ARTICLE

Genetic history from the Middle Neolithic topresent on the Mediterranean island of SardiniaJoseph H. Marcus et al.#

The island of Sardinia has been of particular interest to geneticists for decades. The current

model for Sardinia’s genetic history describes the island as harboring a founder population

that was established largely from the Neolithic peoples of southern Europe and remained

isolated from later Bronze Age expansions on the mainland. To evaluate this model, we

generate genome-wide ancient DNA data for 70 individuals from 21 Sardinian archaeological

sites spanning the Middle Neolithic through the Medieval period. The earliest individuals

show a strong affinity to western Mediterranean Neolithic populations, followed by an

extended period of genetic continuity on the island through the Nuragic period (second

millennium BCE). Beginning with individuals from Phoenician/Punic sites (first millennium

BCE), we observe spatially-varying signals of admixture with sources principally from the

eastern and northern Mediterranean. Overall, our analysis sheds light on the genetic history

of Sardinia, revealing how relationships to mainland populations shifted over time.

https://doi.org/10.1038/s41467-020-14523-6 OPEN

*A full list of authors and their affiliations appears at the end of the paper.

NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 |www.nature.com/naturecommunications 1

1234

5678

90():,;

http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-020-14523-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-020-14523-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-020-14523-6&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-020-14523-6&domain=pdfhttp://orcid.org/0000-0002-0923-9881http://orcid.org/0000-0002-0923-9881http://orcid.org/0000-0002-0923-9881http://orcid.org/0000-0002-0923-9881http://orcid.org/0000-0002-0923-9881www.nature.com/naturecommunicationswww.nature.com/naturecommunications

The whole-genome sequencing in 2012 of “Ötzi”, an indi-vidual who was preserved in ice for over 5000 years nearthe Italo-Austrian border, revealed a surprisingly high levelof shared ancestry with present-day Sardinian individuals1,2.Subsequent work on genome-wide variation in ancient Europeansfound that most “early European farmer” individuals, even whenfrom geographically distant locales (e.g., from Sweden, Hungaryand Spain) have their highest genetic affinity with present-daySardinian individuals3–6. Accumulating ancient DNA (aDNA)results have provided a framework for understanding how earlyEuropean farmers show such genetic affinity to modernSardinians.

In this framework, Europe was first inhabited by Paleolithicand later Mesolithic hunter-gatherer groups. Then, starting about7000 BCE, farming peoples arrived from the Middle East as partof a Neolithic transition7, spreading through Anatolia and theBalkans8,9 while progressively admixing with local hunter-gatherers10. Major movements from the Eurasian Steppe, begin-ning about 3000 BCE, resulted in further admixture throughoutEurope11–14. These events are typically modeled in terms of threeancestry components: western hunter gatherers (“WHG”), earlyEuropean farmers (“EEF”), and Steppe pastoralists (“Steppe”).Within this broad framework, the island of Sardinia is thought tohave received a high level of EEF ancestry early on and thenremained mostly isolated from the subsequent admixture occur-ring on mainland Europe1,2. However, this specific model forSardinian population history has not been tested with genome-wide aDNA data from the island.

The oldest known human remains on Sardinia date to ~20,000years ago15. Archeological evidence suggests Sardinia was notdensely populated in the Mesolithic, and experienced a popula-tion expansion coinciding with the Neolithic transition in thesixth millennium BCE16. Around this time, early Neolithic pot-tery assemblages were spreading throughout the western Medi-terranean, including Sardinia, in particular vessels decorated withCardium shell impressions (variably described as ImpressedWare, Cardial Ware, Cardial Impressed Ware)17, with radio-carbon dates indicating a rapid westward maritime expansionaround 5500 BCE18. In the later Neolithic, obsidian originatingfrom Sardinia is found throughout many western Mediterraneanarcheological sites19, indicating that the island was integrated intoa maritime trade network. In the middle Bronze Age, about 1600BCE, the “Nuragic” culture emerged, named for the thousands ofdistinctive stone towers, known as nuraghi20. During the lateNuragic period, the archeological and historical record shows thedirect influence of several major Mediterranean groups, in par-ticular the presence of Mycenaean, Levantine and Cypriot traders.The Nuragic settlements declined throughout much of the islandas, in the late 9th and early 8th century BCE, Phoenicians ori-ginating from present-day Lebanon and northern Palestineestablished settlements concentrated along the southern shores ofSardinia21. In the second half of the 6th century BCE, the islandwas occupied by Carthaginians (also known as Punics), expand-ing from the city of Carthage on the North-African coast ofpresent-day Tunisia, which was founded in the late 9th century byPhoenicians22,23. Sardinia was occupied by Roman forces in 237BCE, and turned into a Roman province a decade later24.Throughout the Roman Imperial period, the island remainedclosely aligned with both Italy and central North Africa. After thefall of the Roman empire, Sardinia became increasingly autono-mous24, but interaction with the Byzantine Empire, the maritimerepublics of Genova and Pisa, the Catalan and Aragonese King-dom, and the Duchy of Savoy and Piemonte continued toinfluence the island25,26.

The population genetics of Sardinia has long been studied, inpart because of its importance for medical genetics27,28. Pioneering

studies found evidence that Sardinia is a genetic isolate withappreciable population substructure29–31. Recently, Chiang et al.32

analyzed whole-genome sequences33 together with continentalEuropean aDNA. Consistent with previous studies, they found themountainous Ogliastra region of central/eastern Sardinia carries asignature of relative isolation and subtly elevated levels of WHGand EEF ancestry.

Four previous studies have analyzed aDNA from Sardiniausing mitochondrial DNA. Ghirotto et al.34 found evidence formore genetic turnover in Gallura (a region in northern Sardiniawith cultural/linguistic connections to Corsica) than Ogliastra.Modi et al.35 sequenced mitogenomes of two Mesolithic indivi-duals and found support for a model of population replacementin the Neolithic. Olivieri et al.36 analyzed 21 ancient mitogenomesfrom Sardinia and estimated the coalescent times of Sardinian-specific mtDNA haplogroups, finding support for most of themoriginating in the Neolithic or later, but with a few coalescingearlier. Finally, Matisoo-Smith et al.37 analyzed mitogenomes in aPhoenician settlement on Sardinia and inferred continuity andexchange between the Phoenician population and broader Sar-dinia. One additional study recovered β-thalessemia variants inthree aDNA samples and found one carrier of the cod39 mutationin a necropolis used in the Punic and Roman periods38. Despitethe initial insights these studies reveal, none of them analyzegenome-wide autosomal data, which has proven to be useful forinferring population history39.

Here, we generate genome-wide data from the skeletal remainsof 70 Sardinian individuals radiocarbon dated to between 4100BCE and 1500 CE. We investigate three aspects of Sardinianpopulation history: First, the ancestry of individuals from theSardinian Neolithic (ca. 5700–3400 BCE)—who were the earlypeoples expanding onto the island at this time? Second, thegenetic structure through the Sardinian Chalcolithic (i.e., CopperAge, ca. 3400–2300 BCE) to the Sardinian Bronze Age (ca.2300–1000 BCE)—were there genetic turnover events through thedifferent cultural transitions observed in the archeological record?And third, the post-Bronze Age contacts with major Mediterra-nean civilizations and more recent Italian populations—have theyresulted in detectable gene flow?

Our results reveal insights about each of these three periods ofSardinian history. Specifically, our earliest samples show affinityto the early European farmer populations of the mainland, thenwe observe a period of relative isolation with no significant evi-dence of admixture through the Nuragic period, after which weobserve evidence for admixture with sources from the northernand eastern Mediterranean.

ResultsAncient DNA from Sardinia. We organized a collection ofskeletal remains (Supp. Fig. 1) from (1) a broad set of previouslyexcavated samples initially used for isotopic analysis40, (2) theLate Neolithic to Bronze Age Seulo cave sites of central Sardi-nia41, (3) the Neolithic Sites Noedalle and S’isterridolzu42, (4)the Phoenician-Punic sites of Monte Sirai23 and Villamar43,(5) the Imperial Roman period site at Monte Carru (Alghero)44,(6) medieval remains from the site of Corona Moltana45,(7) medieval remains from the necropolis of the Duomo of SanNicola46. We sequenced DNA libraries enriched for the completemitochondrial genome as well as a targeted set of 1.2 millionsingle nucleotide polymorphisms (SNPs)47. After quality control,we arrived at a final set of 70 individuals with an average cov-erage of 1.02× at targeted SNPs (ranging from 0.04× to 5.39× perindividual) and a median number of 466,049 targeted SNPscovered at least once per individual. We obtained age estimatesby either direct radiocarbon dating (n= 53), previously reported

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6

2 NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 | www.nature.com/naturecommunications

www.nature.com/naturecommunications

radiocarbon dates (n= 13), or archeological context and radio-carbon dates from the same burial site (n= 4). The estimatedages range from 4100 years BCE to 1500 years CE (Fig. 1, Supp.Data 1A). We pragmatically grouped the data into broad periods:Middle/Late Neolithic (‘Sar-MN’, 4100–3500 BCE, n= 6), EarlyCopper Age (‘Sar-ECA’, 3500–2500 BCE, n= 3), Early MiddleBronze Age (‘Sar-EMBA’, 2500–1500 BCE, n= 27), and Nuragic(‘Sar-Nur’, 1500–900 BCE, n= 16). For the post-Nuragic sites,there is substantial genetic heterogeneity within and among sites,and so we perform analysis per site when grouping is necessary(‘Sar-MSR’ and ‘Sar-VIL’ for the Phoenician and Punic sites ofMonte Sirai, n= 2; and Villamar, n= 6; ‘Sar-ORC002’ for aPunic period individual from the interior site of S’Orcu ’e Tueri,n= 1; ‘Sar-AMC’ for the Roman period site of Monte Carru nearAlghero, n= 3; ‘Sar-COR’ for the early medieval individualsfrom the site of Corona Moltana, n= 2; and ‘Sar-SNN’ for themedieval San Nicola necropoli, n= 4). Figure 1 provides anoverview of the sample.

To assess the relationship of the ancient Sardinian individuals toother ancient and present-day west Eurasian and North-Africanpopulations we analyzed our individuals alongside publishedautosomal DNA data (ancient: 972 individuals9,10,13,48–50; modern:1963 individuals from outside Sardinia7 and 1577 individualsfrom Sardinia32,33). For some analyses, we grouped the modernSardinian individuals into eight geographic regions (see inset inpanel c of Fig. 2 for listing and abbreviations, also see Supp.Data 1E) and for others we subset the more isolated Sardinianregion of Ogliastra (‘Sar-Ogl’, n= 419) and the remainder (‘Sar-non Ogl’, n= 1158). As with other human genetic variationstudies, population annotations are important to consider in theinterpretation of results.

Similarity to western mainland Neolithic populations. Wefound low differentiation between Middle/Late Neolithic Sardi-nian individuals and Neolithic western mainland Europeanpopulations, in particular groups from Spain (Iberia-EN) and

southern France (France-N). When projecting ancient individualsonto the top two principal components (PCs) defined by modernvariation, the Neolithic ancient Sardinian individuals sit betweenearly Neolithic Iberian and later Copper Age Iberian populations,roughly on an axis that differentiates WHG and EEF populations,and embedded in a cluster that additionally includes NeolithicBritish individuals (Fig. 2). This result is also evident in terms ofgenetic differentiation, with low pairwise FST ≈ 0.005–0.008,between Middle/Late Neolithic and Neolithic western mainlandEuropean populations (Fig. 3). Pairwise outgroup-f3 analysisshows a similar pattern, with the highest values of f3 (i.e., mostshared drift) being with Western European Neolithic and CopperAge populations (Fig. 3), gradually dropping off for populationsmore distant in time or space (Supp. Fig. 10).

Ancient Sardinian individuals are shifted towards WHGindividuals in the top two PCs relative to early NeolithicAnatolians (Fig. 2). Analysis using qpAdm shows that a two-way admixture model between WHG and Neolithic Anatolianpopulations is consistent with our data (e.g., p= 0.376 for Sar-MN, Table 1), similar to other western European populations ofthe early Neolithic (Supp. Table 1). The method estimates ancientSardinian individuals harbor HG ancestry (≈17 ± 2%) that ishigher than early Neolithic mainland populations (includingIberia, 8.7 ± 1.1%), but lower than Copper Age Iberians (25.1 ±0.9%) and about the same as Southern French Middle-Neolithicindividuals (21.3 ± 1.5%, Table 1, Supp. Fig. 13, ± denotes plusand minus one standard error).

In explicit models of continuity (using qpAdm, see Methods) thesouthern French Neolithic individuals (France-N) are consistentwith being a single source for Middle/Late Neolithic Sardinia (p=0.38 to reject the model of one population being the direct sourceof the other); followed by other western populations high in EEFancestry, though with poor fit (qpAdm p-values < 10−5, Supp.Table 2). France-N may result in improved fits as it is a bettermatch for the WHG and EEF proportions seen in Middle/LateNeolithic Sardinia (Supp. Table 1). As we discuss below, caution is

b n = 961

a

n = 70

4000 3000 2000

106

105

104# S

NP

s co

vere

d01000 1000

Age [years BCE]

Early copper age Early bronze age Nuragic Phoenician/Punic

Roman MedievalMiddle late neolithic

COR/SNNMSR/VIL AMC

Fig. 1 Number of SNPs covered, sampling locations and ages of ancient individuals. a The number of SNPs covered at least once and age (mean of 2σradiocarbon age estimates) for the 70 ancient Sardinian individuals. b The sampling locations of ancient Sardinian individuals and a reference dataset of 961ancient individuals from across western Eurasia and North Africa (with “jitter” added to prevent overplotting; see Supp Data 1E for exact locations).

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6 ARTICLE

NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 |www.nature.com/naturecommunications 3

www.nature.com/naturecommunicationswww.nature.com/naturecommunications

100020003000400050006000

Age [years BCE]

Sardinia(Middle Neolithic toNuragic sites)

Sardinia(Sidore et al. 2015

contemporarysample)

Sardinia(Post-Nuragic

sites)

a Modern groups

Olbia-Tempio (n = 18)Sassari (n = 87)Nuoro (n = 75)Oristano (n = 84)Ogliastra (n = 419)M. Campidano (n = 72)Cagliari (n = 289)Carbonia-I. (n = 41)Unknown (n = 492)

b d Ancient groups

cAncient Sardinian individualAncient individual

A Modern group (median)0.02

0.00

PC

2 (0

.15%

)

–0.02

0.000

–0.04

0.00

–0.01

PC

2 (0

.15%

)

–0.02

–0.03

–0.02 0.00 0.02

Anatolia-N(n = 24)

Balkans-EN(n = 17)

CE-EN(n = 39)

Iberia-EN(n = 10)

France-N(n = 4)

Sar-MN(n = 6)

Sar-ECA(n = 3)

Sar-EMBA(n = 27)

Iberia-LCA(n = 24)

Iberia-BA(n = 48)

Iberia-ECA(n = 15)

GB-EN(n = 14)

GB-LN(n = 34)

CE-MNCA(n = 14)

GB-EBA(n = 71)

CE-EBA(n = 147)

Anatolia-BA(n = 3)

Minoan-BA(n = 10)

Myc-BA(n = 4)

Balkans-IA(n = 3)

Balkans-BA(n = 46)

GB-LBA(n = 34)

CE-LBA(n = 30)

Sar-Nur(n = 16)

Balkans-MNCA(n = 118)

Greece-N(n = 6)

Anatolia-CA(n = 1)

Morocco-EN(n = 4)

Morocco-LN(n = 3)

PC1 (0.30%)

"WesternHunter/Gatherer"(WHG)

“Steppe”

"EarlyEuropeanFarmer"(EEF) "Ancient

NorthAfrican"

"Neolithic/Copper AgeWest Europe"

–0.050 –0.025 0.025

Fig. 2 Principal components analysis based on the Human Origins dataset. a Projection of ancient individuals’ genotypes onto principal component axesdefined by modern Western Eurasians and North Africans (gray labels, see panel (c) for legend for all abbreviations but ‘Can', for Canary Islands). b Zoominto the region most relevant for Sardinia. Each projected ancient individual is displayed as a transparent colored point in panel (a) and (b), with the colordetermined by the age of each sample (see panel (d) for legend). In panel (b), median PC1 and PC2 values for each population are represented by three-letter abbreviations, with black or gray font for moderns and a color-coded font based on the mean age for ancient populations. Ancient Sardinianindividuals are plotted as circles with edges, color-coded by age, and with the first three letters of their sample ID (which typically indicates thearcheological site). Modern individuals from the Sidore et al. sample of Sardinia are represented with gray circles and modern individuals from thereference panel with gray squares. See Fig. 5 for a zoomed in representation with detailed province labels for Sardinian individuals. The full set of labels andabbreviations are described in Supp. Data 1E, F. c Geographic legend of present-day individuals from the Human Origins and our Sardinian referencedataset. d Timeline of selected ancient groups. Note: The same geographic abbreviation can appear multiple times with different colors to represent groupswith different median ages.

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6

4 NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 | www.nature.com/naturecommunications

www.nature.com/naturecommunications

necessary as there is a lack of aDNA from other relevantpopulations of the same period (such as mainland Italian Neolithiccultures and neighboring islands).

For our sample from the Middle Neolithic through the Nuragic(n= 52 individuals), we were able to infer mtDNA haplotypes foreach individual and Y haplotypes for 30 out of 34 males. ThemtDNA haplotypes belong to macro-haplogroups HV (n= 20),JT (n= 19), U (n= 12), and X (n= 1), a composition broadlysimilar to other European Neolithic populations. For Yhaplotypes, we found at least one carrier for each of three majorSardinia-specific Y founder clades (within the haplogroups I2-M26, G2-L91, and R1b-V88) that were identified previously basedon modern Sardinian data51. More than half of the 31 identified Yhaplogroups were R1b-V88 or I2-M223 (n= 11 and 8,

respectively, Supp. Fig. 6, Supp. Data 1B), both of which arealso prevalent in Neolithic Iberians14. Compared with most otherancient populations in our reference dataset, the frequency ofR1b-V88 (Supp. Note 3, Supp. Fig. 6) is relatively high, but as weobserved clustering of Y haplogroups by sample location (Supp.Data 1B) caution should be exercised with interpreting our resultsas estimates for island-wide Y haplogroup frequencies. The oldestindividuals in our reference data carrying R1b-V88 or I2-M223were Balkan hunter-gatherer and Neolithic individuals, and bothhaplogroups later appear also in western Neolithic populations(Supp. Figs. 7–9).

Continuity from the Middle Neolithic through the Nuragic.We found several lines of evidence supporting genetic continuity

Ancient ModernSardinia

Ancient Eurasia/North Africa

2.3

2.6

5.5

14

11.4

2.3

−1.6

2.4

9.9

7.2

2.6

−1.6

1

10.9

8.2

5.5

2.4

1

9.8

7.1

6.7

5.1

4.9

5.9

14.1

11.2

7.6

5.4

6.4

5.6

15.5

12.7

11.7

5

10.4

9.9

19.6

16.5

9.7

7

9.7

10.3

15.5

12

10.3

7.5

10.2

10.8

17.3

13.6

13.5

8.3

12.7

13.2

18.3

14.7

50.7

47.4

49.2

49.1

44.4

39.2

33.1

29.5

29.6

29.7

27.6

22.9

84.3

79.3

80.5

81

77

72.1

76.7

76

73.6

74.8

79

75.5

69.8

69.2

66.6

67.8

73.3

69.6

7.4

6.1

6.6

7

13.1

9.9

8.8

8.3

7.4

7.7

17.9

14.9

7.5

6.2

6.2

6.6

16.2

13.5

8.4

4.2

5.3

5.6

13.2

9.9

11.8

6.7

10.2

11.3

20.8

17.8

23

20.3

22.1

22.4

25.4

21

19.7

17.3

18.5

19

21.6

17.3

34.9

31.8

33.1

33.4

31.1

26.2

18.5

15.1

18.9

17.7

21.6

17.6

17.1

7.7

12.4

13.6

16.6

12.3

25.8

18.4

23.9

23.7

19.9

15.3

27.2

20.3

26.1

26.7

28.5

23.9

188.5

175

184.6

183.5

180.5

175.4

35

20.4

23.3

30.9

30.1

26.2

17.9

15.5

15.2

14.6

11.1

7

20.8

16.8

17.1

17.2

14.9

11

14.9

11.7

12.1

11.4

8.8

4.9

15.9

12.5

14

13.5

9.6

5.4

19.8

17.4

17.6

16.7

12.4

7.9

18.2

13.8

15.4

14.4

10.2

6

21.8

16.9

19.3

17.8

13.3

8.7

18.4

14.3

16

15.2

10.6

6.2

27.3

24.3

25.4

24.9

20.5

15.9

27.5

23

25.6

25.1

20.1

15.6

19.8

17.5

17.5

16.8

12.5

7.8

23

19.5

20.3

20.1

15.5

10.8

28

24.3

25.7

25

19.3

14.6

27.2

23.2

24.8

24.1

18.5

13.8

25.2

21.7

23.2

23

16.8

12.2

29.5

27.6

28.1

27.5

23

18.4

44.8

41.1

43.1

43.1

38.8

34.3

27.8

24.3

25.5

24.6

19.4

14.7

23.9

20.7

21.7

21

15.1

10.4

25.3

21.2

22.8

22.1

16.9

12

14

9.9

10.9

9.8

2.7

11.4

7.2

8.2

7.1

2.7

Sar−MN

Sar−ECA

Sar−EMBA

Sar−Nur

Sar (Ogl)

Sar (non Ogl)

22.5

22.6

22.5

22.3

22.3

22.5

22.6

22.6

22.3

22.3

22.6

22.6

22.6

22.4

22.3

22.5

22.6

22.6

22.4

22.4

22.5

22.5

22.5

22.5

22.3

22.3

22.5

22.5

22.5

22.5

22.3

22.3

22.5

22.5

22.5

22.5

22.3

22.3

22.5

22.4

22.5

22.5

22.3

22.3

22.4

22.5

22.5

22.5

22.3

22.3

22.4

22.4

22.4

22.4

22.2

22.2

21.8

21.7

21.8

21.8

21.8

21.8

22

22

22

22

22

22

21.7

21.7

21.7

21.7

21.7

21.7

22.1

22.1

22.2

22.2

22

22

22.1

22.1

22.2

22.2

22

22

22.4

22.4

22.5

22.4

22.3

22.2

22.5

22.5

22.5

22.5

22.3

22.3

22.4

22.5

22.5

22.5

22.2

22.2

22.4

22.4

22.4

22.4

22.2

22.2

22.5

22.5

22.5

22.5

22.3

22.3

22.1

22.1

22.1

22.1

22

22

22.1

22.1

22.2

22.2

22

22

21.8

21.8

21.8

21.8

21.8

21.8

22.3

22.3

22.3

22.3

22.2

22.2

21.8

21.9

21.9

21.9

21.8

21.8

21.8

21.9

21.8

21.8

21.8

21.8

21.9

21.9

21.9

21.9

21.8

21.8

20.3

20.2

20.3

20.3

20.3

20.3

21.8

21.7

21.7

21.7

21.6

21.6

22.1

22.1

22.2

22.2

22.1

22.1

22.2

22.2

22.3

22.3

22.2

22.2

22.1

22.1

22.2

22.2

22.1

22.1

22.1

22.1

22.2

22.2

22.1

22.1

21.9

21.9

22

22

21.9

21.9

22

22

22

22.1

22

22

22

22

22

22

22

22

22.1

22.1

22.1

22.1

22.1

22.1

21.8

21.8

21.8

21.9

21.8

21.8

21.8

21.9

21.9

21.9

21.9

21.9

21.9

21.9

22

22

21.9

21.9

21.9

21.9

21.9

21.9

21.9

21.9

21.8

21.9

21.9

21.9

21.9

21.9

21.6

21.6

21.7

21.7

21.7

21.7

21.6

21.6

21.6

21.6

21.6

21.6

21.1

21.1

21.1

21.1

21.1

21.1

20.9

21

21

21

20.9

20.9

21.5

21.5

21.5

21.5

21.5

21.5

21.9

21.9

21.9

21.9

21.9

21.9

21.7

21.7

21.7

21.8

21.7

21.7

22.3

22.3

22.4

22.4

22.3

22.3

22.3

22.3

22.4

22.3

Sar−MN

Sar−ECA

Sar−EMBA

Sar−Nur

Sar (Ogl)

Sar (non Ogl)

Sar

−M

NS

ar−

EC

AS

ar−

EM

BA

Sar

−N

urS

ar (

Ogl

)S

ar (

non

Ogl

)G

B−

LNF

ranc

e−N

GB

−E

NIb

eria

−E

CA

Iber

ia−

EN

Bal

kans

−E

NIb

eria

−LC

AC

E−

EN

CE

−M

NC

AIb

eria

−B

AA

nato

lia−

NM

inoa

n−B

AIr

on_G

ates

−H

GW

HG

−H

GC

E−

EB

AG

B−

EB

AS

tepp

e−M

LBA

Myc

−B

ALe

vant

−B

AC

anaa

nite

Iran

−C

AM

oroc

co_L

NS

tepp

e−E

MB

AE

HG

−H

GM

oroc

co_E

NS

pani

shT

usca

nS

icili

anG

reek

Fre

nch

Tur

kish

_Jew

Mal

tese

Cyp

riot

Tur

kish

Mor

occa

n_Je

wB

asqu

eLe

bane

seJo

rdan

ian

Pal

estin

ian

Dru

zeB

edou

inA

Tun

isia

n_Je

wLi

byan

_Jew

Tun

isia

nS

ahar

awi

21.7

21.9

22.1

22.3

22.5

f3 x 100

0.0

20.0

40.0

60.0

80.0

Fst x 1000

Modern Eurasia/North Africa

Fig. 3 Genetic similarity matrices. We calculated FST (upper panel) and outgroup-f3 (lower panel) of ancient Sardinian (Middle/Late Neolithic to Nuragicperiods) and modern Sardinian individuals (grouped into within and outside the Ogliastra region) with each other (left), various ancient (middle), andmodern populations (right) of interest. The full sharing matrices can be found in Supp. Figs. 10/11, where we also include post-Nuragic sites. For ancientgroups, sample sizes and time-spans are provided in Fig 1d. The full set of labels and abbreviations are described in Supp. Data 1E, F.

Table 1 Results from fitting models of admixture with qpAdm for Middle Neolithic to Nuragic period.

Proxy source populations Admixture fractions Standard error

Target a b c p-value a b c a b c

A Sar-MN WHG Anatolia-N – 0.376 0.177 0.823 – 0.014 0.014 –Sar-ECA WHG Anatolia-N – 0.268 0.161 0.839 – 0.020 0.020 –Sar-EMBA WHG Anatolia-N – 0.049 0.161 0.839 – 0.007 0.007 –Sar-Nur WHG Anatolia-N – 0.134 0.163 0.837 – 0.009 0.009 –

B Sar-MN WHG Anatolia-N Steppe 0.265 0.177 0.823 0.000 0.016 0.023 0.026Sar-ECA WHG Anatolia-N Steppe 0.18 0.164 0.836 0.000 0.023 0.032 0.036Sar-EMBA WHG Anatolia-N Steppe 0.032 0.162 0.838 0.000 0.009 0.012 0.013Sar-Nur WHG Anatolia-N Steppe 0.089 0.163 0.837 0.000 0.010 0.014 0.016

C France-N WHG Anatolia-N Steppe 0.093 0.213 0.787 0.000 0.018 0.023 0.027Iberia-EN WHG Anatolia-N Steppe 0.243 0.087 0.913 0.000 0.012 0.017 0.019Iberia-LCA WHG Anatolia-N Steppe 0.045 0.251 0.749 0.000 0.012 0.015 0.018Iberia-BA WHG Anatolia-N Steppe 6.0 × 10−3 0.239 0.689 0.072 0.010 0.014 0.016CE-EN WHG Anatolia-N Steppe 0.656 0.046 0.954 0.000 0.007 0.010 0.012CE-LBA WHG Anatolia-N Steppe 0.105 0.128 0.403 0.468 0.008 0.011 0.013

(A) Two-way models of admixture for ancient Sardinia using Western Hunter-Gatherer (WHG) and Neolithic Anatolia (Anatolia-N) individuals as proxy sources. (B) Three-way models of admixture forancient Sardinia using Western Hunter-Gatherer (WHG), Neolithic Anatolia (Anatolia-N), and Early Middle Bronze Age Steppe (Steppe-EMBA, abbreviated Steppe in table) individuals as proxy sources.(C) Three-way models for select comparison populations on the European mainland. Full results are reported in Supp. Info. 4.

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6 ARTICLE

NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 |www.nature.com/naturecommunications 5

www.nature.com/naturecommunicationswww.nature.com/naturecommunications

from the Sardinian Middle Neolithic into Bronze Age andNuragic times. Importantly, we observed low genetic differ-entiation between ancient Sardinian individuals from varioustime periods (FST= 0.0055 ± 0.0014 between Middle/Late Neo-lithic and late Bronze Age, Fig. 3). Furthermore, we did notobserve temporal substructure within the ancient Sardinianindividuals in the top two PCs—they form a coherent cluster(Fig. 2). In stark contrast, ancient individuals from mainlandregions such as central Europe show large movements over thefirst two PCs from the Neolithic to the Bronze Age, and also havehigher pairwise differentiation (e.g., FST= 0.0200 ± 0.0004between Neolithic and Bronze Age individuals from centralEurope, Supp. Fig. 11). A qpAdm analysis cannot reject a modelof Middle/Late Neolithic Sardinian individuals being a directpredecessor of Nuragic Sardinian individuals (p= 0.15, Supp.Table 2, also see results for f4 statistics, Supp. Data 2). OurqpAdm analysis further shows that the WHG ancestry propor-tion, in a model of admixture with Neolithic Anatolia, remainsstable at 17 ± 2% through the Nuragic period (Table 1A). Whenusing a three-way admixture model, we do not detect significantSteppe ancestry in any ancient Sardinian group from the Middle/Late Neolithic to the Nuragic, as is inferred, for example, in laterBronze Age Iberians (Table 1B, Supp. Fig. 13). Finally, in a five-way model with Iran Neolithic and Moroccan Neolithic samplesadded as sources, neither source is inferred to contribute ancestryduring the Middle Neolithic to Nuragic (point estimates arestatistically indistinguishable from zero, Supp. Fig. 14).

From the Nuragic period to present-day Sardinia: signatures ofadmixture. We found multiple lines of evidence for gene flowinto Sardinia after the Nuragic period. The present-day Sardinianindividuals from the Sidore et al. sample are shifted from theNuragic period ancients on the western Eurasian/North-AfricanPCA (Fig. 2). Using a “shrinkage” correction method for theprojection is key for detecting this shift (see Supp. Fig. 23 for anevaluation of different PCA projection techniques). In theADMIXTURE results (Fig. 4), present-day Sardinian individuals

carry a modest “Steppe-like” ancestry component (but generallyless than continental present-day European populations), and anappreciable “eastern Mediterranean” ancestry component (alsoinferred at a high fraction in other present-day Mediterraneanpopulations, such as Sicily and Greece) relative to Nuragic periodand earlier Sardinian individuals.

To further refine this recent admixture signal, we consideredtwo-way, three-way, and four-way models of admixture withqpAdm (Table 2, Supp. Figs. 15–18, Supp. Tables 3–5). We findthree-way models fit well (p > 0.01) that contain admixturebetween Nuragic Sardinia, one northern Mediterranean source(e.g., individuals with group labels Lombardy, Tuscan, French,Basque, Spanish) and one eastern Mediterranean source (e.g.,individuals with group labels Turkish-Jew, Libyan-Jew, Maltese,Tunisian-Jew, Moroccan-Jew, Lebanese, Druze, Cypriot, Jorda-nian, Palestinian) (Table 2C, D). Maltese and Sicilian individualscan provide two-way model fits (Table 2B), but appear to reflect amixture of N. Mediterranean and E. Mediterranean ancestries,and as such they can serve as single-source proxies in two-wayadmixture models with Nuragic Sardinia. For four-way modelsincluding N. African ancestry, the inferences of N. Africanancestry are negligible (though as we show below, forms of N.African ancestry were already likely present in the easternMediterranean components).

Because of limited sample sizes and ancestral source mis-specification, caution is warranted when interpreting inferredadmixture fractions; however, the results indicate that complexpost-Nuragic gene flow has likely played a role in the populationgenetic history of Sardinia.

Refined signatures of post-Nuragic admixture and hetero-geneity. To more directly evaluate the models of post-Nuragicadmixture, we obtained aDNA from 17 individuals sampled frompost-Nuragic sites. The post-Nuragic individuals spread across awide range of the PCA, and many shift towards the “eastern” and“northern” Mediterranean sources posited above (Fig. 2). Weconfidently reject qpAdm models of continuity from the Nuragic

Sar

−E

MB

A

Sar

−N

ur

Sar

−V

IL

Iber

ia−

EN

GB

−E

N

CE

−E

N

Bal

kans

−E

N

Ana

tolia

−N

Iber

ia−

BA

Iber

ia−

LCA

Iber

ia−

EC

A

CE

−E

BA

Min

oan−

BA

Ste

ppe−

MLB

A

Ste

ppe−

EM

BA

Iron

_Gat

es−

HG

WH

G−

HG

Mor

occo

_Ibe

rom

auru

sian

Sar

dini

an Olb

Sas

Nuo Ogl Ori

Cam Car

Cag

Spa

nish

Bas

que

Fre

nch

Tus

can

Sic

ilian

Tun

isia

n_Je

w

Mor

occa

n_Je

w

Mal

tese

Liby

an_J

ew

Gre

ek

Tur

kish

_Jew

Tur

kish

Cyp

riot

Leba

nese

Jord

ania

n

Pal

estin

ian

Dru

ze

Bed

ouin

A

Moz

abite

Egy

ptia

n

Tun

isia

n

Alg

eria

n

Sah

araw

i

Anc

ient

Mod

ern

Sar

−M

NS

ar−

EC

A

Sar

−O

RC

002

Sar

−M

SR

Sar

−A

MC

Sar

−C

OR

Sar

−S

NN

Fra

nce−

N

Myc

−B

AIr

an−

CA

Iran

−N

EH

G−

HG

Mor

occo

_EN

Mor

occo

_LN

Fig. 4 Admixture coefficients estimated by ADMIXTURE (K= 6). Each stacked bar represents one individual and color fractions depict the fraction of thegiven individual's ancestry coming from a given “cluster”. For K= 6 (depicted here), Sardinian individuals up until the Nuragic share similar admixtureproportions as other western European Neolithic individuals. Present-day as well as most post-Nuragic ancient Sardinian individuals have elevated Steppe-like ancestry (blue), and an additional ancestry component prevalent in Near Eastern/Levant populations (orange). An ancient North-African component(green) appears at low fraction in many present-day Mediterranean populations, and somewhat stronger in samples from the Sardinian Punic site Villamar.ADMIXTURE results for all K= 2,… , 11 are depicted in the supplement (Supp. Fig. 19).

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6

6 NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 | www.nature.com/naturecommunications

www.nature.com/naturecommunications

period for all of these post-Nuragic samples, apart from a samplefrom S’Orcu ’e Tueri (ORC002, Table 2E, Supp. Table 6). TheADMIXTURE results concur, most post-Nuragic individualsshow the presence of novel ancestry components not inferred inany of the more ancient individuals (Fig. 4).

Consistent with an influx of novel ancestry, we observed thathaplogroup diversity increases after the Nuragic period. Inparticular, we identified one carrier of the mtDNA haplogroupL2a at both the Punic Villamar site and the Roman Monte Carrusite. At present, this mtDNA haplogroup is common acrossAfrica, but so far undetected in samples from Sardinia36. We alsofound several Y haplogroups absent in our Neolithic trough theNuragic period sample (Supp. Fig. 6). R1b-M269, at about 15%within modern Sardinian males51, appears in one Punic (VIL011)and two Medieval individuals (SNN002 and SNN004). We alsoobserved J1-L862 in one individual from a Punic site (VIL007)and E1b-L618 in one medieval individual (SNN001). Notably, J1-L862 first appears in Levantine Bronze Age individuals within theancient reference dataset and is at about 5% frequency inSardinia today.

We used individual-level qpAdm models to further investigatethe presence of these new ancestries (Supp. Data 3). In addition tothe original Neolithic Anatolian (Anatolia-N) and HunterGatherer (WHG) sources that were sufficient to model ancientSardinians through the Nuragic period, we fit models withrepresentatives of Steppe (Steppe-EMBA), Neolithic Iranian(Iran-N), and Neolithic North-African (Morocco-EN) ancestryas sources. We observe the presence of the Steppe-EMBA (pointestimates ranging 0–20%) and Iran-N components (pointestimates ranging 0–25%) in many of the post-Nuragicindividuals (Supp. Fig. 14). All six individuals from the PunicVillamar site were inferred to have substantial levels of ancientNorth-African ancestry (point estimates ranging 20–35%, Supp.Fig. 14, also see ADMIXTURE and PCA results, Figs. 2 and 4).When fit with the same five-way admixture model, present-daySardinians have a small but detectable level of North-Africanancestry (Supp. Fig. 14, also see ADMIXTURE analysis, Fig. 4).

Models with direct continuity from Villamar to the present arerejected (Table 2F, Supp. Table 6). In contrast, nearly all the otherpost-Nuragic sites produce viable models as single sources for the

Table 2 Results from fitting models of admixture with qpAdm and Sardinian aDNA as sources.

Proxy source populations Admixture fractions Standard error

Target a b c p-value a b c a b c

A Sar-ECA Sar-MN – – 0.175 – – – – – –Sar–EMBA Sar-ECA – – 0.769 – – – – – –Sar-Nur Sar-EMBA – – 0.765 – – – – –Cagliari Sar-Nur – –

modern Sardinians (e.g., Sar-COR qpAdm p-values of 0.16 and0.261 for Cagliari and Ogliastra, respectively; Sar-SNN qpAdm p-values of 0.037 and 0.016, similarly Table 2F, Supp. Table 6). Wefound some evidence of substructure: Sar-ORC002 (from aninterior site) is more consistent with being a single source forOgliastra than Cagliari, whereas Sar-AMC shows an oppositepattern (Supp. Table 6).

We also carried out three-way admixture models for eachpost-Nuragic Sardinian individual using the Nuragic sample asa source or outgroup, and potential sources from variousancient samples that are representative of different regions ofthe Mediterranean. We found a range of models can be fit foreach individual (Supp. Tables 7–8). For the models withNuragic as a source, by varying the proxy populations, one canobtain fitted models that vary widely in the inferred Nuragiccomponent (e.g., individual COR002 has a range from 4.4 to87.8% across various fitted models; similarly, individualAMC001, with North-African mtDNA haplogroup U6a, had arange form 0.2 to 43.1%, see Supp. Tables 7–8). TheORC002 sample had the strongest evidence of Nuragic ancestry(range from 62.8 to 96.3%, see Supp. Tables 7–8). Further, theVIL, MSR, and AMC individuals can be modeled with NuragicSardinian individuals included as a source or as an outgroup,while the two COR and ORC002 individuals can only bemodeled with Nuragic individuals included as a source. Oneindividual from the medieval period San Nicola Necropoli(SNN001) was distinct in that we found their ancestry can bemodeled in a single-source model as descendant of a populationrepresented by present-day Basque individuals (Supp. Table 8).When we apply the same approach to present-day Sardinianindividuals, we find models with the Nuragic sample as anoutgroup fail in most cases (Supp. Table 9). For models thatinclude Nuragic as a possible source, each present-dayindividual is consistent with a wide range of Nuragic ancestry.The models with the largest p-values return fractions of Nuragicancestry that are close to, or higher than 50% (Supp. Table 9),similar to observed in our population-level modeling (Table 2).

Fine-scale structure in contemporary Sardinia. Finally, weassessed our results in the context of spatial substructure within

modern Sardinia, as previous studies have suggested elevatedlevels of WHG and EEF ancestry in Ogliastra32.

In the PCA of modern west Eurasian and North-Africanvariation, the ancient Sardinian individuals are placed closest toindividuals from Ogliastra and Nuoro (see Figs. 2 and 5a). At thesame time, in a PCA of just the modern Sardinian sample, theancient individuals project furthest from Ogliastra (Fig. 5b).Interestingly, individual ORC002, dating from the Punic periodand from a site in Ogliastra, projects towards Ogliastraindividuals relative to other ancient individuals.

Further, in the broad PCA results, the median of the province ofOlbia-Tempio (northeast Sardinia) is shifted towards mainlandpopulations of southern Europe, and the median for Campidano(southwest Sardinia) shows a slight displacement towards theeastern Mediterranean (Fig. 5a). A three-way admixture model fitwith qpAdm suggests differential degrees of admixture, with thehighest eastern Mediterranean ancestry in the southwest (Carbonia,Campidano) and the highest northern Mediterranean ancestryin the northeast of the island (Olbia, Sassari, Supp. Fig. 17). Theseobservations of substructure among contemporary Sardinianindividuals contrast our results from the Nuragic and earlier,which forms a relatively tight cluster on the broad PCA (Fig. 2) andfor which the top PCs do not show any significant correlations withlatitude, longitude, or regional geographic labels after correcting formultiple testing (Supp. Figs. 24–33).

DiscussionOur analysis of genome-wide data from 70 ancient Sardinianindividuals has generated insights regarding the population his-tory of Sardinia and the Mediterranean. First, our analysis pro-vides more refined DNA-based support for the Middle Neolithicof Sardinia being related to the early Neolithic peoples of theMediterranean coast of Europe. Middle/Late Neolithic Sardinianindividuals fit well as a two-way admixture between mainlandEEF and WHG sources, similar to other EEF populations of thewestern Mediterranean. Further, we detected Y haplogroups R1b-V88 and I2-M223 in the majority of the early Sardinian males.Both haplogroups appear earliest in the Balkans among Meso-lithic hunter-gatherers and then Neolithic groups9 and later inEEF Iberians14, in which they make up the majority of Y

Sas

Sas

–0.025 0.000 0.025 0.050 0.075

PC1 (0.16%)

–0.04

–0.02

0.00

0.02

0.04

PC

2 (0

.10%

) Sas

Cam

Cag

Cag

Sas

Sas

Sas

Sas

Sas

SasSas

Sas

Sas

Sas

Sas

Sas

Sas

Sas

Sas

Sas

Sas

SasSas

Sas Sas

Olb

Olb

Sas

Sas

Sas

Sas

Sas

Sas

SasSas

Sas

Nuo

Sas

Sas

Olb

Sas

Sas SasSas

Sas

Nuo

Nuo

Sas

Sas

Nuo

Nuo

Cag

Ori

Sas

Sas

Sas

SasSas Sas

SasSas

Sas

Ori

Nuo

Nuo

Nuo

Sas

Sas

Olb

Sas

OlbSas

Olb

Sas

Olb

Cag

Nuo

Ori

Nuo

CagCag

Cag

Car

Cam

Cag

CagCag

Ori

Cam

Sas

Sas

Cag

Ori

Ori

Ori

Cam

Nuo

Ori

Cag

Cag

Ogl

Cam

Nuo

Nuo

Sas

Cam

Nuo

Nuo

Cag

Cam

Ori

Ori

Nuo

Cam

CagCamOri

Ori

Cag

Car

Cag

Cag

Car

Cam

Ori

Ori

Cag

Cam

Cam

Ori

CarOri

Ori Cag

Nuo

Nuo

Car

Cag

Cam

Olb

Olb

Cag

Car

Car

SasCamCam

CamCam Cag

Sas

Nuo

Ori

Car

Olb

Nuo

Ori

Cam

Sas

Olb

Olb

Sas

Sas

Nuo

Olb

Cag

Sas

Nuo

NuoNuo

Nuo

Car

Car

Car

Cam

Cam

Sas

Car

Nuo Cag

Car

OlbOri

Sas

Olb

Sas

Sas

Cag

CagCagSas

Sas

Cam

CamCag

Cag

Cag

Olb

Sas

Cam

Ori

Car

Cam

Nuo

Cag

Cag

Ori

Ori

Ori CamCam

OriOri

Cag

Cag

Ori

Ori Cag

Cag

Ori

Ori

Ori

CagOri

Ori

OriOri

Nuo

Cag

CamCam

Cam

Cam

Cag

Cag

Cag

Cag

Cag

Cag

Cam

Cam

Cam

CamCamCam

Cag

Car

Car

Cag

CagCar

Cag

CamCag

Cag

Cag

CamCam

CagCam

Cam

Nuo

Cam

OriCar

Ori

Nuo

Nuo

Nuo

Ori

Cam

Cam

Ori

Ori

Cag

Cam

Cag

Nuo

NuoCar

Cag

Ori

Ori

Cag

CamSas

Nuo

Cag

Cag

Cam

Ori

OriOri

OriCag

Cam

Cag

Cam

Sas

Ori

Cag

OriOri

Cam

Ori

Ori

Nuo

Nuo

Ori

Ori

Ori

Cam

CagNuo

NuoNuo

Nuo

CagOriOri

Cag

Cag CagOri

Ori

Cam

Cam

Sas

Ori

Ori

Cam

Cam

Cag

CagCag

Cam

Nuo

Nuo

Sas

Nuo

CamOri

Ori

Nuo

Nuo

Nuo

Cam

Cam

Cag

Car

Car

Car

Car

Cag

Cag

Car

Car

Cam

Cam

Car

Nuo

Cag

Car

Car

Car

Car

Car

Car

Nuo

Nuo

Cam

Cam

Ori

Cag

Olb

SasSas

CamCam

Cag

CagCag

Cag

Cam

Ori

Cag

Ori

Ori

Cag

Nuo

NuoCam

Cam

Ori

Nuo

Cag

Cag

Ori

Ori

Nuo

Nuo

Nuo

Sas

Cag

Cag

Cag

Cag

Cag

Car

Car

Car

Ori

Ori

NuoNuo

Ori

Ori

Ori

Olb

Ogl

Ori

Ogl

Ogl

Ogl

Cag

Ogl

Ogl

Ogl

Ogl

Nuo

Ogl

Ogl

Nuo

Ogl

Ogl

Ogl

Nuo

Cag

Ogl

Ogl

CagCag

Sas

Ogl

Nuo

Car

SasOgl

Ogl

Ori

Sas

Ogl

Car

Nuo

Ori

Ogl

Ogl

Ogl

OglCag

Ogl

Ori

Nuo

Ogl

Ogl

Ogl

Cag

Ogl

Cag

Cam

Ogl

Ogl

Ogl

Car

OglSas

Ogl

Ogl

Nuo

Nuo

Ogl

Ogl

Ogl

Ori

Ogl

Cag

Ogl

Cag

Ogl

OglOgl

Ori

OglOgl

Ogl

Ogl

OglOglOgl

NuoOgl

Ogl

Ogl

Nuo Ogl

Ogl

Ogl

Ori

Ogl

Nuo

Ogl

Ogl

Ogl

Ogl

Nuo

Ogl

Ogl

Ogl

Ogl

Cag

Cag

Nuo

Ogl

Car

Ogl

Ogl

Nuo

Cag Ogl

Ogl

Ogl

Ogl

Ogl

Nuo

Ogl

Ori

Ogl

Ogl

Cam

Sas

Ogl

Ogl

Cam

Ogl

Cam

SasNuo

Ogl

Ogl

Ogl

Ogl

Ori

Ogl

Ogl

OglOgl

Ogl

Ogl

Ogl

Ogl

Ogl

Ori

Ogl

Ogl

Ogl

Car

Sas

Ogl

Ogl

Ogl

Car

Nuo

Sas

Ogl

ORC

SNN

ORRoRRR

gg

C

g

C

OlbagOlbg bag

SUC

SEC

SUA

VIL

C

guoo

guoggo

S12

VIL

FIL

SUA

b

SSaS VILasasSSasS LILasas

b

SS LLLL

CammCam

O

SuoNu S

OOOO

COri

C

oNuo

ammaro

ammarao

Nuo

mma ma

CarCCa

MSR

SEC

SEC

AMC

SEC

ORC

SUCORC

OOOOOOOO

SEECEOOOOOOOROOROO

SEECECC

OO CCCCC rio

CCMA8NOERCRRCC

CCCUUUUoUUCCRCRRCRR

CNNCNNNCCCNNuNNuCNNNNCNOooooooooooooooooooCCCCCCCCCC

CCCCUUUCCCCCCCCCCCCCCNuCNuNNCCCNCCCNNCN

NONNooooNNoouuuouoNNNNoooooRRUooUU

RRRR OOoooooooooooooOOoooooo

uO

mmmmmmmCCCCCCCCCCSCCmmmCCCCCCCC

SooooSS

O

Car

O

Car

a

C

CaCa

C

Ca

arRRRORRO

arOOO

SUC

MA1

SID

SUC

ISB

MA7

asssSas

OO

RRRROMRRRRRRRROROROROROORRRRRORROOOOOOOO

aggagg

amN

mmNu

amammmu

mmmmmu

SUC

MSR

PJU

MA8

SUA

SSSOOOO

SSSOOOONNCNNOOOOCCOOOOONCCOOOOOOCCOOOOOOOCCOOOOOONNCCOOOOOOOOOOOOOOOSSSSNNCNCCSSSSIDSISSSS

EEENNEENNNNNNEENENNCNS

UUUCCCCEESESENNNNSSESS NNNNNNEENNEENNNNNNNNNNNNNNN SuuSNNuNNuSuuSNNNNuuNNuuuSuuSNNNNuuCNNCuuCC NNNNNNNNNNNNCCCUCUUCCCCEESESENNS NCNNCNNNNNNNNNNCNNNNNNNNNNNNsss

OOSSS

OOROROOOOOOOOOOOOOOOOOOOOOOOSooooSS

S DUUSSSSS SSooooS

CCCAAAUAUAAAAaaaaaaMMM

a JJJJPPJgggJJPJggggMMMMMM

CaCaCaagagagaaggJUJPJJggJPJJggPgg UggCCCAUAAUUAA

OOOOMOOOOOOOMMMMRRRRRROOOROUU

N

gg SEECUUCCCCUUUCUUUUUCU NCCCNCCCC

gUUUg SSSSSESESSgUUUUU

DDD og uugg uouoSESESEEuuSSSSSggD

Nu

Eoo

EECCCUUUUUCUCNNCCNNCCCCUUooSoUooUUoo

DD EE

N

ORC

VIL

ORC

ORC

CaCCCCCCCC C77CCCC77CRC77AA77CRCRCCCCCCCC777CCCCCCCCCC77UUUUU777AAA7A7A

UJUAAAUU

RCRCRARRRRCJJJJAAAUJJ

AA CaUUUUUUJUCUUUU

oSSSSSSMAMAEEEEooMooAMAooAMooAMAooAMA

mmSSSESESS

AAAAAAAAESEEEEEEEEEESESAAAAA

Nuog

oNggg

NuoNooN

SaCa

SasSCa

asSID

LON

MA8

ISC

i CMAAriri AMAAANNA8MAMAM NN88ANN8A888

Ci CCaN

Ci

ggMrrMAMAMgggg NN88NNA8A8NN88

O

gggg

Ol

gOlbgg

sC O

CsC Oi

CC

OOO

O

sag mmmmm

aggCa

gmmgmgg

OriO

rri

OO2OO2OOOO2OO

iA11A1AA122221111111

ririi

C

iAAA

S

C

O

s

S

Ca

O

s

NCag

mmmISSSISIIISSIIDDDDDDDDD

NNmmDmmmmDDDDDDDDDDORNNRRNNOO

gggRggOgOgOgOgOOOggggOOggggDDgDDDDDDDDDDDggDDDDDDmmo NNNmm

agNNN

gN

gCa

SSSISIDIS DDDDDDDDDD RNuNRRRNNgggggggg uSgOOOgggOggggggDDDDDD

NN

S

NN

SEC

SNN

SUA

SUC

SUC

NOE

SUC

VIL

ORC

Ori OrrrriOr riri

CagOri

E

OrOag

NiO

EEEEEE

OrCCC

OOCCCC

OOooOOOOOO

rririrriiri ECCEESSiSiiSiirirrrSii22CCCCCoCCCCCCCEECCCNCCCCCNNNCCCNNCCRC

ENNSSSNNNSSS uoouENuNEuEouuuouooiRi

RESSSSSSS ooCooEECEEESEESS

SS1S1S12S12S1S12SSSSSCECSEESS22

iSSiiCCCC

NNCCCCCCCCCNNNNCCNNNNNCCCCuuUNuN

AAASSSSSS11SSS1S11111SSS1S1S OO2OOOO222O2OO22OOO22222rRrrRr

COOOOOOCC

CCCooCCCCCOO8OOOOOOOOOOOO8OOOO8OOOOOOOORCRCCu

oooooooouuuouoooRRRooooooooooooo

SCCCCCCC

CCCCCCCCCCCCCCCCCUCCCCCCCCUUUCCUuUuuUUuUUNNuuuuuuuu

MMMAMAMAASSSSSSSMMCC OOOO2OOOOOOA1AA1OO2OOOO222

CCCCAAAAA2

asC

SaaCLSaSaLLLILssLSSILssL

O

CaCLLLVILssLLLLLLLLLL

SS

CaaSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSSMMLONOuOuONNuuMMuuuMMuuuuN oNNuouuNNuuuouuuNuNuuuuuuuuouoouoouuououououoouooNNuuuuNuu

C SSSSSSSSSSSSCCSSSSmSSSSSSoM

SESSSSSSSSSSSSooMMuouououoooooooooooooooooooooooooooooooooooooooooo

s

S N

Sas

assN

Sas

as

CamOCOCaOUUCUCCCCaCa

rm

CCCEEEEE

CC

UUUUCUCC

SCaa

SCaa

Saaa

Saa

OrrCaCaSUSSririSCaCaSSSSOCC

Or

UUSSU

Caoag

agmm

agSSSggSSSmag mmgg mS

agg mm

CUUSUUo

Ca

CUU

mag mma mmUU

S12

AMC

MA1

ORC

SNN

MA8

AMC

VIL

SUA

LON

MA7

COR

MA1

SUA

OOri

OCOrOrCCCCCCOO

CO

CCCMAOrOrArii

OOri

OUOOCOO

O

OrOrOrAAiiAriri

OriNuo

g

iNu

COriCaaCaraCCCCC

OCCa

Sa

SCCOCO

SSaCa

Sas

SaCa

CC

Carmm

CNLNN

Cr SCSSSNSSNNNN

C

CCa

Cmm

CaLNN

CSSUNNSSNNNN

SUC

S12

COR

SNN

VIL

ORC

CagCar

Nuo

Ogl

Olb

Ori

Sas

CamCag

Nuo

–0.02 –0.01 0.00 0.01

PC1 (0.30%)

–0.03

–0.02

–0.01

–0.00

PC

2 (0

.15%

)

Ogl

Nuo

AnaBal

Bal

Bal

Bal CE

CE

FraGBGB

Gre

IbeIbe

Ibe

Ibe

Ita

MinMyc

Sic

Sas

Cag

Ber

Can

Cro

Fre

FreS

Mal

Sic

Spa

SpaN

Tus

Ogl

Ogl

?

?Ori

?

Car

?

Ogl

Ori

?Ori

OglOgl

Ogl

?

??

?

Ogl

Ogl

Car

Cag

Sas

Olb

?Cag

Ogl

Cag

?

Ogl

Ogl

Cag

Cag?

?

Olb

Cag

Cam?

Ori

OglCam

Cag

Cam

Nuo

Nuo

NuoCag

Sas?

?

Cam

Nuo?

Ogl

?

Cam

??

?

Ogl

Ori

?

?

Cag

Ori

OglOgl

Ori

Car

?

Car

Cam

Ogl

?

Cag

Ori

?

Cag

Ori

CagCam

Car

?

?

?

Cag

Ogl

?

Ogl

Cam

Sas

Ogl

?

Ori

Nuo

Ogl

?

Cag

Sas

?

Sas

?

Ogl

Ogl

Ogl

?

Ogl

Cag

Ogl

Ori

Car

Cag

Ogl

?Cag

Ogl

Cam

Sas ?Sas

CagOgl

?

?

Ogl

??

?

Cam

Ogl

?

Car

Cag

Cag

Cag

Ori

Ogl

Nuo

Cag

?Ogl

Cag

?

?

Ori

Cag

Cag

?

?

?

Ogl

?

?

?

?

Car

Ogl

?

? ?

Cag

??

Cag

Nuo

Cag

?

Cag

?

Ogl

Ogl

Cag

Cag

Ogl

CagOgl

?

Ogl

Ogl?

Ogl

CagSas

?

?

?Ori

Ogl

??

??

?

?

Ori

??

Ogl

?

Ogl

Cag

?

Ori

?

Sas

?

Cag

Ogl

Cag

Cam

?

Cag

Cag

Cag

?

Ogl

?

Ori

OriCag

Ogl

?

?Ori

Cag

Cag

Cag

Ogl

Cag

Cag

Ogl

Cag

Ogl?

Ogl

Cag?

Cag

Ogl

Ogl

CagCag

? Ogl

Nuo

?

Ori

Cag

Sas

Ori

?

Ogl

?

Ori

Nuo

Ogl

Sas

Nuo Ogl

Ori

Cag

Cag

Ogl

Nuo?

Cag

?Cag

Ogl

Cag

Cag

Cag

Ogl

Cag

Ogl

Cag

Ogl

Car

?

Cag

?

?

OglOgl

Ogl

Ori

Ogl

Ogl

Ogl

?

Cag

?Ori

Cag

?

Ogl

Ori

?

Ogl

Ogl

Nuo

Cag

Cag

?

Cag

Cag

Nuo

Cag

Ogl

Ori

?

Cag

Ori

CagCag

Cag

Cag

?

Ogl

?

CagCam

Cag

Ogl

Ori

?Ogl

Nuo

Cag

Ogl

OglOgl

Ogl

Cam

Olb

Ogl

Cag

?

?Cag

Ogl

Cag

Sas

?

Ogl

?

Ogl

Ogl

Cag

?

Ogl

?

Cag

Olb

Ogl

?

?

Ogl

Ogl

?

?

Ogl

CamCag

?

Cam

Ori?

Ogl

Cag

Cag?

CamSas

Cag

?Cag

Cag

Cam

Ogl

Car

Cag?

CagCam

Ogl

Cag

OglOgl

Ogl

Cag

Olb

Olb

OriCam

Cam

?

OglOgl

Nuo

Ogl

Cag

?

?

?

?

Ori

Cag

Ogl

?

Cag

?

Ogl

Sas

?

?

Cag

Nuo

?

Ogl

Cag

CamOgl

Ogl

Ogl

?Car

Ori

Ogl

?

?

?

?

Cam

Sas

Ogl

Cag

Ogl

?

?

?

?

Cag

CamCag

Ogl

?

?

Car

Ogl

?

Ogl

Cag

Cag

Cam

? Sas

?

?

Ogl?

Cag

Ogl

?

Sas

Nuo

Sas

Ori

Ogl

?

?

Cag

OglOgl

?

Ori

Cam

Ogl

Ogl

Ogl

Ogl

?

?

OglOgl

Olb

Cam

Cag

Ogl

Ogl

Cag

?

?

Ori

Sas

Cag

Cag

?

?

Cag

Nuo

Nuo

?

?

Olb

Cag

Sas

Ori

?

Cag

Ogl

?

NuoCag

Ogl

SasCag

?

Cag

SasOgl

?

Ogl

Cag

Sas

Sas

?

?

Ogl

Cag

Sas Cam

?

? Sas

Cag

Cam

Ogl

Cag

Cag

Cag

?

Ogl

Nuo

?

?Cag

?

Ori

Ogl

Ori

OglCam

?

Ori?

Ogl

Cag

?

Sas

Ogl

Ogl

Ogl

Ogl

Sas Cag

Cam

Sas

?

Ori?

Cam

Ogl

Car

Cag

? Ogl

Ogl

Ogl

?

Ori

Ogl

Cag?

Sas?

Ogl

Ori

Cag

Ogl

Cag

?

Cam

Cag

?

Ogl

?

Ogl

Ori

?

? ?

Ogl

?

CamCar

Cag

Car

?

Cag

Nuo

Ogl

Ogl

Ogl

Nuo

Cag?

Nuo

?

Nuo

OglOgl

CamOgl

Nuo

Ogl

Cag

Ogl

Ogl

Ogl

?Sas

?

?

Nuo

Ogl

Cag

Sas

Sas

?

Ori

OglNuo

Ogl

?

Cam

Sas

?

Cag

Ogl

NuoOgl

Olb

OriCag

Cag

Ogl

Cag

Sas

Cag

Ogl

Ori

Ogl

Car

Ogl

Cam

Nuo

?

Ogl

Ogl

Cam

Ogl

Ori

?

Cag

Ogl

Cag

?

Olb

Ogl

Nuo

OglOri

Cam

?

Cag

Cag

Cag

Ogl

?

Cag

Cag

Ogl

?

Ogl

Ogl

Cag

Cag

Ogl

Ogl

?

Ogl

Ogl

Cag

?

??

Ogl

Cag

Ogl

Ori

Nuo

Sas

Cag

Ogl

Nuo?

??Sas

Ogl

Ogl

Ori?

?

?

Ogl

?

Ogl

?

Ogl

OglCag

?

?

Cag

Ogl

CamCag

Ogl

Cag?

?

?

Ogl

Ogl

Ogl

Car

Ori

?

OglOri

Cag

Ogl

Ogl

Cag?

?

?

Nuo

Ogl

Ori

?

Ogl

Ogl

OglCag

?

Cag?

Cag

Ogl

Ogl

Sas

Cag

Car

Nuo

?

Ogl

Cag

?

Ogl

Sas

Car

?

Ogl

Ori

?

Ogl

Cag

Ogl

Cag

?

?

?

Nuo

Ogl

Ogl

?

Ogl

?

Ogl

Ori

Ori

Cag

?

Ogl

Cag

Ogl

OglOgl

?

?

Car

?

Ori

CagCag

Ogl

?

Cam

?

Cam

CarCag

Ogl

Ori

Ogl

Sas

?

?

Cag

Ogl

Cag

Ori

Ogl

??

?

?

Ori

Ori

?

?

NuoCag

Ogl

CagCag

?

Car

?

?

Cam

?

Cag

?

?

Cag

Cag

?

?

?

?

Cag

Cag

?

Sas

Ogl

Ogl

Ogl

Cag

Ogl

?

Ogl

Ori

?

?

?

Cag

Ogl

Cag

Cag

Cam

Ogl

?

Ogl

Ogl

?

Ogl

Sas

Ogl?

Cag

Ogl

Nuo

?

?

?

??

?

Ogl

?

Ogl

Cag

Sas

Ogl

?

?

Car

?

Ogl

OglOgl

Ogl

Cam?

Ori

?

Cag

??

?

?

Cam

?

Nuo

Sas

?

Car

?

?

Cam

Cag

Ogl

CarOriSas

?

?

Cag

Sas

Nuo

Car

?Ogl

OglCag

Ogl

Cag

Ogl

?

Nuo

Ogl

Ogl

?

CamCag

Cag

Ogl

CagCag

Ogl

Ogl

?

?

Ori

?

?Sas

?

Ogl

Cag

Cag

?

Sas?

Nuo Cag

?

?

Ogl

Ogl

Cag

Ogl

? Car

Car

Cam

?

Ori

Car

?

Cag

Cag

Ogl

Cam

?

?

?

?

?

Cag

Car

Ogl

Cam

Ogl

Ogl

?

?

Ogl Cam

?

NuoOgl

Nuo

Ogl

Ogl

Cag

??

Cag

Sas

Nuo

Car

?

Cag

Cag

?

Nuo

?

Car

Ogl

Cag

?

Ogl ?

?

?

Ogl

?Ogl

Ogl

Ori

Ogl

?

Cag

Sas

?

Ogl

?

?

Sas

Cag

?

Ogl

Cag

Ogl

Ogl

Ori

?

Nuo

OglOgl

?

Ogl

Ogl

?

Nuo

Cag

Ogl

Ogl

Sas

?

Cag?

?

?

Cam

?

Ori?

Ogl

Ogl

?

Cam

Cam

Sas

Ori

Ogl SasOgl

Ogl

?

Nuo

Ogl

Ogl

Cag

? Sas

Cam

?

Sas

Ogl

?Ori

Cag

Cag

?

?Ogl?

?

Ogl

Cag

?

Ogl

CagCam

Ori

Ogl

Car

?

Cag

?Cag ?

Nuo

Olb

?

Ogl

Cag

Cag

Ogl

?

Cag

Sas CamOgl

Cag

Ogl

Ogl

CagOgl Car

Ori

?

OglOgl

Cag

Cag

Cag

Cam

Cag

Ogl

OglOgl

Cag

Cag

Cag

OglOgl

Ogl

Cag

?

Nuo

Cam

Ori

Cag

Ogl

Ogl

?

Sas

?

Cag

?

Ogl

Ogl

Nuo

Nuo

Ogl

?

??

Ori

?

Sas

Ogl

Cag

?

?

?

Cag

?

Ogl

Car

Sas

Cag

Cag

Ogl

Cag

Ogl

Ogl

?Car

Ogl

?

Ogl

Sas

?

?

Ogl

Nuo

?

Sas

Ogl

?

Cag

Ogl

NuoOgl

Ogl

Ogl

Ogl

Car

?

?

Sas

Ogl

Ogl

Olb

Cag

Ogl

?

?

Cag

Cag

Ogl

Ogl

Sas

Ogl

?

Ogl

Cam

?

?

Nuo

Cag

Ori

Ogl

Ogl

Ogl

?

?

?

Ogl

Nuo

?Sas

Ogl

Olb

?

Ori

?

Sas

?

?

?

Ogl

?

?

Sas

Ogl

?

Sas

?

Olb

?

Ogl Ogl

Cag

Nuo

?

?

Cag

?

Cag

Ogl

Cam

Cag

?

Cag

?

?

Ogl

?

?

Olb

?

?

?Sas

?

Nuo

?

Ogl

Nuo

Nuo

?

Ogl

Cag

?

Sas

Ogl

Cag

Car

Ogl

?

Sas

?

?

Olb

Cag

?

??

OglOgl

Cag

Sas

Ogl

CagOri

Ogl

?

?

?

??

Nuo Ogl

?

Ogl

Cag

Nuo

Ogl?

Cam

Ogl

OglCag

?

Ogl

Cag

Ori

CagOgl

Ogl

Ogl

Cag

Cag

Ogl

Ogl

Ogl

?

Ogl

Ogl

Nuo ?

Cag?

?

?

Ogl

Nuo

Ogl

?

?

Ogl

?

?

?

Ogl

Ogl

Nuo

?Nuo

Sas

?

Ogl

Cag

Ogl

Ogl

Cag

Cag

?

Car

Cag

?

??

Ogl

Cag

Cag

Cag

?

Cam

Ogl

?

Ogl

?

Ogl

?

Cag

Ori

Ogl

?

?

Ogl

?

?

Ogl

Cag

Car

?

?

?

CagSas

?

Sas

Sas

Nuo

Ogl

Cag

Sas

Ogl

??

?

Sas

Cag

Ogl

?

?

?

Ogl

Ori Cag

Ogl

?

?

Ogl?

Cag

?

?

? Ogl

?

Sas

?

Nuo

Ogl

Cam

Cag

Car

Sas

Ogl

?

Ogl

?

?

Ogl

?

?

?

? Sas

Ogl

Cag

Cag

Ogl

?

Ogl

?Cam

?

Sas

Nuo

?

?Ogl

Nuo

Ogl

Sas

?

Nuo

Cag

Ogl

Cag

?

?

Ogl

Sas

?

CagOgl

Nuo

?

Ogl

Cag

Olb

?

Cam

Ogl

Sas Cag

?

Cag

Ori

?

?

Cam

Cam

Ogl

Cag

Ogl

Ori

Ogl

Cag?

Cam

Ogl

?

?

?

Cam

?

Cag

Ogl

Ogl

?

Ogl

Ogl

Ogl

?

Ogl

Cam

OglOglOgl

Olb

Ogl

?

?

Sas

Cag

Cag

aaa

Cam

Ogl

?

Ogl

?

Cag

Ogl

NOE

LON

LON

MSR

SEC

VIL

S12

NOE

raaaraF aaaaSUAMA1

SIDMA11

SSMA1A1

SSMA

OO

ORC

SUC

SUC

SSC

SSSSRCSS

rFrFF araaararararrraraFFF aaararaarra

lC

OO

Cal

bIbORC

MA1

MA8

FIL

SUC

SUC

COR

MSR

MA7

SEC

SUAUAAAAAAAUAAAAU

FFFF

alal

OCC

C?

bgS

b

g?

ag?

S

Ca C?

C?

CCCCCCCCCCCCC

BBBBBBBBBBBBBBBBBBBBBBBBGGBGGBBBBORC

MA1

SUA

MA8SUC

VIL

AMC

COR

AMC

S12

AMC

ORC

VIL

SUA

CCCCC SOO

SSSS

OOg

SEC

SUC

SUC

VIL

SUC

SID

ORCCCCCCCSSS

CCCCCCOOSE

DSIDSIDCDD

CDDOSSSOSS

CCDD

CD

CCDDDI CCORRCIDIDII CSI

CIIC

OSSOSSCCCCCCCC

GGGGBBBBBBBGGBBBBBGBGGGBBBBGBBG

CCCCCCCCCCCCCC

OOE

ORC

ORC

PJU

ISC

SUAORC

RCORRCRCRCORORRRC

ON

O CCOORORRR CCCCCCRRRCRCCC

A88MA88

??

??

MD MDBBBBBBGBBGGG

SSCRCCC

MM CC

SS

CCCCCSUSUSSSSSSSCCCCCCCCC

SSSUUCCCUUSUS

CCCUUSSSUSSSS

CCCCSUCCCCCCCCCCCCCCCCCCCCCCCCCCCCCCC

RCCC

ORC

MA7

SNN

SNN

S12

MA8

MA8

SUCSEC

ISB

NN

CCRCC

NNNN

CRRCC

OOglOUCSUCUC

OglOglCCC

OOOSOO

CCCCC

SSSS SSSSSISCSSSSSUSUUCUSSSISUUUCUCUCUCCSSSSSSSSSSISUCUUCCUCC

ISS

SSEESUSUSUU

SEEUUUU

ISSSSSSS CCCCCCCCCCS CUCCCCCCCCCC

A8888

ISSSSSBECCUCCUCCSEUCUCUCCSS

S12

SEC

SNN

SUAORC

CagCar

NuoOgl

Olb

OriSas

Cam

ba

PCA from Fig 2with focus onSidore et al. 2015samples

PCA ofSidore et al. 2015samples* withancients projected

Ogl

Fig. 5 Present-day genetic structure in Sardinia reanalyzed with aDNA. a Scatter plot of the first two principal components from Fig. 2a with a zoom-inon present-day Sardinia diversity in our sample33. Median PC values for each Sardinian region are depicted as large circles. b PCA results based on present-day Sardinian individuals, subsampling Cagliari and Ogliastra to 100 individuals to avoid effects of unbalanced sampling. In both panels, each individual islabeled with an abbreviation that denotes the source location if at least three grandparents were born in the same geographical location (“small” three-letter abbreviations) or if grand-parental ancestry is missing with question mark. We also projected each ancient Sardinian individual on to the top two PCs(points color-coded by age, see Fig. 1 for the color scale).

ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6

8 NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 | www.nature.com/naturecommunications

www.nature.com/naturecommunications

haplogroups, but have not been detected in Neolithic Anatoliansor more western WHG individuals. These results are plausibleoutcomes of substantial gene flow from Neolithic populationsthat spread westward along the Mediterranean coast of southernEurope around 5500 BCE (a “Cardial/Impressed” ware expan-sion, see Introduction). We note that we lack autosomal aDNAfrom earlier than the Middle Neolithic in Sardinia and from keymainland locations such as Italy, which leaves some uncertaintyabout timing and the relative influence of gene flow from theItalian mainland versus from the north or west. The inferredWHG admixture fraction of Middle Neolithic Sardinians washigher than that of early mainland EEF populations, which couldsuggest a time lag of the influx into Sardinia (as HG ancestryincreased through time on the mainland) but also could resultfrom a pulse of initial local admixture or continued gene flowwith the mainland. Genome-wide data from Mesolithic and earlyNeolithic individuals from Sardinia and potential source popu-lations will help settle these questions.

From the Middle Neolithic onward until the beginning of thefirst millennium BC, we do not find evidence for gene flow fromdistinct ancestries into Sardinia. That stability contrasts withmany other parts of Europe which had experienced substantialgene flow from central Eurasian Steppe ancestry starting about3000 BCE11,12 and also with many earlier Neolithic and Copperage populations across mainland Europe, where local admixtureincreased WHG ancestry substantially over time10. We observedremarkable constancy of WHG ancestry (close to 17%) from theMiddle Neolithic to the Nuragic period. While we cannot excludeinflux from genetically similar populations (e.g., early Iberian BellBeakers), the absence of Steppe ancestry suggests genetic isolationfrom many Bronze Age mainland populations—including laterIberian Bell Beakers13. As further support, the Y haplogroup R1b-M269, the most frequent present-day western European hap-logroup and associated with expansions that brought Steppeancestry into Britain13 and Iberia14 about 2500–2000 BCE,remains absent in our Sardinian sample through the Nuragicperiod (1200–1000 BCE). Larger sample sizes from Sardinia andalternate source populations may discover more subtle forms ofadmixture, but the evidence appears strong that Sardinia wasisolated from major mainland Bronze Age gene flow eventsthrough to the local Nuragic period. As the archeological recordshows that Sardinia was part of a broad Mediterranean tradenetwork during this period19, such trade was either not coupledwith gene flow or was only among proximal populations ofsimilar genetic ancestry. In particular, we find that the Nuragicperiod is not marked by shifts in ancestry, arguing againsthypotheses that the design of the Nuragic stone towers wasbrought with an influx of people from eastern sources such asMycenaeans.

Following the Nuragic period, we found evidence of gene flowwith both northern and eastern Mediterranean sources. Weobserved eastern Mediterranean ancestry appearing first in twoPhoenician-Punic sites (Monte Sirai, Villamar). The northernMediterranean ancestry became prevalent later, exemplified mostclearly by individuals from a north-western Medieval site (SanNicola Necropoli). Many of the post-Nuragic individuals could bemodeled as direct immigrants or offspring from new arrivals toSardinia, while others had higher fractions of local Nuragicancestry (Corona Moltana, ORC002). Substantial uncertaintyexists here as the low differentiation among plausible sourcepopulations makes it challenging to exclude alternate models,especially when using individual-level analysis. Overall though,we find support for increased variation in ancestry after theNuragic period, and this echoes other recent aDNA studies in theMediterranean that have observed fine-scale local heterogeneityin the Iron Age and later14,52–54.

In addition, we found present-day Sardinian individuals sitwithin the broad range of ancestry observed in our ancientsamples. A similar pattern is seen in Iberia14 and central Italy54,where variation in individual ancestry increased markedly in theIron Age, and later decreased until present-day. In terms of thefine-scale structure within Sardinia, we note the median positionof modern individuals from the central regions of Ogliastra andNuoro on the main PCA (Fig. 5a) are less shifted towards novelsources of post-Nuragic admixture, which reinforces a previousresult that Ogliastra shows higher levels of EEF and HG ancestrythan other regions32. At the same time, in the PCA of withinSardinia variation (Fig. 5b), differentiation of Ogliastra fromother regions and other ancient individuals is apparent, likelyreflecting a recent history of isolation and drift. The northernprovinces of Olbia-Tempio, and to a lesser degree Sassari, appearto have received more northern Mediterranean immigration afterthe Bronze Age; while the southwestern provinces of Campidanoand Carbonia carry more eastern Mediterranean ancestry. Both ofthese results align with known history: the major Phoenician andPunic settlements in the first millennium BCE were situatedprincipally along the south and west coasts, and Corsican shep-herds, speaking an Italian-Corsican dialect (Gallurese), immi-grated to the northeastern part of Sardinia55.

Our inference of gene flow after the second millennium BCEseems to contradict previous models emphasizing Sardinianisolation12. These models were supported by admixture tests thatfailed to detect substantial admixture32, likely because of sub-stantial drift and a lack of a suitable proxy for the NuragicSardinian ancestry component. However, compared with otherEuropean populations50,56, we confirm Sardinia experiencedrelative genetic isolation through the Bronze Age/Nuragic per-iod. In addition, we find that subsequent admixture appears toderive mainly from Mediterranean sources that have relativelylittle Steppe ancestry. Consequently, present-day Sardinianindividuals have retained an exceptionally high degree of EEFancestry and so they still cluster with several mainland EuropeanCopper Age individuals such as Ötzi2, even as they are shiftedfrom ancient Sardinian individuals of a similar time period(Fig. 2).

The Basque people, another population high in EEF ancestry,were previously suggested to share a genetic connection withmodern Sardinian individuals32,57. We observed a similar signal,with modern Basque having, of all modern samples, the largestpairwise outgroup-f3 with most ancient and modern Sardiniangroups (Fig. 3). While both populations have received someimmigration, seemingly from different sources (e.g., Fig. 4,ref. 14), our results support that the shared EEF ancestry com-ponent could explain their genetic affinity despite their geo-graphic separation.

Beyond our focal interest in Sardinia, the results from indivi-duals from the Phoenician-Punic sites Monte Sirai and Villamarshed some light on the ancestry of a historically impactfulMediterranean population. Notably, they show strong geneticrelationships to ancient North-African and eastern Mediterra-nean sources. These results mirror other emerging ancient DNAstudies37,58, and are not unexpected given that the Punic center ofCarthage on the North-African coast itself has roots in the easternMediterranean. Interestingly, the Monte Sirai individuals, pre-dating the Villamar individuals by several centuries, show lessNorth-African ancestry. This could be because they harbor earlierPhoenician ancestry and North-African admixture may havebeen unique to the later Punic context, or because they wereindividuals from a different ancestral background altogether.Estimated North-African admixture fractions were much lower inlater ancient individuals and present-day Sardinian individuals, inline with previous studies that have observed small but significant

NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-14523-6 ARTICLE

NATURE COMMUNICATIONS | (2020) 11:939 | https://doi.org/10.1038/s41467-020-14523-6 |www.nature.com/naturecommunications 9

www.nature.com/naturecommunicationswww.nature.com/naturecommunications

African admixture in several present-day South Europeanpopulations, including Sardinia32,59,60.

As ancient DNA studies grow, a key challenge will be fine-scalesampling to aid the interpretation of shifts in ancestry. Oursample from Sardinia’s post-Nuragic period highlights the com-plexity, as we simultaneously observe examples of individuals thatappear as novel immigrant ancestries (e.g., from Villamar and SanNicola) and of individuals that look more contiguous to the pastand to the present (e.g., the two Corona Moltana siblings, theORC002 individual, several of the Alghero Monte Carru indivi-duals). This variation is likely driven by differential patterns ofcontact—as might arise between coastal versus interior villages,central trading centers versus remote agricultural sites, or evenbetween neighborhoods and social strata in the same village. Wealso note that modern populations are collected with differentbiases than ancient individuals (e.g., the sub-populations sampledby medical genetics projects33 versus the sub-populations that areaccessible at archeological sites). As such, caution should beexercised when generalizing from the sparse sampling typical formany aDNA studies, including this one.

With these caveats in mind, we find that genome-wide ancientDNA provides unique insights into the population history ofSardinia. Our results are consistent with gene flow being minimalor only with genetically similar populations from the MiddleNeolithic until the late Bronze Age. In particular, the onset of theNuragic period was not characterized by influx of a distinctancestry. The data also link Sardinia from the Iron Age onwardsto the broader Mediterranean in what seems to have been aperiod of new dynamic contact throughout much of the Medi-terranean. A parallel study focusing on islands of the westernMediterranean provides generally consistent results and bothstudies make clear the need to add complexity to simple modelsof sustained isolation that have dominated the genetic literatureon Sardinia52. Finally, our results suggest some of the currentsubstructure seen on the island (e.g., Ogliastra) has emerged dueto recent genetic drift. Overall, the history of isolation, migration,and genetic drift on the island has given rise to an unique con-stellation of allele frequencies, and illuminating this history willhelp future efforts to understand genetic-disease variants pre-valent in Sardinia and throughout the Mediterranean, such asthose underlying beta-thalassemia and G6PD deficiency.

MethodsArcheological sampling. The archeological samples used in this project derivefrom several collection avenues. The first was a sampling effort led by co-authorLuca Lai, leveraging a broad base of samples from different existing collections inSardinia, a subset of which were previously used in isotopic analyses to understanddietary composition and change in prehistoric Sardinia40. The second was from theSeulo Caves project41, an on-going project on a series of caves that span the MiddleNeolithic to late Bronze Age near the town of Seulo. The project focuses on thediverse forms and uses of caves in the prehistoric culture of Sardinia. The Neolithicindividuals from Sassari province as well as the post-Nuragic individuals werecollected from several co-authors as indicated in Supplementary InformationSection 1. The third was a pair of Neolithic sites Noedalle and S’isterridolzu42. Thefourth are a collection of post-Nuragic sites spanning from the Phoenician to theMedieval time. All samples were handled in collaboration with local scientists andwith the approval of the local Sardinian authorities for the handling of arche-ological samples (Ministero per i Beni e le Attività Culturali, Direzione Generaleper i beni Archeologici, request dated 11 August 2009; Soprintendenza per i BeniArcheologici per le province di Sassari e Nuoro, prot. 12993 dated 20 December2012; Soprintendenza per i Beni Archeologici per le province di Sassari e Nuoro,prot. 10831 dated 27 October 2014; Soprintendenza per i Beni Archeologici per leprovince di Sassari e Nuoro, prot. 12278 dated 05 December 2014; Soprintendenzaper i Beni Archeologici per le Provincie di Cagliari e Oristano, prot. 62, dated 08January 2015; Soprintendenza Archeologia, Belle Arti e Paesaggio per le Provinciedi Sassari, Olbia-Tempio e Nuoro, prot. 4247 dated 14 March 2017; Soprinten-denza per i Beni Archeologici per le Provincie di Sassari e Nuoro, prot. 12930 dated30 December 2014; Soprintendenza Archeologia, Belle arti e Paesaggio per leProvincie di Sassari e Nuoro, prot. 7378 dated 9 May, 2017; Soprintendenza per iBeni Archeologici per le Provincie di Cagliari e Oristano, prot. 20587, dated 05October 2017; Soprintendenza Archeologia, Belle Arti e Paesaggio per le Provincie

di Sassari e Nuoro, prot. 15796 dated 25 October, 2017; Soprintendenza Arche-ologia, Belle Arti e Paesaggio per le Provincie di Sassari e Nuoro, prot. 16258 dated26 November 2017; Soprintendenza per i Beni Archeologici per le province diSassari e Nuoro, prot. 5833 dated 16 May 2018; Soprintendenza Archeologia, BelleArti e Paesaggio per la città metropolitana di Cagliari e le province di Oristano eSud Sardegna, prot. 30918 dated 10 December 2019). For more detailed descriptionof the sites, see Supplementary Information Section 1.

Initial sample screening and sequencing. The ancient DNA (aDNA) workflowwas implemented in dedicated facilities at the Palaeogenetic Laboratory of theUniversity of Tübingen and at the Department of Archaeogenetics of the MaxPlanck Institute for the Science of Human History in Jena. The only exception wasfor four samples from the Seulo Cave Project which had DNA isolated at theAustralian Centre for Ancient DNA and capture and sequencing carried out in theReich lab at Harvard University. Different skeletal elements were sampled using adentist drill to generate bone and tooth powder, respectively. DNA was extractedfollowing an established aDNA protocol61 and then converted into double-stranded libraries retaining62 or partially reducing63 the typical aDNA substitutionpattern resulting from deaminated cytosines that accumulate towards the mole-cule’s termini. After indexing PCR62 and differential amplification cycles, the DNAwas shotgun sequenced on Illumina platforms. Samples showing sufficient aDNApreservation where captured for mtDNA and ≈1.24 million SNPs across thehuman genome chosen to intersect with the Affymetrix Human Origins array andIllumina 610-Quad array47. The resulting enriched libraries were also sequenced onIllumina machines in single-end or paired-end mode. Sequenced data were pre-processed using the EAGER pipeline64. Specifically, DNA adapters were trimmedusing AdapterRemoval v265 and paired-end sequenced libraries were merged.Sequence alignment to the mtDNA (RSRS) and nuclear (hg19) reference genomeswas performed with BWA66 (parameters –n 0.01, seeding disabled), duplicateswere removed with DeDup64 and a mapping quality filter was applied (MQ≥30).For genetic sexing, we compared relative X and Y-chromosome coverage to theautosomal coverage with a custom script. For males, nuclear contamination levelswere estimated based on heterozygosity on the X-chromosome with the softwareANGSD67.

After applying several standard ancient DNA quality control metrics, retainingindividuals with endogenous DNA content in shotgun sequencing >0.2%, mtDNAcontamination

in our study. We processed these samples through the same pipeline and filtersdescribed above, resulting in a reference dataset of 972 ancient samples.Throughout our analysis, we used a subset of n= 1,088,482 variant