Repeated plague infections across six generations of Neolithic Farmers

The study addresses the long-debated cause of the Neolithic demographic decline (ca. 5300–4900 cal BP) in Europe, contrasting hypotheses of agricultural crisis versus spread of an early form of plague. Prior work identified Late Neolithic and Bronze Age Yersinia pestis lineages (LNBA and LNBA') and, more recently, earlier diverging pre-LNBA lineages from Scandinavian and Baltic individuals, suggesting the presence of ancestral plague in the Neolithic. In Scandinavia, the demographic decline overlaps with cultural transitions such as the disappearance of the TRB (Buren/Neolithic/Tircherbrue) culture and changes in megalith-building. Mortuary practices, social organization (e.g., kin-based burial hypotheses), and selection criteria for megalithic interments remain contested. This study aims to clarify social structure and the frequency, spread, and potential impact of plague within Middle Neolithic Scandinavian communities by integrating population-scale ancient genomics, kinship reconstruction, and pathogen screening across multiple megalithic sites.

The paper situates its research within evidence for prehistoric plague genomes largely from the LNBA period (4700–2400 cal BP), divided into lineages differing by the presence/absence of a key gene associated with flea-borne transmission. Newer studies (Rascovan et al., Susat et al.) identified earlier diverging pre-LNBA Y. pestis lineages in Scandinavia and Latvia, implying the presence of ancestral plague forms during the Neolithic, possibly via sporadic contacts. In Sweden, the TRB culture overlapped temporally and geographically with the Pitted Ware Culture (PWC), offering plausible sources of hunter-gatherer ancestry in mixed individuals. The paper references prior findings on European Steppe-related ancestries (Corded Ware/Battle Axe) emerging post-4800 BP, admixture dates, and patrilineal/patrilocal social structures documented archaeologically (e.g., Hazleton North), providing a comparative framework for interpreting kinship and disease spread in megalithic communities.

- Sampling and sequencing: 174 samples from ancient Scandinavian human skeletal remains were screened; after data processing and consolidation, 108 distinct individuals from nine burial structures (seven Falbygden passage graves, one west-coast Swedish site, one Danish site) formed the final dataset. Sex determination and ancestry were inferred from genomic data. - Population genetics: Constructed a panel of 1,430 ancient shotgun-sequenced genomes as references. Performed PCA to assess genetic affinities and temporal associations. Y-chromosome haplogroups were assigned. Identity-by-descent (IBD) sharing analyses clustered individuals into fine-scale groups, including multiple Falbygden-specific clusters and a broader Danish/Swedish cluster. Sr isotope data supplemented mobility/origin inferences for select individuals. DATES analysis estimated Steppe–Farmer admixture timing. - Kinship and pedigrees: Pairwise relatedness estimates classified first-, second-, and third-degree relationships, enabling reconstruction of one large and three smaller pedigrees. Bayesian chronological modeling of radiocarbon dates constrained temporal spans of pedigree branches (~150–180 years). - Spatial burial analysis: Mapped skeletal elements and individuals’ positions within the Frälsegården chamber to test associations between kinship, sex, generation, and burial location, noting special treatments and possible secondary burial practices. - Pathogen screening: Metagenomic screening of sequencing data targeted known human pathogens. Yersinia reads were authenticated and quantified. Plague-positive samples were categorized by coverage (tentative <0.1×, lower 0.01–1.5×, higher >1.5×). For higher-coverage Y. pestis genomes (1.5×–10.5×), SNP calling supported phylogenetic reconstruction alongside ancient and modern references. - Pan-genomics: To overcome single-reference biases, reads were mapped to a variation graph built from 26 complete assemblies spanning the Y. pseudotuberculosis complex (including Y. pestis, Y. pseudotuberculosis, Y. similis). Gene presence/absence, especially around virulence-associated ypm loci (ypmA/B/C), was profiled to infer ancestral gene content and potential virulence differences among strains. - Co-infections: Identified instances of co-detection (e.g., Y. pestis with Y. enterocolitica) and cataloged other pathogens where present. - Interpretation: Integrated phylogeny, coverage tiers, kinship timing, and burial placement to infer the number and timing of infection events and their spread within pedigrees.

- Prevalence and distribution: Y. pestis was detected in approximately 17% of the 108 individuals examined, with infections spanning southern Scandinavia, including Falbygden, the Swedish west coast, and Zealand (Denmark). Frälsegården showed the highest rate (13/47; 28% when excluding Steppe-related individuals). - Multiple infection events: Evidence indicates three distinct plague infection events within roughly 120 years, traced across a multi-generational pedigree at Frälsegården. - Plague phylogeny: Five higher-coverage Y. pestis genomes (1.5× to 10.5×) enabled SNP-based phylogenetic placement. Most clustered with LNBA-related strains. FRA013 and FRA006 were identical to the previously published Gökhem2 genome from the same site and period. FRA020 differed at three positions (designated strain C vs. strain B cluster), while FRA012 was markedly basal (29–34 SNP differences) relative to B/C and falls near early diverging lineages (designated strain A, pre-LNBA-like). - Strain structure: Three Falbygden-associated strains identified (A, B, C). Strain A detected in the earliest pedigree generation (including the right-side progenitor FRA021 and an early unrelated individual), suggesting initial spread in generation one. Strains B and C appear in generations three to five; strain C is concentrated in the left-side subfamily with identical higher-coverage genomes, consistent with a swift outbreak. - Co-infection: At least one individual (FRA013) showed co-infection with Y. pestis (3.9×) and Y. enterocolitica (1.8×). - Pan-genomic insights: Early diverging strains retained up to ~50 kb of ancestral gene content present in Y. pseudotuberculosis/Y. similis but absent in later Y. pestis strains. Distinct combinations at the ypm locus were identified: strains A and B resembled ancestral ypmB-like organization, whereas strain C showed a previously unobserved gene combination at the ypm locus, potentially linked to increased virulence. - Social structure and kinship: One large pedigree spanning ~150–180 years was reconstructed at Frälsegården, with clear associations between kinship, sex, generation, and burial location. Patterns indicate patrilineal and patrilocal organization, male-line influence on burial placement, and instances of polygynous reproduction and female mobility. - Population structure: Most individuals (n≈96) carried Neolithic Farmer ancestry (TRB/Middle Neolithic), with minor groups showing Steppe-related ancestry (Steppe I ~4400 cal BP, Steppe II ~4000–3000 cal BP). Two individuals showed substantial hunter-gatherer ancestry, likely from PWC admixture. Y haplogroups aligned with ancestry: Farmers largely I2; Steppe groups R1a/R1b/I1. Admixture dating suggests Steppe–Farmer gene flow around ~4750 cal BP, prior to widespread CWC arrival. - Sex ratio: Slight male bias (≈58% male) overall, with site-specific variation.

The integration of pathogen genomics with kinship and spatial burial data reveals repeated plague incursions into a single community over multiple generations. The earliest event (strain A) likely spread during the first generation of the Frälsegården pedigree without causing lineage extinction, while later events (strains B and C) affected subsequent generations, with strain C potentially driving a concentrated outbreak in the left-side subfamily, inferred from high infection frequency and identical genomes. Pan-genomic evidence of ancestral Y. pseudotuberculosis gene content and distinct ypm locus configurations suggests variability in virulence and transmission pathways among early plague strains, potentially including non-flea, human-associated routes. The observed infection rate (~17%) argues against plague being rare or purely zoonotic spillover in Neolithic Scandinavia. Social organization appears patrilineal and patrilocal, with burial placement reflecting kinship and sex, informing how disease could have propagated through close kin networks. Collectively, these findings support a role for plague as a contributing factor to the Neolithic decline in Scandinavia, acting alongside cultural and demographic shifts and possibly other pathogens.

By combining population-scale ancient genomics, kinship reconstruction, and pathogen screening, the study documents pervasive and repeated plague infections in a Neolithic Scandinavian community, identifying three strains (A, B, C) and multiple infection events over roughly 120 years. Early diverging lineages retain ancestral gene content and show novel ypm locus configurations, implying potential differences in virulence and transmission from later historical plague. The work reconstructs detailed pedigrees and demonstrates patrilineal/patrilocal social organization linked to burial practices. These results provide a high-resolution snapshot of disease dynamics and social structure near the end of the Neolithic, indicating that plague likely contributed to regional demographic decline. Future research should expand geographic and temporal sampling, apply improved pathogen capture and long-read/pan-genomic approaches to resolve gene-content variation, and integrate osteological/epidemiological modeling to better quantify virulence, transmission routes, and mortality impacts.

- Sampling and preservation: The dataset is restricted to well-preserved individuals interred in megalithic tombs and a stone cist, potentially biasing infection frequency estimates relative to the broader population. - Detection sensitivity: Pathogen detection depends on tissue type and preservation; qPCR benchmarks suggest substantially higher detection in teeth than bones, implying the true prevalence could exceed the observed 17%. - Coverage constraints: Many plague-positive samples are low coverage, limiting precise phylogenetic placement and precluding discrimination between closely related strains (e.g., B vs. C) in numerous cases. - Chronological resolution: Overlapping calibrated radiocarbon intervals and model uncertainty limit the ability to temporally order infection events within and between pedigree branches. - Functional inference: Virulence assessments rely on gene presence/absence patterns (e.g., ypm locus configurations) without experimental validation; transmission routes are inferred indirectly. - Affiliation of admixture sources: While PWC is a likely source of hunter-gatherer ancestry in some individuals, contributions from other northern HG groups cannot be fully excluded.