population structure of early neolithic population in the South Caucasus

[bicicleur 2]

Despite the localisation of the southern Caucasus at the outskirt of the Fertile Crescent, the Neolithisation process started there only at the beginning of the sixth millennium with the Shomutepe-Shulaveri culture of yet unclear origins. We present here genomic data for three new individuals from Mentesh Tepe in Azerbaijan, dating back to the beginnings of the Shomutepe-Shulaveri culture. We evidence that two juveniles, buried embracing each other, were brothers. We show that the Mentesh Tepe Neolithic population is the product of a recent gene flow between the Anatolian farmer-related population and the Caucasus/Iranian population, demonstrating that population admixture was at the core of the development of agriculture in the South Caucasus. By comparing Bronze Age individuals from the South Caucasus with Neolithic individuals from the same region, including Mentesh Tepe, we evidence that gene flows between Pontic Steppe populations and Mentesh Tepe-related groups contributed to the makeup of the Late Bronze Age and modern Caucasian populations. Our results show that the high cultural diversity during the Neolithic period of the South Caucasus deserves close genetic analysis.


https://www.nature.com/articles/s42...vNIcu1ylIZyf-eOiJ5dBGuLJIOTrTpVctA72xQkuC5cVI

 

[Jovialis]

Guarino-Vignon et al. 2023

The authors used a variety of techniques to study the genetic relationships between Bronze Age populations from the South Caucasus and other parts of Western Eurasia. They found that all Bronze Age individuals from Armenia plot together and are shifted toward the Steppe cluster. In the ADMIXTURE analysis, they all exhibit a red component, absent in the Neolithic Mentesh Tepe individuals but maximised in Steppe populations and present, also, in CHG individuals. Interestingly, individuals from Chalcolithic Armenia (from Areni-1 cave, four of whom are directly dated by C14) do carry this Steppe/CHG component, whereas a Chalcolithic individual from Alkhantepe in Azerbaijan does not. D-statistics of the form D(Mbuti, Steppe Eneolithic; Mentesh Tepe, South Caucasus Bronze Age) are almost all significantly positive (Z-score: +2.1 to +5.

References

Guarino-Vignon et al. 2023
Section 2, paragraphs [3], [7], [8], [11]
Section 3, paragraphs [2], [4], [6]
Section 4, paragraph [6]
Section 5, paragraphs [2], [4]

 

[Jovialis]

[h=2]Untitled[/h]

• Neolithisation took place between 9000 and 7000 BCE. In some places, the Neolithic gained ground through the acculturation of local hunter-gatherers. Assimilation processes took place with a degree of admixture.
• The mechanism of Neolithisation in the South Caucasus remains poorly understood. Mesolithic sites are known at Damjili Cave, unit 5 (Western Azerbaijan) and Kotias Klde Cave (Western Georgia).
• The first settlements attributed to the Early Neolithic period belong to an aceramic culture. Evidence of agriculture and herding remains scarce.
• Shomutepe-Shulaveri culture is the most ancient Caucasus culture with a complete Neolithic package. Found in several clusters of settlements in the northern foothills of the lesser Caucasus, the SSC is characterised by circular mud-brick houses.
• The origins of the SSC are still discussed 10,11. Due to the rapid transition from the aceramic stage to the SSC, population continuity during the Neolithisation process is possible. Several cultural and biological features are nonlocal.
• Mentesh Tepe is located in the Tovuz district of western Azerbaijan and has been excavated between 2007 and 2015. The botanical assemblage is dominated by cereals, especially barley, naked wheat and emmer. Animal remains consist largely of domesticated ones and wild animals are rare.
• Mentesh Tepe is exceptional for the discovery of a collective burial containing around 30 individuals. The number of individuals in the burial, as well as their sex and age bias, suggest a dramatic event such as an epidemic, a famine, or a sudden episode.
• There is no specific orientation or position of the corpses, but some intentional arrangements are visible. The most striking is formed by two juveniles embracing each other. Such an arrangement is rare, but other examples have been found in Neolithic and Protohistoric times.
• Paleogenetic studies of individuals found in the collective burial.

[h=2]Results[/h]

• DNA isolation and sequencing. From the 30 individuals from Structure 342 of the site, we sampled 23 petrous bones. We could only obtain genome-wide data for one female (Individual 7, later called MT7) and two males.
• Mentesh Tepe individuals carry three main components: ca. 30% Neolithic Iran (Iran_N; green), 15% Levant Neolithic (PPN; pale rose) and 55% blue and pink components shared with Anatolian or European Neolithic populations.
• To test for the genetic affinity with earlier populations in Western Eurasia, we performed D-statistics of the form D(Mbuti, Y; Z, MT) The statistic deviates significantly from zero if the pair of Anatolian/ Caucasian/Mesopotamian groups (Z) and Mentesh (MT) do not have the same genetic relation to Western European populations.
• When performing the same D-statistic (Fig. 2c) with MT7, MT23 and MTT001 reunited within a single Mentesh Tepe group (but excluding MT26, as he is related to MT23-cf. infra). , we observe that they share more alleles with the Early European Farmers (Greece_N, Serbia_EN), Anatolian farmers, Anatolian Epipaleolithic individual (Anatolia EP) and Neolithic Levant (PPN). They also deviate from the Neolithic Anatolian populations.
• Using qpAdm 34, we could successfully model the Mentesh Tepe group and each of the individuals as a two-way admixture of TellKurdu_LN and Iran_N. We also modelled Mentesh as a mixture between a Neolithic Anatolian group and a North Mesopotamian group.
• We used DATES 35 to estimate the date of the admixture event. We found that the admixture between the Anatolian source and the Iranian source only took place 15 ± 5 generations before Mentesh Tepe occupancy. With 28 years per generation, this dates the admixture event around 6300 BC.
• We explored the transitions in the genetic structure of Bronze Age populations from the South-Caucasus. Sadly, no ancient DNA could be retrieved from the Mentesh Tepe Chalcolithic levels.
• The PCA shows that all the Bronze Age individuals from Armenia plot together and are shifted toward the Steppe cluster. They all exhibit a red component, absent in the Neolithic Mentesh Tepe individuals but maximised in Steppe populations.
• We used qpAdm with the rotating method 37 and modelled the Chalcolithic and Bronze Age populations found in Armenia. The only fitting model for Areni-1 cave (Chalcolithic Armenia) is an admixture between 25% Steppe and 75% Mentesh.
• Next, genetic relatedness analyses were performed to further investigate potential close familial ties within individuals of structure 342.
• Although READ can infer relatedness from extremely lowcoverage data, a limitation of this method is that it generally requires a cohort of input individuals. We extracted the genotype of the Neolithic MTT001, along with 21 previously published Anatolian individuals from the Late Chalcolithic/Early Bronze-Age Arslantepe (ART) and Tell Kurdu from both Neolithic and Early Chalcolithic period 23.
• Using this approach, we detected elevated levels of genetic relatedness between individuals MT23-MT26. This value equates to P̅ 0 734 ± 0.024 when normalised over the median of ART individuals.
• us of individuals MT7 and MTT001. . The pairs MT23-MT7 and MT26-MT7 and MT26-MT7 were found to have close genetic ties.

[h=2]Discussion[/h]

• Our genetic kinship analysis results regarding pair MT23-MT26 unequivocally indicate that these two individuals share a first order of relatedness. This observation, combined with their close age of death, their shared mitochondrial haplogroups, and the context in which the remains of these two individuals were discovered, is strongly consistent with the hypothesis that they are siblings.
• Mentesh Tepe represents a highly homogeneous population or an extended family uniting people related by more than 3°or more. On a regional scale, a certain degree of homogeneity is observed.
• The genetic data from Mentesh Tepe shows that this SSC site shared one ancestry probably inherited from the Mesolithic Caucasian Hunter-Gatherers, one from the early Anatolian farmers and one from North-eastern Mesopotamia.
• Mentesh Tepe appears as a site that closely followed the process of migration and admixture of Anatolian farmers and of Mesopotamian populations.
• The genetic data from Mentesh Tepe also helps understanding better the subsequent periods.
• The North Caucasus Chalcolithic individual used as a proxy for the Chalcolithic and Neolithic Caucasus shows a genetic profile that differs from the Mentesh Tepe one.
• Chalcolithic and Late Bronze Age groups in the South Caucasus can be modelled as a mixture of Mentesh Tepe-related populations and the Steppe population. Steppe ancestry is not spread homogeneously during the Chalcolithic period.
• An archaic ancestry associated with CHG re-emerges during the Early Bronze Age/Kura-Araxes period. We also observed some Steppe ancestry in two Kura-Araxes groups.
• Analysis of ancient DNA from the Mentesh Tepe Neolithic funerary pit reveals an uneven genetic homogeneity inside this Shomu-Shulaveri culture archaeological site. The Shomu-Shulaveri culture took place between two intense population events.

[h=2]Methods[/h]

• Archaeological sampling. We obtained ethical authorisation from the Institute of Archaeology and Ethnography of Azerbaijan to study samples from Mentesh Tepe. Extraction was attempted on the 30 samples.
• Ancient DNA extraction was performed using a protocol adapted from ref. 46. 50-200 mg of petrous bone was powdered by drilling and incubated in lysis buffer for 14 h. After centrifugation, 1 ml of supernatant was recovered and purified. Only samples with positive mitochondrial DNA amplification were used for library preparation.
• Ancient DNA extract was converted to a TruSeq Nano Illumina library using the manufacturer's protocol with slight modifications that account for the ancient DNA damage. The adaptor mix was gently pre-heated before adding ligase enzyme.
• genomic capture was performed using the myBaits Expert Whole Genome Enrichment (WGE) kit.
• Aliquots of the amplified libraries were first pooled and sequenced on MiSeq instrument (2 × 75 bp) on the IGenSeq platform.
• Data processing. All of our newly sequenced samples were specifically preprocessed through EAGER-v1.92.37 47. From the raw shotgun sequencing data, sequencing adaptors were thus clipped, and pairedend reads were merged. Reads were then aligned against the 1000Genomes-phase2 reference genome.
• To evaluate contamination, we used AuthentiCT 52, and for the male samples, we calculated contamination using the X chromosome 53 with ANGSD 54.
• We selected 3529 published ancient human genomes from Eurasia. DNA sequencing data were generated with whole genome shotgun or hybridisation capture technics. We analysed ancient data with 1587 Eurasian individuals.
• PCA. We ran PCA with smartpca 55 using the HO-dataset, on 1390 European and Middle Eastern individuals, and we projected all the ancient samples. ADMIXTURE. We computed ADMIXTURE 33 analysis using the HO dataset, where all populations were downsampled to a maximum of 20 individuals.
• We tested a rotating group with Iran_GanjDareh_N, CHG, Russia_Caucasus_ Eneolithic, PPN, Barcın_N, Anatolia_TellKurdu_LN IRQ_Nemrik9_PPN and TUR_SE_Mardin_PPN. We then run the same qpAdm analysis for the Mentesh group form by MT23, MT7 and MTT001 and all the South-Caucasus Neolithic (SCN) individuals in one group.

[h=2]DATES.[/h]

• We used DATES v75318 35 to estimate the time of admixture events between Iran_N and Anatolia_N in the Mentesh group. We assumed 29 years per generation 13.
• Y chromosome and mitochondrial haplogroups were determined by comparison of vcf files to databases.
• We used PLINK-v1.90b6.16 58,59 to filter out anything but biallelic autosomal SNP positions carrying a minor allele frequency greater or equal to 5% from our 1240K dataset. The resulting dataset was transposed to the tped/tfam format and used as input for READ. We re-applied this protocol a second time using six individuals from the Late Neolithic and Early Chalcolithic site of Tell Kurdu.
• We used TKGWV2 to substantiate our initial results. The final BAM files from all Mentesh individuals were downsampled to around 1.8M reads beforehand. We then performed the analysis using the 22M nonfixed biallelic variant set provided by Fernandes et al.
• The dataset used for PCA and ADMIXURE includes 597,573 SNPs for 5116 individuals. When an analysis targets Mentesh as a group, it means three individuals grouped together. READ analysis has been replicated twice.