Content uploaded by João Luís Cardoso
Author content
All content in this area was uploaded by João Luís Cardoso on Nov 09, 2022
Content may be subject to copyright.
rf-
Jt
✓
=
f
;t/1· tf
�•h
I
�·
I
t
I[ • �
RESEARCH ARTICLE
◥
DOMESTICATION
The genomic history and global expansion
of domestic donkeys
Evelyn T. Todd
1
, Laure Tonasso-Calvière
1
, Loreleï Chauvey
1
, Stéphanie Schiavinato
1
,
Antoine Fages
1
, Andaine Seguin-Orlando
1
, Pierre Clavel
1
, Naveed Khan
1,2
, Lucía Pérez Pardal
3,4
,
Laura Patterson Rosa
5
, Pablo Librado
1
, Harald Ringbauer
6
, Marta Verdugo
7
, John Southon
8
,
Jean-Marc Aury
9
, Aude Perdereau
9
, Emmanuelle Vila
10
, Matilde Marzullo
11
, Ornella Prato
11
,
Umberto Tecchiati
11
, Giovanna Bagnasco Gianni
11
, Antonio Tagliacozzo
12
, Vincenzo Tinè
13
,
Francesca Alhaique
12
, João Luís Cardoso
14,15
, Maria João Valente
16
, Miguel Telles Antunes
17
,
Laurent Frantz
18,19
, Beth Shapiro
20,21
, Daniel G. Bradley
7
, Nicolas Boulbes
22
, Armelle Gardeisen
23
,
Liora Kolska Horwitz
24
, Aliye Öztan
25
, Benjamin S. Arbuckle
26
, Vedat Onar
27
, Benoît Clavel
28
,
Sébastien Lepetz
28
, Ali Akbar Vahdati
29
, Hossein Davoudi
30
, Azadeh Mohaseb
28,30
,
Marjan Mashkour
28,30,31
, Olivier Bouchez
32
, Cécile Donnadieu
32
, Patrick Wincker
9
,
Samantha A. Brooks
33
, Albano Beja-Pereira
3,4,34,35
, Dong-Dong Wu
36,37
, Ludovic Orlando
1
*
Donkeys transformed human history as essential beasts of burden for long-distance movement, especially
across semi-arid and upland environments. They remain insufficiently studied despite globally expanding
and providing key support to low- to middle-income communities. To elucidate their domestication history,
we constructed a comprehensive genome panel of 207 modern and 31 ancient donkeys, as well as
15 wild equids. We found a strong phylogeographic structure in modern donkeys that supports a single
domestication in Africa ~5000 BCE, followed by further expansions in this continent and Eurasia and
ultimately returning to Africa. We uncover a previously unknown genetic lineage in the Levant ~200 BCE,
which contributed increasing ancestry toward Asia. Donkey management involved inbreeding and the
production of giant bloodlines at a time when mules were essential to the Roman economy and military.
Domestic donkeys (Equus asinus)have
facilitated the movement of goods and
people for millennia, enabling trade and
transport across a broad spectrum of
landscapes (1). Despite their importan ce
to ancient pastoral societies, little is known
about the deep history of donkeys and the im-
pact of human management on their genomes.
This is most likely due to their undervalued
status and loss of utility in modern industrial-
ized societies. Donkeys are, however, extraordi-
nary working animals and remain essential for
developing communities, especially in semi-
arid environments (2). Understanding their
genetic makeup is not only key to assessing
their contribution to human history but also
to improving their local management in the
future.
The current archaeological record of early
donkeys is limited (1,3), which makes their
domestic origins and spread through the
world contentious. The reduced body size of
zooarchaeological ass remains in Egypt at El
Omari (4800 to 4500 BCE) and Maadi (4000 to
3500 BCE) has been interpreted as early evi-
dence of domestication (4–7). Carvings on the
LibyanpalettethatwerefoundinAbydos,
Egypt (3200 to 3000 BCE), depict lines of walk-
ing asses, cattle, and sheep, which also suggests
a domestication context (8,9). Together with
contemporary remains from the same region
that show morphological evidence for load ca r-
rying (10), these findings suggest that donkeys
could have been first domesticated within a
range extending from the Northeastern Sahara,
theNileValley,theAtbaraRiver,andtheRed
Sea Hills to Eritrea. In this model, donkeys were
domesticated by pastoralists to assist with mo-
bility around 5500 to 4500 BCE b eca use of th e
large-scale aridification of the Sahara (1). Inde-
pendent evidence based on patterns of mito-
chondrial (11,12) and nuclear sequence variation
(13) also points to the African origins of the don-
key, owing to their closer proximity to African
wild asses (Equus africanus spp.) than to Asian
wild asses (Equus hemionus spp.).
However, candidate regions outside of Africa
are also proposed as alternative domestication
centers. In Ash Shuman (Yemen), for example,
ass remains of disputed domestic status predate
those from Egypt by 2000 years (~6500 BCE)
(14). Likewise, textual, iconographic, and zoo-
archaeological material indicates a possible
additional center in Mesopotamia during the
fourth and third millennia BCE (15–19), a
context in which first-generation hybrids of
donkeys and Syrian onagers have been iden-
tified genetically (20). The segregation of
mitochondrial variation in two main clades
may also support a dual domestication process,
for which the Nubian wild ass (Equus africanus
africanus) is securely identified as the progen-
itor of Clade I (11,12). As for the ancestor of
Clade II, it could either be the extinct Atlas
wild ass (Equus africanus atlanticus), which
was endemic to Northern Africa, or another un-
described subspecies that potentially ranged
outside of Africa. Whether a single, maternally
RESEARCH
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 1of9
1
Centre d’Anthropobiologie et de Génomique de Toulouse (CAGT), CNRS UMR 5288, Université Paul Sabatier, Toulouse 31000, France.
2
Department of Biotechnology, Abdul Wali Khan University,
Mardan 23200, Pakistan.
3
CIBIO, Centro de Investigação em Biodiversidade e Recursos Genéticos, InBIO Laboratório Associado, Campus de Vairão, Universidade do Porto, Vairão 4485-661,
Portugal.
4
BIOPOLIS Program in Genomics, Biodiversity and Land Planning, Campus de Vairão, Universidade do Porto, Vairão 4485-661, Portugal.
5
Department of Animal Science, Sul Ross State
University, Alpine, TX 79830, USA.
6
Department of Archaeogenetics, Max Planck Institute for Evolutionary Anthropology, Leipzig 04103, Germany.
7
Smurfit Institute of Genetics, Trinity College
Dublin, Dublin D02 PN40, Ireland.
8
Earth System Science Department, University of California, Irvine, CA 92697, USA.
9
Genoscope, Institut de biologie François Jacob, CEA, Université d’Evry,
Université Paris-Saclay, Evry 91042, France.
10
Laboratoire Archéorient, Université Lyon 2, Lyon 69007, France.
11
Dipartimento di Beni Culturali e Ambientali, Università degli Studi di Milano, Milan
20122, Italy.
12
Bioarchaeology Service, Museo delle Civiltà, Rome 00144, Italy.
13
Soprintendenza archeologia belle arti e paesaggio per le province di Verona, Rovigo e Vicenza, Verona 37121, Italy.
14
ICArEHB, Campus de Gambelas, University of Algarve, Faro 8005-139, Portugal.
15
Universidade Aberta, Lisbon 1269-001, Portugal.
16
Faculdade de Ciências Humanas e Sociais, Centro de
Estudos de Arqueologia, Artes e Ciências do Património, Universidade do Algarve, Faro 8000-117, Portugal.
17
Centre for Research on Science and Geological Engineering, Universidade NOVA de
Lisboa, Lisbon 1099-085, Portugal.
18
Palaeogenomics Group, Department of Veterinary Sciences, Ludwig Maximilian University, Munich 80539, Germany.
19
School of Biological and Behavioural
Sciences, Queen Mary University of London, London E1 4DQ, United Kingdom.
20
Department of Ecology and Evolutionary Biology, University of California, Santa Cruz, CA 95064, USA.
21
Howard
Hughes Medical Institute, University of California, Santa Cruz, CA 95064, USA.
22
Institut de Paléontologie Humaine, Fondation Albert Ier, Paris / UMR 7194 HNHP, MNHN-CNRS-UPVD / EPCC
Centre Européen de Recherche Préhistorique, Tautavel 66720, France.
23
Archéologie des Sociétés Méditéranéennes, Université Paul Valéry - Site Saint-Charles 2, Montpellier 34090, France.
24
National Natural History Collections, Edmond J. Safra Campus, Givat Ram, The Hebrew University, Jerusalem 9190401, Israel.
25
Archaeology Department, Ankara University, Ankara 06100,
Turkey.
26
Department of Anthropology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599, USA.
27
Osteoarchaeology Practice and Research Center and Department of Anatomy,
Faculty of Veterinary Medicine, Istanbul University-Cerrahpaşa, Istanbul 34320, Turkey.
28
Archéozoologie, Archéobotanique, Sociétés, Pratiques et Environnements, Muséum National d'Histoire
Naturelle, Paris 75005, France.
29
Provincial Office of the Iranian Center for Cultural Heritage, Handicrafts and Tourism Organisation, North Khorassan, Bojnord 9416745775, Iran.
30
Archaezoology
section, Bioarchaeology Laboratory of the Central Laboratory, University of Tehran, Tehran CP1417634934, Iran.
31
Department of Osteology, National Museum of Iran, Tehran 1136918111, Iran.
32
GeT-PlaGe - Génome et Transcriptome - Plateforme Génomique, GET - Plateforme Génome & Transcriptome, Institut National de Recherche pour l’Agriculture, l’Alimentation et l’Environnement,
Castaneet-Tolosan Cedex 31326, France.
33
Department of Animal Science, UF Genetics Institute, University of Florida, Gainesville, FL 32610, USA.
34
DGAOT, Faculty of Sciences, Universidade do
Porto, Porto 4169-007, Portugal.
35
Sustainable Agrifood Production Research Centre (GreenUPorto), Universidade do Porto, Vairão 4485-646, Portugal.
36
State Key Laboratory of Genetic
Resources and Evolution, Kunming Institute of Zoology, Chinese Academy of Sciences, Kunming, Yunnan 650201, China.
37
Kunming Natural History Museum of Zoology, Kunming Institute of
Zoology, Chinese Academy of Sciences, Kunming, Yunnan 650223, China.
*Corresponding author. Email: ludovic.orlando@univ-tlse3.fr
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
inherited marker captures the whole complex-
ity of underlying ancestries can, however, be
questioned after recent results from other ani-
mals [e.g., horses (21)]. Furthermore, previous
analyses of nuclear genetic variation in African
and non-African donkeys have failed to dis-
entangle their origins (13,22). Overall, this
lack of consensus between genetic and archae-
ological data means that the geographic and
temporal origin of donkeys and whether they
were do mesticated more than once remain
uncertain. The global spread of donkeys is also
unclear, as their worldwide patterns of genomic
diversity lack extensive characterization.
Modern donkeys originated in Africa and
spread into Eurasia
To address these issues, we sequenced 49 mod-
ern donkey genomes from underrepresented
regions and combined these with 158 public-
ly available genomes to capture worldwide di-
versity (Fig. 1A and table S1) (13,23–25). We
constructed a fine-scale recombination map
from genomes that encompassed all phyloge-
netic groups, which we used to phase 13,013,551
variants (fig. S1 and tables S3 and S4). Prin-
cipal component analysis (26)revealedstrong
geographical substructuring, with donkeys
from Af rica , Europe, and Asia forming distinct
genetic clusters (Fig. 1B and figs. S2 and S3).
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 2of9
Fig. 1. Modern donkey dataset and population evolutionary history.
(A) Number and geographical distribution of modern donkey samples (n= 207).
Pie charts show the ADMIXTURE proportion of domestic ancestry (gray),
African wild ass ancestry (white), and kiang ancestry (black) averaged across all
individuals from each country (56). For visualization, the total surface of each
pie chart is scaled to 2%. (B) Smartpca (57) of modern donkeys, with the imputed
ancient samples in black. PC, principal component. (C) Treemix phylogeny of
modern domesticates (excluding individuals with high wild introgression, n= 201)
(27). Node supports are estimated from 100 bootstrap pseudo-replicates
(confidence <90% in red). The percentage values indicate admixture proportions
inferred from Treemix (27). (D) SMC++ demographic trajectories (colored)
and split time estimates (black) for pairs of main geographic regions (28),
repeating the analysis on two datasets of three individuals per population
(the second dataset is shown in semitransparency). Modern donkeys are colored
and shaped according to geographical location and continents in all panels.
Country Codes: Algeria (ALG), Brazil (BRA), China (CHI), Canary Islands (CYK),
Denmark (Eas), Egypt (EGY), Spain (ESP), Ethiopia (ETH), Ghana (GHA), Iran
(IRA), Ireland (IRE), Kazakhstan (KAZ), Kenya (KEN), Kyrgyzstan (KYR),
Mauritania (MAU), Mongolia (MON), Nigeria (NIG), Oman (OMA), Portugal
(PTG), Saudi Arabia (SAU), Senegal (SEN), Somalia (SOM), Sudan (SUD), Syria
(SYR), Tibet (TIB), Turkmenistan (TKM), Turkey (TUK), Tunisia (TUN), Yemen
(YEM), Croatia (YUC), Macedonia (YUM).
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
A Treemix phylogenetic reconstruction that
grouped modern donkeys according to sampl-
ing locations (27) confirms the earliest split
between African (Clade A) and mostly non-
African donkeys (Clade B) (Fig. 1C). Further
structure within Clade A separates donkeys
from the Horn of Africa (Ethiopia and Somalia)
plus Kenya from those from Western Africa
(Ghana, Mauritania, Nigeria, and Senegal).
Within Clade B, we find another major diver-
gence between European and Asian donkeys,
with east-to-west affinities in Europe from
the Balkans (Croatia and Macedonia) to Iberia,
Denmark, and Ireland. Conversely, Asian sub-
populations show west-to-east substructur-
ing from Iran and Central Asia to China and
Mongolia. Combined, these findings suggest
expansions from a central source into both
continents.
In Clade B, some of the most basal donkeys
are from the Southern Arabian Peninsula
(Om an and Yemen), whereas the single don-
key from Saudi Arabia analyzed here shows
European affinities, which is indicative of a
secondary translocation. Similarly, the Pega
donkey from Brazil is nested within Iberia,
which mirrors the colonization history of the
Americas. Clade B also includes donkeys
from Nubia (Egypt and Sudan) that show
affinities to the Levant (Syria) and Anatolia
(Turkey), as well as donkeys from Maghreb
(Tunisia), with closer genetic proximity to
European subpopulations. This suggests gene
flow into Africa from donkeys that were native
to Anatolia and the Levant, but not to the
Arabian Peninsula. Overall, this phy logenetic
reconstruction is compatible with both mod-
els of donkey domestication: a unique origin
in Africa followed by dispersals out and back,
or dual origins in Africa and the Southern
Arabian Peninsula.
The unique origin model posits a demo-
graphic expansion in Africa first and subse-
quent waves into Europe and Asia. By contrast,
dual origins would result in an earlier split of
demographic trajectories between African and
Eurasian subpopulations, given their deep phy-
logenetic divergence. To test these, we first per-
formed demographic modeling using SMC++
(28), which revealed the first expansion around
5200 BCE (7186 ± 742 years ago), in line with
archaeological evidence of domestication oc-
curring at this time (Fig. 1D and fig. S6). Addi-
tionally, when modeled from a possible African
source, SMC++ trajectories indicated more re-
cent and nearly coincidental expansions into
Asia around 2600 BCE (4573 ± 577 years ago)
and Europe around 2800 BCE (4806 ± 671 years
ago) (Fig. 1D). This is in line with the unique
origin model and the earliest archaeological
evidence of donkeys in Asia (Iranian Plateau
and the Indus Valley) and Europe (Portugal,
Greece, and Cyprus) in the mid-to-late third
millennia BCE (29–34).
Furthermore, the Yemen and Oman sub-
populations do not branch basal to Clade B
according to fineSTRUCTURE (35), in contrast
to the expectations of the dual origins model,
but within Asian subpopulations (Fig. 2, A and
B). Lastly, the pairwise genetic distances be-
tween Ethiopia and non-African subpopula-
tions were greater than those from Yemen (fig.
S5). They both increased linearly with geo-
graphic distances and supported identical
dispersal rates (fig. S5; p= 0.775), which is in
line with a single wave of expansion at a con-
stant pace. Therefore, our analyses support
an early domestication in Africa that spread
at an even rate into the Arabian Peninsula
and Eurasia and flowed back into Nubia and
Maghreb. Modern subpopulations from the
Horn of Africa and Kenya so far best represent
the descendants of earliest donkeys.
Ancient donkey genomes reveal early and
rapid dispersal into Asia and secondary
contacts between Europe and Western Africa
The patterns of genetic variation within mod-
ern subpopulations may reflect recent breed-
ing history rather than early domestication
(36). Additionally, they could underrepresent
the contribution of lineages that were once
important but have since declined (37). Dating
population splits also assumes constant yet
unknown generation intervals. To address
these caveats and validate the domestica-
tion hi story reconstructed above, we gener-
ated a genomic time series spanning the past
~4000 years that included 31 ancient don-
keys from 11 different sites, which ranged
from the Atlantic shores (Portugal) to Central
Asia (Eastern Iran/Turkmenistan) (Fig. 3A and
table S2).
Ancient genomes sequenced to 0.77- to
5.05-fold coverage (table S6) were analyzed by
using two complementary methods: pseudo-
haploidization following (21) that resulted in
4,833,570 nucleotide transversion sites and
genotype imputation following (38), at 7,161,029
polymorphic sites present at >5% frequency in
modern donkeys (38,39). Imputation accuracy
was confirmed by the high consistency rates
between imputed and observed genotypes
after down-sampling of high-coverage modern
genomes and downstream analyses that were
largely consistent with those that were based
on pseudo-haploidized data (figs. S7 to S15
and table S6).
The three oldest samples from our dataset
consist of donkeys from Anatolia (Acemhöyük,
Turkey), which were radiocarbon dated to
2564 to 2039 BCE. Their age and phylogenetic
placement within Clade B confirm an early
expansion out of Africa by ~2500 BCE, which
is in agreement with SMC++ time estimates
(Fig. 2A and fig. S10). These samples, as well
as a donkey from Eastern Iran/Turkmenistan
that is affiliated to the Bactria–Margiana Ar-
chaeological Complex (BMAC) (~2050 BCE;
sample Chalow3), branch before the forma-
tion of modern subpopulations from Central
Asia (Kazakhstan, Kyrgyzstan, Turkmenistan)
and Eastern Asia (China, Mongolia, Tibet)
(Fig. 2C). These subpopulations thus di-
verged after 2050 BCE, but potentially be-
fore the radiocarbon age of the donkey from
Doshan Tepe (1049 to 928 BCE), which appears
closer to modern subpopulations from Cen-
tral Asia in one Treemix analysis (Fig. 2D and
fig. S10).
Ancient samples from Iran (Shahr-i-Qumis,
800 BCE to 800 CE), including one Sassanid
(sample AM805), are not more closely related
to Central than to Eastern Asian modern sub-
populations, although their exact phylogenetic
placement remains unclear (Fig. 2, A and B,
and fig. S10). Their fineSTRUCTURE affinities
to modern Iran, Anatolia (Turkey), the Levant
(Syria), and Maghreb (Tunisia) support dif-
ferent genetic ancestry profiles from those
inferred at the nearby site of Doshan Tepe.
This indicates a population turnover in Iran
after ~1000 BCE but before ~500 CE, corre-
sponding to the radiocarbon time interval
of Doshan Tepe and a single specimen from
Shahr-i-Qumis.
Notably, all our ancient specimens from
Europe cluster within modern European do-
mesticates, supporting differentiation within
this continent before the oldest European
samples analyzed (Tarquinia, 803 to 412 BCE,
~2500 years ago; Fig. 3C). However, a donkey
from a Roman context in Marseille, a major
seaport trading center in Southern France
(Centre Bourse Marseille, 0 to 500 CE), dis-
played strong genetic affinities with modern
individuals from Western Africa (Fig. 3, B and
D). Additionally, SNP and haplotype shared-
ness with modern Western Africa were also
found in European donkeys from the Islamic
era in Portugal (Albufeira, 1228 to 1280 CE) and
Roman times in Northern France (Boinville-
en-Woëvre, 200 to 500 CE) (Fig. 3, B and E).
This reveals multiple contacts between Europe
and Western Africa from the Antiquity to
Middle Ages. Despite ancient European don-
keys showing Western African ancestry,
these contacts have affected Western Africa
more than Europe, in line wit h Tre emix
inferring gene flow predominantly in this
direction rather than the reverse (Fig. 1C).
All modern Irish donkeys and the two
Etruscan samples from Tarquinia are devoid
of Western African ancestry. This suggests the
preservation of old European genetic lineages,
at least in some modern subpopulations of
this continent.
Donkey management involved inbreeding and
introgression from divergent lineages
Inbreeding is a common reproductive strategy
for breeding animals with desirable traits (40).
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 3of9
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
To assess whether ancient donkey breeders
made use of inbreeding, we measured the pro-
portion of autosomal runs of homozygosity
(ROH) using three independent techniques,
all of which provided consistent results (figs.
S13 and S14). We detected inbreeding, but no
significant changes in levels between modern
and ancient donkeys (Wilcoxon rank sum test,
p=0.395)(Fig.4,A,B,andC).Conversely,
modern horses show higher inbreeding lev-
els than their ancient counterparts (Wilcoxon
rank sum test, p< 0.001), which mirrors pre-
vious reports of reduced heterozygosity and
increased deleterious mutation load in re-
cent times (Fig. 4, D, E, and F, table S11)
(21,41). Longer ROH tracts are more common
in modern horses and donkeys than in the
past, which is consistent with inbreeding
from closer generations in their genealogies
(Fig. 4, C and F). Overall, our analyses sup-
port recent major changes in reproductive
management inflating inbreeding in horses,
but not in donkeys.
Admixture modeling suggests ongoing intro-
gression from African wild asses into modern
donkeys from Africa and the Southern Arabian
Peninsula (with between 0.24 to 6.99% of ad-
mixture) (Fig. 1A, fig. S4, and table S5). This
is in line with free-ranging local management
practices that allow for continued interbreed-
ing with wild and feral subpopulations (4,42).
The limited but substantial amount of wild
genetic material from kiangs in one modern
donkey from China also supports admixture
between taxa generally regarded as separate
species. This confirms previous reports of
mitochondrial introgression (43)andgenomic
admixture despite different karyotypes (24).
Notably, all but one ancient donkey (sample
Tur168) carried remnants of outgroup mate-
rial (0.21 to 4.15%; Fig. 3B), which poten-
tially results from recent range contractions
of wild subpopulations and ancient manage-
ment practices providing more opportunities
for wild introgression.
The genome of MV242, a donkey from Israel
dating to the Hellenistic period (350 to 58 BCE),
displayed the largest fraction of divergent
genetic material (Fig. 3B, 4.15% ± 0.019). In
Treemix, this sample showed a deeper place-
ment than all donkeys present in our panel,
except the Somali wild ass [Equus africanus
somaliensis (E.a.som)] (Fig. 2E). Significantly
positive f4 (E.a.som, MV242; Horn+Ken, x)
statistics revealed MV242-related genetic an-
cestry in some modern subpopulations (x), espe-
cially toward Central and Eastern Asia (Fig. 5E).
This ancestry was already present in the BMAC
sample from Iran (sample Chalow3, Fig. 5F),
which indicates contact ~2050 BCE at the
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 4of9
Fig. 2. Haplotype sharedness and phylogenetic placement of ancient
European donkeys. (A) Haplotype sharedness clustering of modern (n= 168) and
ancient donkeys (n= 31) reconstructed using fineSTRUCTURE (35). Modern
domesticates are colored following Fig. 1, and ancient individuals are numbered
according to Fig. 3A. Cluster supports are shown in as probabilities on each node
if >0.8. MV242 placement is incongruent with Treemix (Fig. 2E) because of the
limited representation of divergent ancestries in the modern reference panel
used for imputation. (B) Co-ancestry matrix based on haplotype sharedness.
Co-ancestry values are averaged for co-clustered individuals. (Cto E) Treemix
phylogenies of three ancient specimens shown in black (C: Chalow3, D: Doshan
Tepe, and E: MV242) placed within the subpopulations defined in Fig. 1C (27).
Branches that are not scaled are shown as dashed lines. s.e., standard error.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
latest. It was, however, absent in Acemhöyük
at that time, which suggests that the MV242
divergent lineage ranged into Eastern Iran/
Turkmenistan, but not Turkey. This lineage
also left genetic material in modern Anatolia,
the Levant, Nubia, andMaghreb,butnotin
Western Africa, which is consistent with don-
keys carrying MV242-related ancestry flowing
back into some African regions. In addition,
this ancestry was also present in Southwest-
ern European subpopulations (Canary Islands,
Portugal, and Spain), but neither in the mod-
ern Balkans and Ireland nor in any ancient
European sample analyzed here (Fig. 5, E and
F). Combined, our results suggest a range for
the MV242-related lineage from the Levant
into Asia rather than Europe and Africa.
Despite its divergent genetic makeup, MV242
carries a mitochondrial haplotype character-
istic of Clade II (Fig. 5A). Our tip-calibrated
coalescent analyses revealed that the time to
the most common recent ancestor of that
Clade was 32,226 BCE and not 332,580 to
142,980 BCE (Fig. 5B), as previously reported
(12,44). Because the same holds true for
Clade I, both clades could have coexisted in
sympatry 25,000 years later when donkeys
were first domesticated (Fig. 5B). Additionally,
no phylogeographic structure is apparent in
patterns of mitochondrial variation, both in
modern and ancient subpopulations, as ancient
specimens from Asia and Europe, sometimes
from the same archaeological sites, were placed
across both clades (Fig. 5A). Y-chromosomal
variation was also associated with little, if any,
population structure (Fig. 5, C and D). Com-
bined, our results dismiss mitochondrial DNA
and the Y-chromosome as reliable markers of
domestication history in donkeys.
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 5of9
Fig. 3. Ancient donkey dataset, genetic affinities to outgroups, and
modern donkeys. (A) Geographical distribution, estimated age, and sample
names of ancient donkeys (n= 31). Pie charts represent the proportion
of individuals with dun coat color (white), heterozygotes (gray), and derived
coat color (red) at each site. Genotype probabilities ≥0.99 are denoted
with ** and ≥0.9 with *. (B) Heatmap displaying outgroup f3-statistics in the
form of (modern, ancient; kiang) (58). Bar charts represent the proportion
of wild ancestry (kiang, onager, zebra, E.a.som) in each ancient individual
with standard errors estimated from ADMIXTURE with 100 bootstrap pseudo-
replicates (56). (Cto E) Treemix phylogenies of ancient specimens from
three archaeological sites shown in black [C: Tarquinia (Tarquinia214,
Tarquinia501), D: Bourse (BourseB, BourseC), E: Albufeira 1x1] placed within
the subpopulations defined in Fig. 1C (27). Branches that are not scaled are
shown as dashed lines. s.e., standard error.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
Romans bred improved donkeys for producing
mules essential for their military power
and economy
Beyond documenting domestication history
at the global scale, our genomic dataset also
included three jennies (females) and six jacks
(males) from the same archaeological site
(Boinville-en-Woëvre) (Fig. 3A). These were
found in a dedicated farming area of a Roman
villa, providing insights into local management
practices in Roman Northern France (200 to
500 CE). One jack (sample GVA349) appeared
particularly inbred, with long ROH indicative
of recent consanguinity (Fig. 4A), and was ge-
netically related to four jacks and one jenny
(family group GVA1, including samples GVA125,
GVA347, GVA348, GVA349, GVA353, and
GVA354; table S10). Additionally, two jennies
showed genetic relatedness coefficients equiv-
alent to full siblings (family group GVA2, in-
cluding samples GVA355 and GVA358; table
S10). This indicates breeding management
within close kin, potentially aimed at select-
ing for desirable traits. Genotype imputation
at TBX3 (13) revealed the presence of dun and
derived colored coats, but no evidence for the
dominant alleles associated with white spots
or long hair was found in the sequence align-
ments at KIT (45)andFGF5 (fig. S11 and tables
S7 to S9) (46). The latter two phenotypes are,
however, relatively common in modern breeds
from France, which suggests post-Roman se-
lection for these traits.
The abundance of donkeys at Boinville-en-
Woëvre stands as an exception in Roman
France, as mules dominated all other assemb-
lagesfromthistime(47). Contemporaneous
Roman sites report mules of a large and uni-
form size, indicating selective breeding in
the parental species for expensive animals of
exceptional stature [Varro (2, 6)] (48). Mor-
phometric m easurements previously revealed
five donkeys from family group GVA1 as giant
(148 to 156 cm at the withers) (47). We found
that sample GVA359 had a similarly large size
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 6of9
Fig. 4. Inbreeding in domestic donkeys and horses. (A) Distribution of total ROH length in modern versus ancient donkeys. (B) Total length of ROH in donkey
genomes through time. (Cand D) Same as (A) and (B), but for 79 modern and 75 ancient horses. ROH tracts were identified using the program ngsF-HMM (59).
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
(144 cm) and genetic affinities to Western
Africa. This may indicate restocking to enhance
body size from distant bloodlines carrying
divergent ancestry or from wild populations.
Outgroup admixture was significantly higher
at Boinville-en-Woëvre than in other ancient
donkeys except the divergent MV242 specimen
(p= 0.045). Significantly negative f4(kiang,
MV242; Fiumarella1, Boinville-en-Woëvre)
statistics support restocking into family group
GVA1 only, from a lineage more divergent
than MV242 (Fig. 5G). Additionally, f4(kiang,
E.a.som; Fiumarella1, GVA1) statistics reject
unbalanced allele sharedness between E.a.som
andGVA1,rulingoutrestockingfromE.a.som
or more divergent populations (Fig. 5H). Com-
bined, these findings uncover a lineage, phylo-
genetically intermediate between MV242 and
E.a.som, that contributed to the genetic makeup
of some Roman donkeys at Boinville-en-Woëvre.
Together with the evidence of genetic related-
ness and inbreeding, this suggests Boinville-en-
Woëvre as a likely mule production center that
maintained the bloodlines of giant donkeys
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 7of9
Fig. 5. Uniparental marker phylogenies and introgression of divergent
lineages. (A) Mitochondrial phylogeny constructed using IQ-TREE (60)with100
bootstrap pseudo-replicates marked with a black triangle if >90%. (B) Posterior
distributions of the time to the most recent common ancestors of all mitochondrial
haplotypes, Clade I and Clade II labeled with their modes. (Cand D)Sameas(A)
and (B) for the Y-chromosome. (Eto H)f4-statistics(58)exploringthegenetic
contribution of divergent lineages into modern and ancient donkeys. Z scores were
corrected for multiple testing, and red bars with asterisks show p<0.05.
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
that were selected through familial breeding
and restocking. This center may illustrate how
Romans sustained the eno rmous demand for
mules, which is documented in the nearby
Rhine frontier(49), and fueled transportation
networks throughout the Empire (47).
Discussion
Our study solves long-standing debates about
donkey domestication. We support domesti-
cation starting from a unique African source
~5000 BCE. Donkeys subsequently spread
into Eurasia from ~2500 BCE, and Central and
Eastern Asian subpopulations differentiated
~2000 to 1000 BCE. Genetic affinities char-
acteristic of modern Western Europe were
already formed by 500 BCE. After early do-
mestication, African donkeys further differ-
entiated in the West and the Horn of Africa
plus Kenya but also received streams of ge-
netic ancestry from Western Europe as well as
a region encompassing the Levant, Anatolia,
and Mesopotamia. Donkey domestication in-
volved limited but detectable wild introgres-
sion. It did not entail inflated inbreeding in
recent times, in contrast to horses. The pro-
cesses of donkey and horse domestication
differed substantially, as horses were domes-
ticated twice (50) and rapidly spread across
Eurasia from the lower Don-Volga region
~2000 BCE (21). Their regional differentia-
tion remained relatively limited owing to
strong connectivity at continental distances
early on and until oriental bloodlines were
propagated throughout the world during the
past 1000 years (41,51). The extent to which
the different domestication trajectories of
donkeys and horses were only driven by their
respective roles in human societies or also re-
flected management practices adapted to their
respective mating and social behavior (52)re-
mains to be explored.
This work clarifies global patterns of donkey
domestication and movements but also high-
lights many directions for future research. For
example, it remains unknown whether domes-
tic donkeys only dispersed out of Africa by
land through the Sinai Peninsula or across
the Red Sea from Ethiopia to Yemen. Addi-
tionally, modern subpopulations from the Horn
of Africa plusKenya were found to be the first
expanding. This may suggest early domesti-
cation there or that donkeys domesticated
elsewhere in Africa entering the region more
recently. Further research is needed to clarify
the timing of pastoral spread into the Red Sea
Sudanese region and the Horn of Africa. Cur-
rent dates range from ~2500 BCE in Ethiopia
and Eritrea (53)to~3000BCEinNorthern
Kenya (54). Donkeys are not present in the
archaeological record of Western Africa before
the beginning of the common era, either (55),
which postdates by 3000 years the time when
donkey populations from the Horn of Africa
plus Kenya and Western Africa are inferred
to have split genetically. This may indicate an
early yet undocumented arrival in the region
or a slow migration westward, only reaching
the modern range later. Improving the current
African archaeological record thus appears
paramount to refining the exact context un-
derlying early donkey domestication and sub-
sequent population movements.
Further genomic studies in other regions
would also largely benefit the understanding
of donkey diversity and history. Resolving the
genetic structure of equine remains from the
third millennium BCE of Southwest Asia will
be challenging because of postmortem DNA
decay but essential to mapping the geographic
range of the divergent lineage identified here
(sample MV242), as well as to understanding
dispersal mechanisms in greater detail. The
same holds true for Chalcolithic and Bronze
Age Europe, which remain genetically undocu-
mented in our dataset, and onwards. Develop-
ing genetic knowledge of ancient European
donkeys will further clarify patterns of exchange
across the Mediterranean region, including
during and after Roman times, as revealed in
this study. It will also provide insights into
the dispersal mechanisms underpinning the
genetically supported presence of donkey
remains in Portugal ~2200 BCE (33). Genetic
characterization of local archaeological sites at
the population scale may uncover additional
mule breeding centers other than the one re-
ported here. This will shed light on the diver-
sity of breeding management strategies that
were developed by Romans to supply their
continental-wide economy and military with
adequate animal resources (49). For now, both
the absence of mules and the rarity of horse
mares at Boinville-en-Woëvre (47)suggestthat
mares were brought in for mating before re-
turning pregnant to their owners. Alternatively,
donkey breeders may have visited other farms
with their jacks to cover mares.
Efforts should continue to characterize the
modern donkey diversity around the world,
especially in Saudi Arabia, which is currently
characterized by a single individual, as well
as in Africa, for which no populations located
south of the Equator have been sampled. Such
efforts may not only refine the historical leg-
acy of past populations into the modern world
but also uncover the genetic basis of desert
adaptations, which could prove invaluable for
future donkey breeding in the face of global
warming.
REFERENCES AND NOTES
1. P. Mitchell, The Donkey in Human History: An Archaeological
Perspective (Oxford Univ. Press, 2018).
2. S. L. Norris, H. A. Little, J. Ryding, Z. Raw, PLOS ONE 16,
e0247830 (2021).
3. B. Kimura, F. Marshall, A. Beja-Pereira, C. Mulligan,
Afr. Archaeol. Rev. 30,83–95 (2013).
4. F. Marshall, Rethinking Agriculture: Archeological and
Ethnoarcheological Perspectives (Left Coast Press, 2007).
5. S. Bokonyi, in Studi di paletnologia in onore di Salvatore M.
Puglisi, M. Liverani, A. Palmieri, R. Peroni, Eds. (Università di
Roma La Sapienza, 1985), pp. 495–499.
6. J. Boesneck, A. Von den Driesch, in El Omari: A Neolithic
Settlement and Other Sites in the Vicinity of Wadi Hof, Helwan,
F. Debono, B. Mortensen, Eds. (Philipp von Zabern, 1990),
pp. 99–107.
7. J. Boesneck, A. von den Driesch, R. Ziegler, in Maadi III: The
Non-lithic Small Finds and the Structural Remains of
the Predynastic Settlement, I. Rizkana, J. Seeber, Eds.
(Philipp von Zabern, 1989), pp. 87–128.
8. K. M. Cialowicz, Les palettes Egyptiennes aux motifs
zoomorphes et sans decoration. Etudes de l'art predynastique
(Jagiellonian University, 1991).
9. D. J. Brewer, D. B. Redford, S. Redford, Domestic Plants and
Animals: The Egyptian Origins (Aries and Phillips, 1994).
10. S. Rossel et al., Proc. Natl. Acad. Sci. U.S.A. 105, 3715–3720
(2008).
11. A. Beja-Pereira et al., Science 304, 1781 (2004).
12. B. Kimura et al., Proc. Biol. Sci. 278,50–57 (2011).
13. C. Wang et al., Nat. Commun. 11, 6014 (2020).
14. M. Cattani, S. Bokonyi, in Essays on the Late Prehistory of the
Arabian Peninsula. Serie Orientale Romana XCIII, S. Cleuziou,
T. Maurizio, Z. Juris, Eds. (Istituto Italiano per l'Africa e l'Oriente,
2002).
15. R. H. Meadow, H. P. Uerpmann, Equids in the Ancient World,
vol. 2 (Dr Ludwig Reichert Verlag, 1991).
16. J. Clutton-Brock, Mammal Rev. 22,79–85 (1992).
17. J. Zarins, R. Hauser, The domestication of equidae in
third-millennium BCE Mesopotamia. Cornell University Studies
in Assyriology and Sumerology 24, XI + 432 (2014).
18. E. Vila, in Equids in Time and Space: Papers in Honour of Véra
Eisenmann, M. Mashkour, Ed. (Oxbow, 2006), pp. 101–123.
19. J. Boessneck, A. von den Driesch, U. Steger, Baghdader
Mitteilingen 15,149–189 (1984).
20. E. A. Bennett et al., Sci. Adv. 8, eabm0218 (2022).
21. P. Librado et al., Nature 598, 634–640 (2021).
22. S. Rosenbom et al., Mol. Phylogenet. Evol. 85,88–96 (2015).
23. G. Renaud et al., Sci. Adv. 4, eaaq0392 (2018).
24. H. Jónsson et al., Proc. Natl. Acad. Sci. U.S.A. 111,
18655–18660 (2014).
25. L. Zeng et al., Zool. Res. 42, 450–460 (2021).
26. A. L. Price et al., Nat. Genet. 38, 904–909 (2006).
27. J. K. Pickrell, J. K. Pritchard, PLOS Genet. 8, e1002967 (2012).
28. J. Terhorst, J. A. Kamm, Y. S. Song, Nat. Genet. 49, 303–309
(2017).
29. D. Reese, Mediterr. Archaeol. Archaeom. Int. J. 13, 289–320
(2013).
30. B. A. Knapp, Biblic. Archaeol. 55,52–72 (1992).
31. S. Ratnagar, Trading Encounters: from the Euphrates to the
Indus in the Bronze Age (Oxford Univ. Press, 2004).
32. M. A. Zeder, in Equids in the Ancient World, R. H. Meadow,
H. P. Uerpmann, Eds. (Reichert, 1986), pp. 164–193.
33. J. L. Cardoso, J. T. Vilstrup, V. Eisenmann, L. Orlando,
J. Archaeol. Sci. 40, 4483–4490 (2013).
34. S. Amiri, M. Mashkour, F. Mohaseb, R. Naseri, in Archaeozoology
of Southwest Asia and Adjacent Areas XIII: Proceedings of the
Thirteenth International Symposium, University of Cyprus,
Nicosia, Cyprus, June 7-10, 2017, J. Daujat, A. Hadjikoumis,
R. Berthon, J. Chahoud, V. Kassianidou, J. D. Vigne, Eds. (2021),
pp. 199–219.
35. D. J. Lawson, G. Hellenthal, S. Myers, D. Falush, PLOS Genet. 8,
e1002453 (2012).
36. L. Girdland Flink et al., Proc. Natl. Acad. Sci. U.S.A. 111,
6184–6189 (2014).
37. L. A. F. Frantz, D. G. Bradley, G. Larson, L. Orlando,
Nat. Rev. Genet. 21, 449–460 (2020).
38. R. Hui, E. D’Atanasio, L. M. Cassidy, C. L. Scheib, T. Kivisild,
Sci. Rep. 10, 18542 (2020).
39. B. L. Browning, S. R. Browning, Am. J. Hum. Genet. 98,
116–126 (2016).
40. T. N. Kristensen, A. C. Sørensen, Anim. Sci. 80, 121–133 (2007).
41. A. Fages et al., Cell 177, 1419–1435.e31 (2019).
42. F. Marshall, L. Weissbrod, in Tracking Down the Past.
Ethnohistory Meets Archaeozoology, G. Grupe, J. Peters,
G. McGlynn, Eds. (M. Leidor, 2009),
pp. 59–79.
43. E. A. Bennett et al., PLOS ONE 12, e0174216 (2017).
44. L. Wang et al., Front. Genet. 12, 759831 (2021).
45. B. Haase, S. Rieder, T. Leeb, Anim. Genet. 46, 321–324 (2015).
46. R. Legrand, L. Tiret, M. Abitbol, Genet. Sel. Evol. 46,65
(2014).
47. S. Lepetz et al., J. Archaeol. Sci. Rep. 40, 103250 (2021).
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 8of9
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
48. G. K. Kunst, Anthropozoologica 31, 109–118 (2000).
49. C. Johnstone, in Feeding the Roman Army: The Archaeology of
Production and Supply in NW Europe, S. Stalibrass, R. Thomas,
Eds. (Oxbow Books, 2008), pp. 128–145.
50. C. Gaunitz et al., Science 360, 111–114 (2018).
51. S. Felkel et al., Sci. Rep. 9, 6095 (2019).
52. F. B. Marshall, K. Dobney, T. Denham, J. M. Capriles, Proc. Natl.
Acad. Sci. U.S.A. 111, 6153–6158 (2014).
53. J. Lesur, E. A. Hildebrand, G. Abawa, X. Gutherz, Quat. Int. 343,
148–158 (2014).
54. E. A. Hildebrand et al., Proc. Natl. Acad. Sci. U.S.A. 115,
8942–8947 (2018).
55. K. MacDonald, R. Hutton MacDonald, in The Origins and
Development of African livestock: Archaeology, Genetics,
Linguistics and Ethnography, R. M. Blench, K. MacDonald,
Eds. (University College London Press, 2000), pp. 8–9.
56. D. H. Alexander, K. Lange, BMC Bioinformatics 12, 246
(2011).
57. N. Patterson, A. L. Price, D. Reich, PLOS Genet. 2, e190 (2006).
58. N. Patterson et al., Genetics 192, 1065–1093 (2012).
59. F. G. Vieira, A. Albrechtsen, R. Nielsen, Bioinformatics 32,
2096–2102 (2016).
60. B. Q. Minh et al., Mol. Biol. Evol. 37, 1530–1534 (2020).
ACKNOWL EDGME NTS
Funding: L.K.H. thanks the director of the Tel es-Safi/Gath
excavation, A. Maeir, for facilitating sampling. The DonkeyBank
collection of modern donkey DNA samples is supported by the
European Union’s Horizon 2020 Research and Innovation
Programme (grant agreement no. 857251). L.P.P. is funded by
national funds from FCT –Fundação para a Ciência e a Tecnologia,
I.P. Genoscope and GeT-PlaGe sequencing platforms are partly
funded by France Génomique National infrastructure, funded
as part of “Investissement d’avenir”program managed by
Agence Nationale pour la Recherche (contract ANR-10-INBS-09),
and their “Grands Projets”funding instrument (BUCEPHALE). This
project has received funding from the CNRS, University
Paul Sabatier (AnimalFarm IRP), and the European Research Council
(ERC) under the European Union’s Horizon 2020 research and
innovation program (grant agreements 885729-AncestralWeave,
295729-CodeX, 853272-PALAEOFARM, and 681605-PEGASUS).
Author contributions: Conceptualization: L.O.; Materials and
Reagents: J.-M.A., A.P., P.W., O.B., C.D., and L.O.; Archaeological
samples and contextual information: M.V., E.V., M.Marzullo, O.P.,
U.T., G.B.G., A.T., V.T., F.A., J.L.C., M.J.V., M.T.A., N.B., A.G., L.K.H.,
A.Ö., B.S.A., V.O., B.C., S.L., A.A.V., H.D., A.M., M.Mashkour, and
L.O.; Sampled modern donkeys: A.B.-P., D.-D.W., L.P.R., and S.A.B.;
Radiocarbon dating: J.S.; DNA sequencing and preparation: L.T.C.,
L.C., S.S., A.F., A.S.-O., P.C., N.K., L.P.P., J.M.A., A.P., O.B., C.D.,
P.W., A.B.-P., and L.O.; Data analysis: E.T.T. and L.O. with input from
P.L. and H.R.; Writing—Supplementary Information: E.T.T., with
input from L.O.; Writing—Main Article: E.T.T. and L.O. with input
from A.F., L.P.R., P.L., E.V., F.A., L.A.F., B.S., L.K.H., B.S.A.,
M.Mashkour, S.A.B., A.B.P., and all coauthors. Competing interests:
The authors declare that they have no competing interests. Data
and materials availability: The sequence data generated in this
study are availablefor download at the European Nucleotide Archive
(accession no. PRJEB52849). The variant data for this study
have been deposited in the European Variation Archive (EVA) at
EMBL-EBI under accession number PRJEB55549. The recombination
map can be downloaded from https://osf.io/k7x8b/. The accession
numbers for each individual sample and all other data used
in this study are included in tables S1, S2, and S11 of the
supplementary materials. License information: Copyright © 2022
the authors, some rights reserved; exclusive licensee American
Association for the Advancement of Science. No claim to original
US government works. https://www.science.org/about/science-
licenses-journal-article-reuse
SUPPLEMENTARY MATERIALS
science.org/doi/10.1126/science.abo3503
Materials and Methods
Figs. S1 to S15
Tables S1 to S11
References (61–136)
MDAR Reproducibility Checklist
View/request a protocol for this paper from Bio-protocol.
Submitted 28 January 2022; accepted 15 August 2022
10.1126/science.abo3503
Todd et al., Science 377, 1172–1180 (2022) 9 September 2022 9of9
RESEARCH |RESEARCH ARTICLE
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022
Use of this article is subject to the Terms of service
Science (ISSN ) 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 © 2022 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim
to original U.S. Government Works
The genomic history and global expansion of domestic donkeys
Evelyn T. ToddLaure Tonasso-CalvièreLoreleï ChauveyStéphanie SchiavinatoAntoine FagesAndaine Seguin-
OrlandoPierre ClavelNaveed KhanLucía Pérez PardalLaura Patterson RosaPablo LibradoHarald RingbauerMarta
VerdugoJohn SouthonJean-Marc AuryAude PerdereauEmmanuelle VilaMatilde MarzulloOrnella PratoUmberto
TecchiatiGiovanna Bagnasco GianniAntonio TagliacozzoVincenzo TinèFrancesca AlhaiqueJoão Luís CardosoMaria
João ValenteMiguel Telles AntunesLaurent FrantzBeth ShapiroDaniel G. BradleyNicolas BoulbesArmelle GardeisenLiora
Kolska HorwitzAliye ÖztanBenjamin S. ArbuckleVedat OnarBenoît ClavelSébastien LepetzAli Akbar VahdatiHossein
DavoudiAzadeh MohasebMarjan MashkourOlivier BouchezCécile DonnadieuPatrick WinckerSamantha A. BrooksAlbano
Beja-PereiraDong-Dong WuLudovic Orlando
Science, 377 (6611), • DOI: 10.1126/science.abo3503
Donkeys’ African origins
Donkeys have been important to humans for thousands of years, being the primary source of work and transport for
many cultures. Unlike horses, little was known about the origin and domestication of donkeys. Todd et al. sequenced
the genomes of modern and ancient donkeys and found evidence of an eastern African origin over 7000 years
ago, with subsequent isolation and divergence of lineages in Africa and Eurasia. They also reveal the imprint of
desertification on divergence among groups and specifics about donkey breeding and husbandry, including selection
for large size and the practice of inbreeding. —SNV
View the article online
https://www.science.org/doi/10.1126/science.abo3503
Permissions
https://www.science.org/help/reprints-and-permissions
Downloaded from https://www.science.org at bibCNRS INSB on September 10, 2022