Social stratification without genetic differentiation at the site of Kulubnarti in Christian Period Nubia

Nubia, situated along the Nile between Aswan and Khartoum, has a deep history of interactions among peoples from Africa and West Eurasia. Archeological and historical evidence document long-standing ties with Egypt and influences from West Eurasia, including the introduction of Christianity and, later, Arab conquests. Prior genetic studies of present-day Nubian groups show a mixture of sub-Saharan African and West Eurasian-related ancestry, much of it linked to migrations during and after the Arab conquest, obscuring earlier population history. Ancient DNA offers a way to directly investigate the ancestry of Nubian populations before Islamic-period migrations. The site of Kulubnarti (Second–Third Cataracts; ~650–1000 CE) includes two Christian cemeteries (R and S) ~1 km apart with indistinguishable burial practices but significant differences in morbidity, mortality, and inferred socioeconomic status. This context raises key research questions: (1) What was the genetic ancestry of Christian Period Nubians at Kulubnarti, and what were the sources and timing of admixture? (2) Were individuals buried in the R and S cemeteries genetically distinct—consistent with strict endogamy or differing origins—or do they represent a single population exhibiting social stratification without genetic differentiation? (3) Was admixture sex-biased, and what does this reveal about social structure and mobility?

Previous archeological and bioarcheological work documented extensive cultural connections between Nubia and Egypt over millennia, with Christian churches, Greek and Coptic inscriptions, and Old Nubian texts reflecting West Eurasian influence mediated via Egypt. Morphological studies (cranial and dental traits) suggested affinities between Kulubnarti and Wadi Halfa populations but lacked resolution to infer fine-scale genetic relationships. Genetic studies of modern northeastern Africans, including Nubians, Beja, and Sudanese Arabs, reveal gradients of West Eurasian-related ancestry correlated with geography and historical contacts, particularly increases associated with the Arab expansions in the late 1st and early 2nd millennia CE. Ancient DNA from Egypt indicates substantial West Eurasian-related ancestry present by the Late Period and Ptolemaic/Roman times, with varying contributions of Nilotic-related ancestry. Broader studies show West Eurasian-related ancestry entered northeastern and eastern Africa in multiple phases over the past several thousand years, including Levant Neolithic/Chalcolithic-related inputs and later Iranian/Caucasus-related components. Against this backdrop, there remained a gap in genome-wide ancient data from Nubia before Islamic-period migrations and uncertainty as to whether social divisions at Kulubnarti reflected genetic differentiation.

Ethical approvals covered the 1979 excavations and subsequent analyses. Petrous bones from 111 individuals from the R and S cemeteries were processed in dedicated cleanrooms (UCD and HMS). DNA was extracted and dual-barcoded double-stranded or dual-indexed single-stranded libraries were prepared with UDG or partial-UDG treatment. In-solution capture targeted the mitochondrial genome and 1,233,013 autosomal SNPs (1240k panel); sequencing was performed on Illumina platforms. Reads were merged, adapters trimmed, and mapped to RSRS mtDNA and hg19; duplicates were removed and terminal bases trimmed to reduce damage artifacts. Authentication criteria included ≥20,000 SNPs hit, characteristic terminal C→T damage, low mtDNA and X-chromosome contamination estimates; 66 individuals (27 R, 39 S) passed QC, with mean coverage ~0.28× at targeted SNPs. Pseudo-haploid genotype calls were generated; one individual from each first-degree pair (4 pairs) was excluded from population-level analyses. Radiocarbon dating (29 individuals) and Bayesian modeling established contemporaneous cemetery use (~660–960 calCE). Population genetic analyses included PCA (ancients projected onto axes computed from modern African and West Eurasian groups), f3 admixture tests, f4 tests to identify ancestry outliers, qpAdm to model ancestry proportions and source candidates (using “09” reference set with Anatolia EBA added), model competition to distinguish distal/proximal West Eurasian-related sources, and qpWave to test clade relationships between cemetery groups and between Kulubnarti and present-day Nubians. DATES estimated admixture timing for pooled groups and individuals. Relatedness inference identified first–third/fourth-degree kinship and inter-cemetery relatives. Runs of homozygosity were inferred with hapROH for individuals with >400k SNPs. Uniparental markers were called: mtDNA haplogroups via HaploGrep on RSRS-aligned data; Y haplogroups via targeted and off-target Y-SNPs and YFull/ISOGG trees; a chi-square test evaluated cemetery differences in Y-lineage distributions. All analyses used autosomal data unless otherwise specified; datasets merged published ancient and modern references as appropriate for each method.

- Generated genome-wide data for 66 Kulubnarti individuals (27 R cemetery, 39 S cemetery); 29 had direct 14C dates confirming contemporaneous use of both cemeteries (~660–960 calCE). Average SNP coverage ~0.28×. - Admixture: Kulubnarti individuals fall on a cline between Nilotic-related and West Eurasian-related ancestries. qpAdm modeling (excluding six outliers) fits a two-way model with Dinka (Nilotic) and Levant_BAIA (West Eurasian-related) proxies, estimating ~57.5 ± 0.3% West Eurasian-related and ~42.5% Nilotic-related ancestry on autosomes. Individual Nilotic-related ancestry for non-outliers: ~35.6–49.0%. - Source of West Eurasian-related ancestry: Model competition indicates Egypt_published (New Kingdom/Late Period/Ptolemaic) is a plausible proximal source; distal ancestry is best represented by Bronze/Iron Age Levant (Levant_BAIA). Ancient Egyptians themselves carry ~5.0 ± 0.7% Dinka-related ancestry. - Cemetery comparison: No systematic genetic differences between R and S cemeteries. qpAdm estimates of Nilotic ancestry overlap (R: 43.2 ± 0.4%; S: 42.3 ± 0.4%); qpWave clade test p = 0.25; f4 clade tests non-significant after multiple testing; FST = 0.0013. Seven inter-cemetery relative pairs (closest second-degree) indicate familial links across cemeteries; enrichment of within-cemetery burial for first/second-degree relatives observed (combined p = 0.021), but third-degree relative distribution indistinguishable from random. - Admixture timing: DATES on pooled Kulubnarti (excluding outliers) estimates admixture 22.2 ± 1.4 generations (~620 ± 40 years) before the average lifetime, placing admixture ~111–265 CE (95% CI) relative to the cemetery mid-date (~810 CE). Cemetery-specific estimates overlap (R: ~21.7 ± 1.9 generations; S: ~22.3 ± 1.7 generations). Individual DATES estimates vary widely (10.4 ± 3.5 to 46.2 ± 11.8 generations), indicating multiple waves/continuous admixture over ~1,000 years. - Sex-biased admixture: West Eurasian-related ancestry is higher on the X chromosome (64.4 ± 1.8%) than autosomes (57.5 ± 0.3%) (Z = 3.78), implying disproportionately female-associated West Eurasian-related ancestry; estimated 68% (95% CI 59–77%) of West Eurasian-related ancestry derives from females. - ROH: Among nine individuals with >400k SNPs (mostly S cemetery), eight show no or low ROH >4 cM; only one individual (~80 cM in >20 cM blocks) indicates close parental relatedness. Pattern suggests a small community connected to a larger metapopulation. - Uniparental markers: 35/63 non–first-degree individuals carry mtDNA haplogroups common in West Eurasia (e.g., H2a, U5b2b5, J2a2e, R0a1, T1a7, U1a1, U3b, N1b1a2), alongside multiple African L lineages (notably L2a1d1; four L5a1b). Y-chromosome haplogroups in 28 unrelated males are predominantly E-M215 branches (n=15), with additional lineages of likely West Eurasian origin; cemetery differences in Y distribution are not significant (p = 0.11). - Present-day Nubians (Mahas, Danagla, Halfawieen) do not form a clade with Kulubnarti (qpWave clade tests p < 1.5e-10), indicating additional post-Christian Period admixture. Re-estimated admixture dates align with Arab conquest–period gene flow (Mahas: 889–1210 CE; Danagla: 855–1095 CE; Halfawieen: 1148–1498 CE).

The study demonstrates that Christian Period Kulubnarti Nubians carried substantial West Eurasian-related ancestry, most similar to Bronze/Iron Age Levantine sources likely mediated through Egypt, together with Nilotic-related ancestry. Despite clear archaeological signals of socioeconomic differences between the R and S cemeteries, genome-wide data reveal no population-level genetic differentiation, and the presence of multiple inter-cemetery relatives supports social division without genetic isolation. Admixture timing estimates and inter-individual ancestry variance indicate prolonged and possibly continuous gene flow spanning roughly a millennium, consistent with historical periods of intense interaction (e.g., Meroitic Kingdom dynamics and long-term Egypt–Nubia connections). The excess of West Eurasian-related ancestry on the X chromosome implies female-biased gene flow into Kulubnarti, potentially reflecting patrilocal residence or female-mediated connectivity with broader populations. Comparisons to present-day Nubians show that subsequent admixture—coincident with Arab expansions—further reshaped regional genetic landscapes, meaning modern data cannot be directly extrapolated to earlier periods without ancient DNA. Collectively, the results address the research questions by showing social stratification at Kulubnarti was not accompanied by genetic stratification, clarifying ancestry sources and timing, and illuminating sex-biased demographic processes.

This work increases genome-wide ancient data from the Nile Valley by reporting 66 Christian Period individuals from Kulubnarti and shows: (1) a two-way admixture between Nilotic-related and West Eurasian-related ancestries, the latter ultimately most like Bronze/Iron Age Levantine ancestry likely introduced via Egypt; (2) no genetic differentiation between socially distinct cemetery groups, with inter-cemetery relatives indicating fluidity; (3) admixture spanning centuries to a millennium; and (4) disproportionately female-associated West Eurasian-related ancestry consistent with sex-biased gene flow. Present-day Nubian groups are not direct descendants of Kulubnarti without later admixture, highlighting significant post-Christian Period gene flow. Future research should expand temporal and geographic sampling in Nubia and Egypt (especially pre-Christian and Meroitic periods), refine models of sex-biased mobility and social structure (e.g., patrilocality), and integrate isotopic, archaeological, and genomic data to resolve fine-scale demographic histories.

- Temporal and geographic scope: Analyses are confined to one site (Kulubnarti) during ~650–1000 CE; lack of older Nubian genomes limits inferences about earlier admixture waves and ancestry sources. - Reference panels: Only three ancient Egyptian individuals were available as proximal references; absence of ancient genomes from plausible African source populations proximal to West Eurasian-related ancestry constrains source resolution. - Data quality: Average coverage is low (~0.28×); some analyses exclude individuals with fewer SNPs or rely on pseudo-haploid calls, which may reduce precision. - Modeling assumptions: DATES assumes a single-pulse admixture model, yielding average dates if admixture was continuous or multi-wave; qpAdm/qpWave rely on specified outgroup sets and model fit thresholds. - ROH analyses were limited to nine individuals with >400k SNPs, reducing power to generalize relatedness patterns. - Cemetery-based Y-chromosome differences were not statistically significant; small sample sizes limit power to detect subtle sex-specific patterns.