Reconstructing past human genetic variation with ancient DNA: case studies from ancient Egypt and medieval Europe.


Recent advances in palaeogenomics have massively improved our knowledge of human history. An explosion in the number of sequenced ancient human genomes has allowed direct exploration of ancestry dispersion, and processes of assimilation of these ancestries (e.g., admixture, replacement). Despite this, some regions and periods remain under-studied. In this thesis, DNA is recovered from populations yet to be characterized in order to use genomics to answer archaeological and anthropological questions.
Despite the rapidly growing availability of ancient genomes from western Eurasia, the Picts of early medieval Scotland (ca. 300-900 CE) are under-represented among these genomes and remain poorly characterised. Chapter 2 describes the genetic relationship of the Picts with Iron Age, early medieval and present-day populations throughout Eurasia. Genome-wide data (0.1 – 16.5X coverage) were retrieved from four individuals representing two Pictish sites (5th-7th centuries CE), Balintore (Easter Ross, northern Pictland) and Lundin Links (Fife, southern Pictland), and mitochondrial genomes from seven individuals from a cemetery at Lundin Links. Using allele frequency-based methods (PCA, ADMIXTURE, f-statistics, qpAdm), their affinities to a reference panel of ancient and modern genomes were assessed. Mitochondrial DNA (mtDNA) haplogroups were reconstructed to gain insight into social organisation at Lundin Links. Two Picts (BAL003 and LUN004) were genetically similar to north-western people dated from the Iron Age and early medieval period; the two remaining Picts (LUN001 and LUN003) showed admixture from a European source and a source related to present-day Native Americans, previously widely spread in eastern Eurasia, with uncertainties concerning the authenticity of this signal. At Lundin Links, the individuals did not share common maternal ancestors, consistent with high female exogamy at the site.
Chapter 3 assess the fine-scale relationship of the Picts with western Europeans dated from the Iron Age onward. A new combined set of imputed ancient European genomes (n = 285, coverage 0.7-16X), including re-imputed diploid genomes dated to the Iron Age through medieval periods from Britain, Iceland, Scandinavia and central Europe, were used to analyse the Pictish genomes with several haplotype-based methods (identity-by-descent, ChromoPainter and FineSTRUCTURE). These high-resolution techniques allow for a refined assessment of the biological relatedness of the Picts to populations of north-western Europe. Overall, the Picts showed a greater relative genetic affinity with Iron Age populations from the British Isles than with ancient Europeans. In addition, the Picts are genetically similar to present-day Welsh, Northern Irish and Scottish populations. The data presented here highlight the advantages of using high-quality DNA sequences, coupled with new haplotype-based analytic tools, to disentangle complex and recent demographic history between ancient populations.

Compared with Europe, palaeogenetics in Africa is poorly studied, in part because DNA degrades faster in tropical and dry environments. Chapter 4 aims to unveil population movements in Egypt and Sudan from the Neolithic onward. DNA was extracted from 94 samples from Armant (Egypt), Nuerat (Egypt) and Ghaba (Sudan) dated from the Early Neolithic to the historic period. Genome-wide data were successfully recovered from one sample from Nuerat sequenced to 0.22X coverage, dated to 2,868-2,492 cal BCE (95.4% probability) - consistent with the 3rd-4th Dynasties of the Old Kingdom. Allele frequency-based analyses (PCA, ADMIXTURE, f-statistics, qpAdm) show a strong genetic affinity of this sample to Levantine Natufians. Compared with genomes dated from the end of the Dynastic period (Third Intermediate Period) and present-day Egyptians, the Nuerat sample did not carry the Caucasus Hunter-Gatherer genetic component that started to spread across West Asia ~4,000 years ago and is widely spread in present-day populations. The presence of this component in Egypt is likely associated with admixture between local Egyptian populations and Bronze Age-related populations from West Asia. This admixture pattern might result from the dominance of Lower Egypt by Canaanite (Levantine) rulers during the Second Intermediate Period (ca. 1,650-1,550 BCE).
Even though optimised wet-laboratory techniques improved retrieval of aDNA, it is still a challenging technique, with a higher proportion of failure than success, especially for samples from tropical and sub-tropical regions. Alternative methods are required. Tooth crown and root traits, like those in the Arizona State University Dental Anthropology System (ASUDAS), are seemingly useful as genetic proxies. Chapter 5 investigates whether such traits can approximate genetic relatedness in the absence of aDNA, among continental and global samples. For 12 African and 32 global populations, Mantel correlations were calculated between the mean measure of divergence (MMD) distances from up to 36 ASUDAS traits, and FST distances from >350,000 single nucleotide polymorphisms (SNPs) among matched dental and genetic samples. A close ASUDAS/SNP association, based on MMD and FST correlations, is evident, with rm-values between 0.84 in Africa and 0.72 globally. Partial MMD/FST correlations controlling for geographic distances are strong for Africa (0.78) and moderate globally (0.4). Relative to prior studies, MMD/FST correlations imply greater dental and genetic correspondence. The implication is that ASUDAS traits can be used instead of genomic markers when the latter are unavailable.