mostly represented by one of the Baltic Sea (1, Fig. 1a–c) by the 7th century AD, and at the Iranian port
lineages of subclade IV-A, hereafter IV-A1 (Supplementary Figs 1, 3), of Siraf by the 8th century AD (17). In the Balkans, IV-C1 persisted
the earliest occurrence of which, in our dataset, dates back to 7700 bc throughout Medieval times up to the present (8). Translocation
in Romania (7) (Supplementary Data 1), and which is still present of cats over even longer distances was observed by the presence
today in European wild (8) and domestic cats5. The occurrence of a of Asian wildcat (F. s. ornata) mtDNA at the Roman–Egyptian
F. s. lybica mitotype in pre-Neolithic southeast Europe indicates that port of Berenike on the Red Sea (1st–2nd century AD; 23, Fig. 1a–c)
the native range of this subspecies extended beyond the Bosporus. and at Medieval coastal sites in Turkey (9, 10).

2 NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION ARTICLES
a b
F. margarita
1
2
3 4 7
8 14
10
5 6 12 13 0.002
9 11
24 15
25
20 16 17
21
22
23 18 19

26

27 29
28
30

31 F. silvestris silvestris
F. silvestris lybica
F. silvestris ornata 1 2 1 2
* *

F. s. ornata
F. s. cafra
F. s. silvestris
F. s. bieti
F. silvestris cafra A B C D E
F. silvestris bieti F. silvestris lybica
(IV) (V)
(II)
(III)
(I)

c Early Neolithic in Early Bronze Age Classical

southeast Europe in southwest Asia Antiquity Roman and Byzantine empire Ottoman Empire

>6500 BC 4500 BC 3100 BC 2000 BC 800 BC AD 500 AD 1300 AD 1900 Modern wild
1
2
3
4

Europe
5
6

8
9

10

11

Southwest Asia
12
13
14
15

16

17
18
19
20
21
22
23
24
Africa

25
26
27
28
29
30
31

Figure 1 | Spatio-temporal representation of cat maternal genealogies. a, Map showing the present-day distribution of Felis silvestris4 with the
geographic range of each subspecies as reported in literature5 and inferred from the data presented herein. b, Tree of mtDNA lineages observed in our
ancient samples and in modern wild and domestic cats from literature5. c, Spatio-temporal depiction of ancient cat haplotypes as depicted with symbols
from the tree in b. Rows represent the approximate geographic provenance of the samples as reported in the map in a whereas the columns pertain to
chronological periods, the limits of which were selected to separate the prehistoric and historical periods evenly, to unambiguously assign each sample to
a single bin and to take historic events into account that could have affected human–cat interactions, as indicated on the timeline above. A dot inside the
symbols indicates AMS-radiocarbon-dated samples; dashed lines inside the symbols indicate incomplete mtDNA profiles; Near Eastern modern wildcats
from literature5 are indicated by grey-shaded bins. Numbers in a and in c represent the approximate geographic locations of the sites from which the
samples are derived as reported in Supplementary Table 5.

NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol 3

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
ARTICLES NATURE ECOLOGY & EVOLUTION

BC

BC
BC

0
0
BC

0
BC
0

0
0
00

50
0
50

00

19
13
0
0

20

80
>6

45

AD

AD
AD
31
Europe

Mackerel Blotched
(TaM/TaM or TaM/Tab) (Tab/Tab)
Southwest Asia
Africa

TaM, wild type Tab, blotched (W841X)

Figure 2 | Spatio-temporal representation of the alleles determining the phenotypic variation in the shape of tabby patterns, mackerel (TaM)
and blotched (Tab). To overcome issues of potential allelic drop-out, each individual is defined by at least one observed allele, except for the
few instances in which both alleles were detected. The image shows a ‘cat under the chair’ with a tabby mackerel marking, typical of F. silvestris lybica
(Anna (Nina) Macpherson Davies, Copy of Wall Painting from Private Tomb 52 of Nakht, Thebes (I, 1, 99–102) Cat Eating Fish. Photo: © Ashmolean
museum, Oxford, UK).

Coat pattern. The domestication process has not markedly changed Here, we show that mitochondrial lineages corresponding to
the morphology of cats, and few traits can be used today to these two purported domestication centres contributed at different
identify wild or hybrid populations. Of the few traits available, times to the gene pool of modern domestic cats. We deduced this
the most widely used is the tabby coat marking16. The transmem- by establishing the ancestral phylogeography of wild cats in the Old
brane aminopeptidase Q (Taqpep) gene is responsible for the World and by observing its reconfiguration through time, which
tabby phenotypic variation in cats, with a single SNP distinguish- reveals the spread of cats through human agency following ancient
ing most of the mackerel and blotched patterns that are character- land and maritime trade routes.
istic of the wild and domestic patterns, respectively13. To develop
a temporal framework for the emergence of a variation in coat Distribution of wildcats. Our aDNA data (Fig. 1a and Supplementary
pattern typical of domestic cats, we investigated the three SNPs Fig. 4) show a clear phylogeographic structure. F. s. silvestris was
in the Taqpep gene13. We found that the recessive allele respon- confined to Europe, whereas F. s. lybica was found in SWA and
sible for the blotched-tabby pattern in 80% of present-day cats southeast Europe. A clear understanding of the present distribu-
(W841X) occurred in our ancient dataset not earlier than the tion of wild F. s. lybica in Anatolia has proven elusive until now
Medieval period in SWA (3%, minimum number of total alleles, owing to a lack of genetic data. It has commonly been assumed
see Methods) (Fig. 2). Thereafter, its frequency increased in Europe, that the native range of the modern European wildcat includes
SWA, and Africa (50% in total), showing late expansion of this typi- Anatolia5,18,19. Our phylogeographic reconstruction demonstrates
cally domestic allele. that mtDNA clade IV, corresponding to F. s. lybica, was predomi-
nant in Anatolia for many millennia beginning in the Neolithic
Discussion period at the latest. Not a single instance of clade I, corresponding to
Zooarchaeological and iconographic evidence for early cat F. s. silvestris, was detected in our samples from SWA (Fig. 1).
domestication. Owing to the paucity of cat remains in the archeo- Nevertheless, we cannot exclude its presence in the wilds of
logical record and the lack of established osteometric features dis- Anatolia, in particular in the forest and mountain refuges of north-
tinguishing remains from wild and domestic F. s. lybica2, current ern Anatolia and the Caucasus.
hypotheses about early cat domestication are grounded in scanty We found two distinct IV-A mitotypes on either side of the
evidence when compared to other domesticated animals. A com- Bosporus. In Anatolia, from around 8000 BC to 800 BC, almost all
plete skeleton found in Cyprus in association with a human burial cats (12 out of 14) belonged to the IV-A* mitotype. By contrast,
dated to around 7500 BC suggests that cats were probably tamed cats carrying a distinct mitotype, IV-A1, were present in southeast
by early Neolithic sedentary communities that had been growing Europe by the beginning of the 8th millennium BC. This suggests that
cereals in SWA, concomitant with the emergence of commensal F. s. lybica was distributed across Anatolia from the early Holocene
rodents3. Similarly, the skeletons of six cats in an elite Predynastic epoch at the latest, prior to the formation of the present-day exten-
cemetery in Egypt, around 3700 BC, may suggest a close cat–human sion of the Black Sea, and that it made its way to southeast Europe
relationship in early ancient Egypt2. before the onset of farming in the Neolithic period. A split in an
The iconography of Pharaonic Egypt constitutes a key source of ancestral Anatolian cat population in the late Pleistocene epoch,
information about the species’ relationship with humans, and has presumably during the Last Glacial Maximum, followed by local
motivated the traditional belief that cat domestication took place in differentiation and/or drift and founder effect, might have been
Egypt1,17. Numerous depictions in Egyptian art from the 2nd mil- responsible for the distribution of distinct clade IV mtDNA lineages
lennium BC document a progressive tightening of the relationship in Anatolia and southeast Europe. F. s. silvestris and F. s. lybica occur
between human and cat, as illustrated in particular by the popu- across different biotopes that include, respectively, temperate wood-
larization of the motif of the ‘cat under the chair’ of women after land and open bushland4. The expansion of open bushlands during
around 1500 BC1,17. the Late Pleistocene epoch might have attracted F. s. lybica into the

4 NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION ARTICLES
Balkans when the Bosporus was a land bridge and the Balkans rep- cat pelts as cloth items25. Spread of the black rat (Rattus rattus)
resented a refuge for warm-adapted species20,21. and house mouse (Mus musculus) by sea routes as early as the
Currently, IV-A1 is found in the European wildcat population Iron age, documented by zooarchaeological and genetic data26,
and also in modern domestic cats5 (Fig. 1). Our data imply that probably also encouraged cat dispersal for the control of these
admixture episodes potentially occurring through time between new pests.
overlapping populations of wild F. s. silvestris and F. s. lybica could Increased translocation as a result of long-distance trade is also
be in part responsible for F. s. lybica mtDNA introgression in pres- witnessed by the finding of Asian F. s. ornata mtDNA in cats from
ent-day European wildcat populations. Conservation programs the Roman Red Sea port of Berenike (1st–2nd century AD) and from
should also take into account past natural admixture when aiming Turkey in the 6–7th century AD. This was probably the result of
at neutering and removing hybrids that are believed to have a role in increasingly intensive and direct trade connections between south
cryptic extirpations of wild F. s. silvestris populations4. Asia and the Mediterranean basin via the Indian Ocean and Red
Sea27, but possibly also via the Silk Road connecting central Asia
Origin and dispersal of domestic cats. Our data show that mito- with Anatolia28. Long-distance maritime routes29, as described for
type IV-A* had a wide distribution stretching across Anatolia from instance in the 1st century AD Periplus of the Erythraean Sea, prob-
west to east throughout the Neolithic, Bronze age and Iron age. ably explain the occurrence of IV-A*, typical of SWA, as far south as
Its range may have extended as far south as the Levant, where we East Africa (30, Fig. 1a–c).
inferred the presence of subclade IV-B. These findings suggest that Upon arrival in these various new locations, introduced cats
in the Fertile Crescent, cats that developed a commensal relation- reconfigured the phylogeographic landscape of the species through
ship with early farming communities during the Neolithic period admixture with local tame or wild cats, leading to a transfer of deeply
carried at least mitotypes IV-A* and IV-B. Mitotype IV-A* later divergent mitochondrial lineages in the domestic pool (IV-D/E and
spread to most of the Old World, representing the Near Eastern possibly III-F. s. ornata although these lineages are found only at
contribution to the mtDNA pool of present-day domestic cats. low frequency in modern domestic cats5). Modern genetic data
This spread may have started as early as around 4400 BC into have shown that admixture with domestic cats still occurs today
southeast Europe, the date of the first appearance of IV-A* in our in European wildcat populations10,16, and intensive conservation
European dataset, and therefore subsequent to the neolithisation programs have been implemented to preserve the integrity of
of Europe. This suggests that the human-mediated translocation of F. s. silvestris4,30.
cats began in prehistoric times, corroborating the interpretation of
the finding of a cat buried around 7500 BC in Cyprus3. We also found Evolution of the tabby pattern. Our study also sheds light on
IV-A* in cat remains from the Roman–Egyptian port of Berenike the late emergence in domestic cats of a key phenotypic trait,
on the Red Sea and in one Egyptian mummy (Fig. 1a–c), which may the blotched coat marking caused by a SNP in the Taqpep gene13.
hint at an introduction of cats from SWA to Egypt. Wildcats exhibit a mackerel-like coat pattern, whereas the blotched
Our data provide the first evidence for an African origin for one pattern is common in many modern domestic breeds13. In our data-
of the mitochondrial lineages of present-day domestic cats, namely set, the first occurrence of the recessive allele W841X, which is asso-
clade IV-C. Indeed, we found the lineages C1 and C* in the majority ciated with the blotched markings, dates to the Ottoman Empire in
of Egyptian cat mummies. These cats were worshipped and, dur- SWA and later increases in frequency in Europe, SWA and Africa
ing the Greco–Roman period, kept in temple precincts to be mum- (Fig. 2). This result is in agreement with the iconography from the
mified17. We show that, despite a local ban on cat trading being Egyptian New Kingdom through the European Middle Ages, where
imposed in Egypt as early as 1700 BC22, cats carrying IV-C mtDNA cats’ coats were mainly depicted as striped, corresponding to the
spread to most of the Old World. The increasing popularity of cats mackerel-tabby pattern of the wild F. s. lybica1,17 (Fig. 2). It was only
among Mediterranean cultures and particularly their usefulness on in the 18th century AD that the blotched markings were common
ships infested with rodents and other pests presumably triggered enough to be associated with the domestic cat by Linnaeus13, and
their dispersal across the Mediterranean22. Indeed, depictions of cats physical traits started to be selected only in the 19th century AD for
in domestic contexts, already frequent during the New Kingdom the production of fancy breeds15. Thus, both our data and recent
in Egypt around 1500 BC (‘cat under the chair’, Fig. 2), are found genomic data11 suggest that cat domestication in its early stages may
on Greek artifacts from as early as the end of the 6th century BC have affected mainly some behavioural features, and distinctive
(Supplementary Methods). The Egyptian cat must have been very physical and aesthetic traits may have been selected for only recently.
popular, as IV-C1 and C* represented more than half of the mater- A similar pattern of late emergence of other phenotypic traits
nal lineages in Western Anatolia during the 1st millennium AD, has been observed in chicken31, but contrasts with what has been
and occurred twice as frequently as the local mitotype IV-A*. This observed in horses, where coat-colour differentiation appeared at
suggests that the Egyptian cat had properties that made it attrac- an early stage of domestication32.
tive to humans, presumably acquired during the tightening of the
human–cat relationship that developed during the Middle and New Conclusive remarks. The comprehensive aDNA genetic study
Kingdoms and became even stronger afterwards1,17. As the most of cats across time and space that we present, provides answers
pronounced genetic changes that distinguish wild and domestic cats to longstanding questions concerning the domestication pro-
are apparently linked to behaviour11, it is tempting to speculate that cess of the cat and contributes to a better understanding of how
the success of the Egyptian cat is underlain by changes in its socia- humans have reshaped global biodiversity through species trans-
bility and tameness. locations23,26,33. By revealing the original phylogeographic dis-
North of the Alps, domestic cats appeared soon after the Roman tribution of wildcats and its profound modification through
conquest, yet remained absent outside the Roman territory until human-mediated dispersal of tamed cats through time, we show
Late Antiquity23. In medieval times it was compulsory for seafar- that both Near Eastern and Egyptian cat lineages contributed
ers to have cats onboard their ships24, leading to their dispersal at different times to the maternal genetic pool of domestic cats,
across routes of trade and warfare. This evidence explains, for with one or other present in the vast majority of present-day cat
example, the presence of the Egyptian lineage IV-C1 at the Viking breeds. Cat domestication was a complex, long-term process
port of Ralswiek (7–11th century AD)24. The expansion of the domes- featuring extensive translocations that allowed admixture events
tic cat may have been fostered by a diversification in their cultural between geographically separated cat populations at different
usage, which in Medieval Europe included the trade of domestic points in time.

NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol 5

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
ARTICLES NATURE ECOLOGY & EVOLUTION

Methods References
Ancient DNA analyses. Ancient DNA analysis was performed in dedicated 1. Engels, D. W. Classical Cats: The Rise and Fall of the Sacred Cat
aDNA facilities in Paris and Leuven from bone, teeth, skin and hair samples (Routledge, 2001).
(the last two when available in Egyptian mummies) of 352 ancient cats. The ages 2. Van Neer, W., Linseele, V., Friedman, R. & De Cupere, B. More evidence for
of the archaeological remains were determined using direct accelerator mass cat taming at the Predynastic elite cemetery of Hierakonpolis (Upper Egypt).
spectrometry (AMS) radiocarbon dating (KIK-IRPA, Belgium), stratigraphic J. Arch. Sci. 45, 103–111 (2014).
associations with AMS dates, and contextual archaeological evidence 3. Vigne, J.-D., Guilaine, J., Debue, K., Haye, L. & Gérard, P. Early taming of the
(Supplementary Data 1). All dates in the text are reported in calibrated radiocarbon cat in Cyprus. Science 304, 259 (2004).
years BC. DNA was also extracted from claws and skin samples of 28 modern 4. Yamaguchi, N., Kitchener, A. C., Driscoll, C. & Nussberger, B.
wildcats from Bulgaria and east Africa (Supplementary Methods). Felis silvestris. IUCN http://dx.doi.org/10.2305/IUCN.UK.2015-2.RLTS.
Amplification of nine mtDNA and three nuclear DNA fragments in the T60354712A50652361.en (2015).
Taqpep gene was preceded by the elimination of carry-over contamination based 5. Driscoll, C. A. et al. The Near Eastern origin of cat domestication. Science
on the dUTP/UNG system34 and carried out in three separate multiplex PCRs. 317, 519–523 (2007).
Phylogenetically informative SNPs in the mtDNA were selected following the 6. Driscoll, C. A., Macdonald, D. W. & O’Brien, S. J. From wild animals to
most up-to-date worldwide cat phylogeny5 (Supplementary Fig. 1). We targeted domestic pets, an evolutionary view of domestication. Proc. Natl Acad. Sci.
42 informative SNPs in nine short regions distributed across the mitochondrial USA 106, 9971–9978 (2009).
ND5, ND6 and CytB genes that recapitulate the most salient features of the 7. Price, E. O. Animal Domestication and Behavior (CABI Publishing, 2002).
phylogeny obtained with longer portions of these genes (Supplementary Fig. 1). 8. Wozencraft, C. in Mammal Species of the World 3rd edn (eds Wilson, D. E. &
The diagnostic SNPs were screened with Pyrosequencing assays (Biotage, Qiagen) Reeder, D. M.) http://go.nature.com/2p5vsxb (Johns Hopkins Univ. Press, 2005).
and sequencing on a PGM Ion Torrent platform (Institut Jacques Monod, Paris) 9. Zeder, M. A. in Biodiversity in Agriculture (eds Gepts, P. et al.) 227–259
of amplicon libraries following the ‘aMPlex Torrent’ workflow and downstream (Cambridge Univ. Press, 2012).
sequence analysis with a bash script described elsewhere14 (Supplementary Code). 10. Mattucci, F., Oliveira, R., Lyons, L. A., Alves, P. C. & Randi, E. European
The aMPlex Torrent approach combines the sensitivity of multiplex PCR with wildcat populations are subdivided into five main biogeographic groups:
the power and throughput of next-generation sequencing14 and made it possible consequences of Pleistocene climate changes or recent anthropogenic
to screen and analyse a large number of poorly preserved ancient samples in fragmentation? Ecol. Evol. 6, 3–22 (2015).
parallel. We obtained mtDNA sequences from 209 out of 352 ancient cats (59%; 11. Montague, M. J. et al. Comparative analysis of the domestic cat genome
Supplementary Data 1, 2), with expectedly lower success rates for old samples from reveals genetic signatures underlying feline biology and domestication.
hot environments (Supplementary Fig. 2). The mtDNA profiles ranged from 286 to Proc. Natl Acad. Sci. USA 111, 17230–17235 (2014).
449 bp, 12 of which were incomplete profiles generated from two to seven mtDNA 12. Axelsson, E. et al. The genomic signature of dog domestication reveals
fragments. The authentication criteria adopted rely on: (i) strict contamination adaptation to a starch-rich diet. Nature 495, 360–364 (2013).
prevention controls including physical containment as well as material and reagent 13. Kaelin, C. B. et al. Specifying and sustaining pigmentation patterns in
decontamination34–36; (ii) extensive replications performed through independent domestic and wild cats. Science 337, 1536–1541 (2012).
PCRs (at least two, but up to eight with one up to three independent DNA 14. Guimaraes, S. et al. A cost-effective high-throughput metabarcoding
extracts). For samples where the low DNA content decreased data reliability, approach powerful enough to genotype ~44 000 year-old rodent remains
we increased the number of replicates and used different PCR assays, multiplex from Northern Africa. Mol. Ecol. Resour. http://dx.doi.org/10.1111/1755-
and simplex PCR, pyrosequencing and the aMPlex Torrent method performed 0998.12565 (2016).
in independent laboratories (Paris and Leuven) so that samples with different 15. Driscoll, C. A., Clutton-Brock, J., Kitchener, A. C. & O’Brien, S. J. The taming
preservation levels could be genotyped with similar reliability. More details about of the cat. Genetic and archaeological findings hint that wildcats became
DNA extraction, amplification, sequencing and the authentication criteria can be housecats earlier—and in a different place—than previously thought. Sci. Am.
found in the Supplementary Methods. 300, 68–75 (2009).
16. Pierpaoli, M. et al. Genetic distinction of wildcat (Felis silvestris) populations
Phylogeographic analyses. Each specimen was assigned to an mtDNA clade using in Europe, and hybridization with domestic cats in Hungary. Mol. Ecol. 12,
the terminology previously proposed5, including specimens with an incomplete 2585–2598 (2003).
profile (shapes with an inner dashed line in Fig. 1c). Owing to our streamlined 17. Málek, J. The Cat in Ancient Egypt, revised edn 144 (British Museum Press,
sequencing assay, some of the subclades and lineages of IV-A and IV-C observed 1993–2006).
in the 2007 study were collapsed into a single haplotype, which we named 18. Spassov, N., Simeonovski, V. & Spiridonov, G. The wild cat (Felis silvestris
IV-A* and IV-C*, respectively (Supplementary Fig. 1). The ancient and modern Schr.) and the feral domestic cat: problems of the morphology, taxonomy,
sequences generated here were aligned to 159 sequences from Driscoll’s study5. identification of the hybrids and purity of the wild population.
A Bayesian tree of 66 unique 286 bp-long haplotypes (Supplementary Fig. 3) Hist. Nat. Bulg. 8, 101–120 (1997).
was constructed as described in detail in the Supplementary Methods. An ML 19. Kitchener, A. C. & Rees, E. E. Modelling the dynamic biogeography of the
tree, obtained as described in the Supplementary Methods, had the same topology. wildcat: implications for taxonomy and conservation. J. Zool. 279,
Frequencies of haplotypes A* and A1 in Anatolia and southeast Europe, and 144–155 (2009).
of clade C in before and after the 8th century AD in SWA, were tested using a 20. Ivanova, S. et al. Magura Cave, Bulgaria: a multidisciplinary study of Late
Fisher’s exact test. Pleistocene human palaeoenvironment in the Balkans. Quat. Int. 415,
86–108 (2016).
Nuclear markers. We typed allelic variations within the Taqpep gene associated 21. Dubey, S. et al. Molecular evidence of Pleistocene bidirectional faunal
with coat-colour pattern differences—W841X, D228N and T139N13. The results exchange between Europe and the Near East: the case of the bicoloured
presented here are intended to be indicative of allele frequencies. Given the low shrew (Crocidura leucodon, Soricidae). J. Evol. Biol. 20, 1799–1808 (2007).
level of independent replications of our assay and the risk of allelic dropout, 22. Faure, E. & Kitchener, A. C. An archaeological and historical review of the
especially in ancient degraded samples, we could not ascertain genotypes, relationships between felids and people. Anthrozoos 22, 221–238 (2009).
except for a few heterozygous samples showing a fairly high number of reads 23. Peters, J. Römische Tierhaltung und Tierzucht. Eine Synthese aus
in at least two independent amplifications (Fig. 2, Supplementary Methods archäozoologischer Untersuchung und schriftlich-bildlicher Überlieferung 444
and Supplementary Data 2). Assuming that none of the alleles is amplified (Verlag Marie Leidorf, 1998).
preferentially, and adopting a conservative strategy that accounted for the 24. Johansson, F. & Hüster, H. Untersuchungen an Skelettresten von Katzen aus
minimum number of alleles observed, our data across the spatial and temporal Haithabu (Ausgrabung 1966 – 1969) 86 (Wachholtz, 1987).
framework showed that 7 out of 67 successfully amplified cat samples possessed 25. Ewing, E. Fur in dress 168 (Batsford, 1981).
at least one mutant Tabby-W841X allele, of which two were heterozygotes (BMT2 26. Jones, E. P., Eager, H. M., Gabriel, S. I., Johannesdottir, F. & Searle, J. B.
and MET9). In 88 cats we could screen the allele D228N and in all instances Genetic tracking of mice and other bioproxies to infer human history.
we observed the wild-type. Among 63 cats successfully screened for T139N, Trends Genet. 29, 298–308 (2013).
we detected the mutant allele (C to A) in three specimens. 27. Sidebotham, S. E. Berenike and the Ancient Maritime Spice Route
(Univ. California Press, 2011).
Code availability. A bash script and accessory fasta and gff files for data 28. Ball, W. Rome in the East: The Transformation of an Empire 523
analysis of the aMPlex Torrent data are provided as Supplementary Code. (Routledge, 2000).
29. Boivin, N., Crowther, A., Helm, R. & Fuller, D. East Africa and Madagascar
Data availability. Sequence data that support the findings of this study have been in the Indian Ocean world. J. World Prehist. 26, 213–281 (2013).
deposited in Dryad (http://dx.doi.org/10.5061/dryad.g4p30). 30. Witzenberger, K. A. & Hochkirch, A. The genetic integrity of the ex situ
population of the European wildcat (Felis silvestris silvestris) is seriously
Received 10 October 2016; accepted 10 March 2017; threatened by introgression from domestic cats (Felis silvestris catus).
published 19 June 2017 PLoS ONE 9, e106083 (2014).

6 NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
NATURE ECOLOGY & EVOLUTION ARTICLES
31. Girdland Flink, L. et al. Establishing the validity of domestication genes École Normale Supérieure, Paris, for granting access to the pyrosequencer;
using DNA from ancient chickens. Proc. Natl Acad. Sci. USA 111, M. Larmuseau and A. Van Geystelen for discussions and assistance with nuclear
6184–6189 (2014). SNP analyses; K. Knaepen, M. Coomans, and A. Giucca for support in laboratory
32. Ludwig, A. et al. Coat color variation at the beginning of horse procedures in Leuven; and J. Nackaerts of the veterinary hospital Kruisbos
domestication. Science 324, 485 (2009). (Wezemaal, Belgium) for providing cat blood samples. We also thank the curators
33. Boivin, N. L. et al. Ecological consequences of human niche construction: of the following collections for facilitating access to the material under their care
examining long-term anthropogenic shaping of global species distributions. and the permission to take tissue samples: the Royal Museum for Central Africa
Proc. Natl Acad. Sci. USA 113, 6388–6396 (2016). (Tervuren, Belgium), the Muséum National d’Histoire Naturelle and Musée du
34. Pruvost, M., Grange, T. & Geigl, E.-M. Minimizing DNA-contamination Louvre (Paris, France), the British Museum and Natural History Museum
by using UNG-coupled quantitative real-time PCR (UQPCR) on degraded (London, UK) and the Bavarian State Collection of Anthropology and Palaeoanatomy
DNA samples: application to ancient DNA studies. Biotechniques 38, (Munich, Germany).
569–575 (2005).
35. Champlot, S. et al. An efficient multistrategy DNA decontamination
procedure of PCR reagents for hypersensitive PCR applications. PLoS ONE 5, Author contributions
e13042 (2010). The project was initiated by W.V.N., E.-M.G., C.O., T.G. and R.D. The ancient DNA study
36. Bennett, E. A. et al. Library construction for ancient genomics: single strand was conceived and designed by T.G., E.-M.G. and C.O. C.O. carried out the molecular
or double strand? Biotechniques 56, 289–300 (2014). laboratory work, with support of S.G. and analysed the data. J.D. generated the aMPlex
Torrent data. The archaeological bone samples were provided by W.V.N., B.D.C., J.P.,
Acknowledgements N.S., M.E.P., N.Bo., A.M.-M., A.Bă., C.B., N.Be., A.Bo., H.B., J.C., A.C., L.L., N.M., H.M.,
This research has been funded by the IAP program (BELSPO), the KU Leuven V.O., M.O., M.O., O.P., E.M.Q.M., J.S., U.W., and W.V.N. and B.D.C. were responsible for
BOF Centre of Excellence Financing on CAS, and the CNRS (T.G. and E.-M.G.). their curation and archaeozoological recording. The authors’ list from A.Ba. to U.W. is
The high-containment laboratory of the Institut Jacques Monod, Paris was supported by in alphabetical order. C.O., E.M.G. and T.G. wrote the paper. W.V.N., B.D.C., J.P., N.S.,
a grant to E.-M.G. from the University Paris Diderot, ARS 2016-2018. The sequencing M.E.P., N.Bo., A.M.-M. contributed to further discussion about the interpretation of the
facility of the Institut Jacques Monod, Paris, and J.D., were supported by grants to data and the outline of the paper. N.Bo. and M.E.P. revised the English. All the authors
T.G. from the University Paris Diderot, the ‘Fondation pour la Recherche Médicale’ gave final approval for publication.
(DGE20111123014), and the ‘Région Ile-de-France’ (grant 11015901). C.O. was
supported by the FWO mobility program (V4.519.11N, K2.197.14N, K2.057.14N). Additional information
Faunal research carried out by J.P. and team in Anatolia received funding by the Supplementary information is available for this paper.
German Research Foundation (DFG PE424/10-1,2). Research by N.Bo. M.E.P., and Reprints and permissions information is available at www.nature.com/reprints.
A.C. was supported by an ERC grant (206148) and UK NERC Radiocarbon Facility
grant (NF/2012/2/4). The archaeological and archaeozoological research conducted Correspondence and requests for materials should be addressed to C.O., T.G. and E.-M.G.
by A.Bă. and A. Bo. was supported by the Romanian National Authority for Scientific How to cite this article: Ottoni, C. et al. The palaeogenetics of cat dispersal in the
Research, UEFISCDI (PN-II-ID-PCE-2011-3-1015 and PN-II-RU-TE-2014-4-0519). ancient world. Nat. Ecol. Evol. 1, 0139 (2017).
Research at Songo Mnara was directed by S. Wynne-Jones and J. Fleisher with support Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in
from the National Science Foundation (BCS1123091) and the Arts and Humanities published maps and institutional affiliations.
Research Council (AH/J502716/1). We thank G. Larson and E. A. Bennett for critical
reading of the manuscript; the Ufficio Beni Archeologici della Provincia Autonoma di
Bolzano for granting access to the archaeological material of Galgenbühel/Dos de la Competing interests
Forca and J. Crezzini for help in sampling; M.-A. Félix, Institut Jacques Monod and The authors declare no competing financial interests.

NATURE ECOLOGY & EVOLUTION 1, 0139 (2017) | DOI: 10.1038/s41559-017-0139 | www.nature.com/natecolevol 7

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
View publication stats