Ancient DNA reveals the origins of the Albanians

new

Admin
Administrator

Posts: 73,167

Ancient DNA reveals the origins of the Albanians Jun 13, 2023 20:23:42 GMT

Quote

Post by Admin on Jun 13, 2023 20:23:42 GMT

Ancient DNA reveals the origins of the Albanians
Leonidas-Romanos Davranoglou, Aris Aristodemou, David Wesolowski, Alexandros Heraclides
doi: doi.org/10.1101/2023.06.05.543790

Abstract
The origins of the Albanian people have vexed linguists and historians for centuries, as Albanians first appear in the historical record in the 11th century CE, while their language is one of the most enigmatic branches of the Indo-European family. To identify the populations that contributed to the ancestry of Albanians, we undertake a genomic transect of the Balkans over the last 8000 years, where we analyse more than 6000 previously published ancient genomes using state-of-the-art bioinformatics tools and algorithms that quantify spatiotemporal human mobility. We find that modern Albanians descend from Roman era western Balkan populations, with additional admixture from Slavic-related groups. Remarkably, Albanian paternal ancestry shows continuity from Bronze Age Balkan populations, including those known as Illyrians. Our results provide an unprecedented understanding of the historical and demographic processes that led to the formation of modern Albanians and help locate the area where the Albanian language developed.

Introduction
During the Iron Age (1100 BCE–150 CE), the Balkans were characterised by remarkable cultural, linguistic, and genetic heterogeneity (1–6). In the western Balkans, “Celtic” cultures such as Hallstatt and La Tène, interacted for centuries with local groups referred to as the “Illyrians” and “Dalmatians” (7, 8). Deep in the Balkan heartland, heterogenous populations named by classical authors as “Dacians”, “Dardanians”, “Moesians”, and “Paeonians” (1, 7–9) bordered nomadic cultures from the Pontic-Caspian steppe known as the “Scythians” (10), while the southeastern part of the peninsula was inhabited by “Thracians” and “Greeks” (2, 11). Balkan peoples also expanded beyond the confines of the peninsula, with “Messapians” migrating to southeast Italy at least since 600 BCE (12). The linguistic and cultural diversity of the Balkans was considerably homogenised during the Hellenistic (2, 9), Roman (9, 13–15), and especially the Migration Period, when Germanic and Slavic-speaking groups massively settled in the region (2, 3, 16, 17). These events ultimately led to the extinction of all palaeo-Balkan languages except Greek and Albanian. The latter is one of the most enigmatic branches of the Indo-European language family, having vexed linguists for more than two centuries (2, 18, 19).

Tracing the origins of Albanians and their language is challenging for several reasons. Only a handful of historical sources comment on the ethnic and linguistic composition of the southwest Balkans during the transition from classical antiquity to Medieval times (500-1000 CE) (20–22), and none of them mention an Albanian-speaking population from the territory of modern Albania. Speakers of Slavic languages are reported to have inhabited what is now southern Albania in the 8th century CE (15, 21), where the frequency of Slavic toponyms also peaks (23), while the same region is characterised by the presence of Greek-speakers at least since the Medieval period (24). The urbanised Medieval populations of the northwest, referred to by contemporary historians as the Romani/Ῥωμᾶνοι (20), are thought to have spoken a variant of vulgar Latin known as West Balkan Romance (14, 15, 25) that persisted at least until the 13th century CE (26). The demographic and linguistic situation in the mountainous interior is unknown, and it is only in the 11th century CE that Albanians appear in the historical record (22), while the earliest surviving written document of their language dates to 1462 CE (27).

A number of linguistic hypotheses have attempted to identify the affinities of the Albanian language and to locate the region where it developed, yet no definitive conclusions have been drawn. The most prominent, mutually exclusive hypotheses can be divided into those arguing for a local west Balkan origin from an Illyrian (28, 29) or Messapic background (19, 30, 31) [which may or may not have been distinct languages (7, 30, 32)], and those proposing a non-local origin from a Daco-Moesian-Thracian background (2, 19, 33) or an unattested Balkan language, whose speakers entered Albania from the central-east Balkans sometime after 400 CE (15, 32, 34, 35). The validity of these hypotheses, although hotly debated, is hard to test, as these ancient languages are poorly recorded, being known only from fragmentary inscriptions, toponyms, and a handful of historical sources (2, 7, 36). Furthermore, all of the ethnonyms of ancient Balkan peoples, such as “Illyrian” and “Thracian”, are likely artificial labels that were coined by ancient and modern authors (37), and may include several related languages with largely obscure geographical limits, intelligibility, and emic identities of their speakers (8, 9, 32). The most recent linguistic hypotheses propose a sister-group relationship of Albanian to Greek or to the Greek-Armenian clade (18, 38, 39), which firmly places the origin of the language in the Balkans but does not pinpoint the location of the proto-Albanian homeland within the peninsula and its potential affiliation to historically attested populations.

Archaeological data on Albania’s Medieval cultures are also inconclusive, especially for the Komani-Kruja complex (ca. 600-800 CE), which has been interpreted as the cultural expression of either a Romanised population (local or intrusive) (8, 40, 41) or an indigenous Albanian-speaking group (42).

Due to the challenges associated with linking archaeological, literary, and linguistic evidence, an archaeogenetic approach may offer novel insights into the origin of the Albanians, their biological relationships to ancient people, and the affinities of their language. Although gene flow is not always accompanied by language shifts [as in the case of Basque (43) and Etruscan (44)], migration is one of the primary vectors of cultural change (45, 46), of which language dissemination is a frequent outcome (3, 17, 47–49). Recent years have witnessed a surge in the palaeogenomic sampling of the Balkan peninsula (3–6, 11, 16), yet the resulting datasets have not been mined to help us understand how migration led to the emergence and spread of new material cultures, communities, and languages in the territory of modern Albania.

www.biorxiv.org/content/10.1101/2023.06.05.543790v1.full

Admin
Administrator

Posts: 73,167

Ancient DNA reveals the origins of the Albanians Jun 14, 2023 21:26:36 GMT

Quote

Post by Admin on Jun 14, 2023 21:26:36 GMT

Figure 1.
Geographical distribution and dating of the examined Balkan populations. A) Archaeological sites annotated by period. B) Analysed ancient individuals from the territory of modern Albania arranged by archaeological site and date (radiocarbon or archaeological chronology).

Results
Complex genetic transformations during the Balkan Stone to Bronze Age transition
To visualise variation in genetic composition, we perform principal components analysis (PCA), where we project previously published ancient individuals from the Balkans and adjacent regions onto present day West Eurasian populations genotyped with the Human Origins array (Fig. S1). Three distinct population clusters can be seen on PC1 and PC2 of our Stone Age-Early Bronze Age PCA (Fig. 2). Hunter-Gatherer (HG) populations, which include Western Hunter Gatherers and Iron Gates Hunter-Gatherers (IGHG), occupy the right side of PC1; Early European Farmers (EEF), who introduced agriculture to Europe, cluster on the bottom left corner of PC2; and Yamnaya steppe pastoralists, who are attributed with the expansion of Indo-European languages to Eurasia and beyond, plot at the uppermost space between PC1 and PC2.

Admin
Administrator

Posts: 73,167

Ancient DNA reveals the origins of the Albanians Jun 15, 2023 20:44:21 GMT

Quote

Post by Admin on Jun 15, 2023 20:44:21 GMT

Figure 2.
PCA of modern West Eurasian samples (grey points) with projection of ancient Neolithic-Early Bronze Age individuals from the Balkans. Dashed lines indicate the two poles of observed ancestry shifts from the Neolithic onwards – one generated by the assimilation of hunter-gatherers (blue line), and another by admixture with incoming Yamnaya-related populations from the Pontic-Caspian steppe (yellow line). Samples clustering within the PCA space enclosed by HG, EEF, and Yamnaya derive variable amounts of ancestry from these populations.

The PCA shows that the single Early Neolithic sample from Albania clusters with contemporary farming populations from Greece and Anatolia (Fig. 2), an affinity that is further supported by outgroup f3-statistics (Fig. S2A), which measure shared genetic drift between a pair of populations, as well as formal f4-statistics qpAdm tests with rotating sources (Table S3). No major genetic changes are observed in samples dating to the Neolithic-Chalcolithic transition (Fig. 2), with the individuals from Albania (Alb NChl) being modelled as deriving their entire ancestry from the preceding Albanian Neolithic, possibly with an additional 5% admixture from an IGHG-like source (Figs. 3A, S3; Table S4).

Figure 3.
Distal qpAdm models for Balkan populations from the Neolithic to post-Medieval period. A) Neolithic and Early Bronze Age population models. B) Bronze-Iron Age models. C) Roman-post-Roman models. Putative ancestral source populations are colour-coded and presented in the top part of the figure. The single model receiving the highest statistical support is presented (all possible models and the criteria used to select the ones shown here are provided in Tables S3-S18).

Unlike the genetic continuity observed during the Neolithic-Chalcolithic transition, the PCA indicates remarkable ancestry shifts in the Balkans during the Early Bronze Age (EBA) (ca. 2600-1800 BCE). Most EBA Balkan samples show a significant change in their PC position and plot towards the Yamnaya (Fig. 2), suggesting the arrival of the steppe cultural package in the region, including Indo-European speech (4, 11, 43, 47, 49, 51–53). The Albania Çinamak EBA sample, dated to 2663-2472 BCE (Fig. 1B; Table S1), together with two samples from Bulgaria Boyanovo EBA (2500-2000 BCE), cluster closest to Yamnaya (Fig. 2), and qpAdm tests model them as having roughly 70% and 80-85% steppe ancestry, respectively, with the remainder of their genome deriving from local farming populations (Fig. 3A; Tables S5, S7). Accordingly, both male samples from Albania Çinamak EBA and Bulgaria Boyanovo EBA carry Y-chromosome haplogroup R1b-M269 (Tables S22, S26), the primary indicator of paternal ancestry from the Pontic-Caspian steppe (6, 49, 53). Importantly, steppe-derived admixture in Albania Çinamak EBA is best modelled with largely unadmixed Yamnaya groups rather than with EEF-and-HG-admixed Corded Ware populations (Fig. S2C; Table S7), suggesting that Indo-European-related ancestry arrived in the Balkans directly from the Pontic-Caspian steppe, in accordance with previous studies (6). This is also supported by the presence of Y-chromosomal haplogroup R1b-Z2103 in the Bulgaria Boyanovo EBA sample, which has been found in the earliest Yamnaya contexts (6, 49, 53), whereas Corded Ware samples are characterised by largely different subclades of haplogroup R1 (R1a-M417, R1b-U106, R1b-L151) (6, 45, 54). The ancestors of Albania Çinamak EBA were likely recent arrivals from the steppe, as this individual lived during the transitional period where the first kurgan burials made their appearance in Albania (55). Contemporary EBA archaeological finds from tumuli in Shkodër (100 km to the west of Çinamak) include artefacts associated with the Vučedol culture (55), revealing links with the northern Adriatic and Pannonia. It should be noted that a male with high levels of steppe ancestry discovered in a Vučedol context in Croatia also belonged to Y-chromosomal haplogroup R1b-Z2103 (56).

Regarding the broader Balkan EBA context, early interactions between incoming nomadic pastoralists and local farming populations were likely complex, as evidenced by the heterogenous levels of steppe admixture among geographically adjacent groups, which range from 0-7% (Greece Manika EBA), 16% (Bulgaria Tell Kran EBA), 30% (Serbia Mokrin EBA Maros) to ca. 45% (Bulgaria Kapetan Andreevo EBA) (Fig. 3A; Table S5). These differences in ancestry are reflected in the PCA, where the abovementioned groups occupy a middle space between largely unadmixed Yamnaya and EEF populations (Fig. 2).

Admin
Administrator

Posts: 73,167

Ancient DNA reveals the origins of the Albanians Jun 16, 2023 19:53:38 GMT

Quote

Post by Admin on Jun 16, 2023 19:53:38 GMT

The establishment of the BA-IA Balkan genetic cline
We next examined the projections of Bronze Age (BA) and Iron Age (IA) Balkan populations upon the principal components generated earlier (Fig. 4). By the Middle-Late Bronze Age (MLBA) and the Iron Age, the genetic ancestry of Balkan populations had become less heterogenous, exhibiting a north-to-south cline that broadly reflects geography (Fig. 4A). Accordingly, the MLBA-IA samples from Çinamak in Albania are clinal between contemporary populations from Croatia and Montenegro on one side of PC1 and PC2, and from North Macedonia and northern Greece on the other side (Fig. 4B). This affinity is further corroborated by f3-statistics (Fig. S2D, E) and Mobest analysis (Fig. S4), which employs an algorithm for spatiotemporal mapping of genetic profiles using bulk aDNA data (57). Formal f4-models with qpAdm using ultimate sources indicate a uniform genetic profile and a resurgence of farmer ancestry across the Central-West Balkans and northern Greece compared with the preceding EBA, as most samples derive around 60% of their ancestry from EEFs, 0-5% from Iran N, and 30-40% from steppe populations (Fig. 3B; Table S5). In the southern end of the Balkan genetic cline, Bronze Age and Iron Age populations of central-southern Greece and Bulgaria have considerably higher EEF (75-80%) and Iran N (5-10%) ancestry, and a much lower proportion of steppe ancestry (15-20%) (Fig. 3B; Table S5).

Figure 4.
PCA of modern West Eurasian samples (grey circles) with projection of ancient Bronze Age and Iron Age individuals from the Balkans. A) PCA of all tested populations, showing a north-to-south cline that mirrors geography. B) Detail of region enclosed in the dotted rectangle in panel A, showing the clustering of populations from Albania (enclosed in solid line).

The close clustering of BA-IA populations from Albania, Croatia, Montenegro, North Macedonia, and northern Greece is also confirmed in proximate qpAdm models, as the Çinamak MLBA-IA samples derive most of their ancestry from the West Balkans (Tables S8-S9), with a possible 15-25% contribution from a southeast Balkan source (Bulgaria EIA, Greece BA Mycenaean) after the Middle Bronze Age (MBA) (Table S9). Based on the above, the MBA-IA populations of a large geographic region spanning northern Greece, North Macedonia and the entire Adriatic coast, including the region of modern Albania, form a uniform genetic cluster with similar admixture proportions (Fig. 3B) that persists for at least 1.500 years and transcends the linguistic boundaries identified by classical authors (7, 9). Our findings are further reinforced by IBD-sharing between certain samples from Albania and North Macedonia (Table S20) (50).

Intriguingly, two mercenaries from the Battle of Himera in Sicily (58) fall within the PCA cluster of BA-IA Albania (Fig, 4B). This position might be coincidental, as these two samples are characterised by high proportions of IGHG-related ancestry (9%), which in the Balkans can only be found in populations from EBA Serbia (Fig. 3B; Table S5). Furthermore, the proximate ancestry sources of these mercenaries are uncertain using both f3-and-f4-statistics (Fig. S5; Table S10), suggesting they might derive from a currently unsampled population, likely from the Central Balkans.

An outlier from Hellenistic North Macedonia might represent an early migrant from the Near East, as he is shifted towards populations from the Caucasus (Fig. 4A), and his ultimate ancestry is successfully modelled with proxies from the Caucasus and the Levant (Fig. 3B; Table S5).

Admin
Administrator

Posts: 73,167

Ancient DNA reveals the origins of the Albanians Jun 18, 2023 20:07:49 GMT

Quote

Post by Admin on Jun 18, 2023 20:07:49 GMT

Albania as a refugium of Iron Age ancestry during the rise and fall of the Roman Empire
It has been previously demonstrated (3, 16) that during Imperial Roman times, or perhaps even the preceding Hellenistic era, as we show here (Fig. 4B), large-scale westward immigration from the Eastern provinces took place, which transformed the genetic landscape of the Empire (59). This demographic shift also took place in the Balkans (3), as can be seen in the PCA of Roman and Post-Roman samples (Fig. 5), where virtually all local populations are shifted from their Bronze Age and Iron Age position towards the Eastern Mediterranean, forming a cluster along what we term as the Anatolian cline (Fig. 5A).

Figure 5.
PCA of modern West Eurasian samples (grey circles) with projection of ancient Roman, Medieval, and post-Medieval individuals from the Balkans. A) PCA of tested populations, highlighting the shift in the PCA coordinates of Roman and post-Roman groups towards the Anatolian (green dashed line) and Balto-Slavic (blue dashed line) clines, compared to their counterparts from the same regions in the preceding Bronze Age and Iron Age (faded grey polygons). B) Detail of region in panel A, showing the tight clustering of populations from Albania over the past 3000 years, including modern individuals (orange polygon). For simplicity, Roman-era populations within and close to this cluster have been removed and can be seen in the previous panel.

Formal f4-statistical tests using qpAdm also support the divergence in Balkan populations between the BA/IA and the Hellenistic/Roman era, by modelling post-Neolithic Anatolian admixture with an Iran N-like reference (Fig. 3C), which became pervasive in Anatolia and the Aegean by the EBA (60), and was also often accompanied by Levantine ancestry (59). Although largely absent in the preceding Iron Age, Iran N-like admixture entered Balkan populations in a multiphased manner, with some Roman-era-Early Medieval populations being modelled as deriving 0-15% of their ancestry from this source (Croatia Roman Beli Manastir and Šćitarjevo, Montenegro Doclea Roman), which in adjacent regions can be as high as 20-30% (Croatia Novo Selo Bunje, Zadar, and Trogir Dragulin) (Fig. 3C; Table S5). These differential ancestry shifts are also corroborated using ADMIXTURE on a subset of our dataset, represented by the appearance of a West-Central Asian component that was absent prior to the Roman period (Fig. S3).

While Anatolian ancestry was becoming established in the Balkans, a second, even more significant demographic transformation began to take place. By 200 CE, individuals related to north-eastern European, Balto-Slavic, and nomadic steppe populations started appearing in the Balkans (Fig. 5A). Such movements peaked during the Migration Period (roughly 350-600 CE) with the mass migration and settlement of Avar, Germanic and Slavic-speaking groups, which transformed the cultural, linguistic, and ethnic composition of the region (3, 9, 17, 61). Admixture with these newcomers was pervasive, as can be seen in the PCA, where most Balkan populations show a significant shift towards Balto-Slavic groups, along what we term the Balto-Slavic cline (Fig. 5A). This admixture event is mirrored in an unprecedented increase of IGHG-related ancestry in most late Roman-Early Medieval Balkan populations, which is most significant in samples located within the Balto-Slavic cline (Fig. 3C; Table S5). Both qpAdm and ADMIXTURE models suggest that from an average of 4-5% in the preceding BA-IA and the early Roman era, IGHG ancestry rose to 13-18% in strongly Balto-Slavic-shifted populations from late Roman, Medieval, and Post-Medieval Croatia, Macedonia, Montenegro, and Serbia (Fig. 3C; Fig. S3; Table S5).