A high-resolution picture of kinship practices in an Early Neolithic tomb

Abstract

To explore kinship practices at chambered tombs in Early Neolithic Britain, here we combined archaeological and genetic analyses of 35 individuals who lived about 5,700 years ago and were entombed at Hazleton North long cairn¹. Twenty-seven individuals are part of the first extended pedigree reconstructed from ancient DNA, a five-generation family whose many interrelationships provide statistical power to document kinship practices that were invisible without direct genetic data. Patrilineal descent was key in determining who was buried in the tomb, as all 15 intergenerational transmissions were through men. The presence of women who had reproduced with lineage men and the absence of adult lineage daughters suggest virilocal burial and female exogamy. We demonstrate that one male progenitor reproduced with four women: the descendants of two of those women were buried in the same half of the tomb over all generations. This suggests that maternal sub-lineages were grouped into branches whose distinctiveness was recognized during the construction of the tomb. Four men descended from non-lineage fathers and mothers who also reproduced with lineage male individuals, suggesting that some men adopted the children of their reproductive partners by other men into their patriline. Eight individuals were not close biological relatives of the main lineage, raising the possibility that kinship also encompassed social bonds independent of biological relatedness.

Extended Data Figure 1.. The Hazleton North chambered tomb.

a, Distribution of human remains in both chambers (adapted from). b, Right humerus from Individual C showing helical fracture (red arrow), tooth marks (yellow arrow) and gnawed proximal and distal ends (white arrows).

Extended Data Figure 2.. Degrees of biological relatedness among individuals at Hazleton North

(Supplementary Information section 2.2). Pairs with fewer than 15,000 overlapping SNPs are indicated with an asterisk.

Extended Data Figure 3.. Using allelic mismatch rates patterns along the chromosomes to differentiate types of relationships for individuals sharing the same amount of DNA.

a, Differentiating between parent-offspring and sibling relationships. Allelic mismatch rate values across sliding windows of 20 Mb, moving by 1 Mb each step. As an example, we show values at chromosome 17 and include for reference a comparison between two unrelated Neolithic individuals from Britain (in brown), and a comparison between one individual and himself (in purple) to show how mismatch rates behave when two chromosomes are shared. The mismatch rate pattern for SP1m-SC1f is compatible with one chromosome shared along the entire chromosome 14 (in fact, along all autosomal chromosomes (Supplementary Table 6)), indicating a parent-offspring relationship. In contrast, the NC7f-SP3m comparison shows regions on chromosome 17 where no chromosome is shared (~65–70 Mb), other regions where two chromosomes are shared (~0–25 Mb) and other regions where one chromosome is shared (~25–60 Mb), compatible with a sibling relationship. b, Comparing DNA sharing patterns between SC9f and her paternal grandparents. We show mismatch rate values at chromosome 2 and include for reference a parent-offspring comparison (SE1m-SP2m; in blue) to show how mismatch rates behave when one chromosome is shared. Two recombination events (one at ~145 Mb and other at ~220 Mb) in SC9f’s father’s gamete result in SC9f’s sharing one chromosome with SC3m from the start of the chromosome to ~145 Mb, one chromosome with SC4f from 145 to 220 Mb and one chromosome with SC3m from 220 Mb to the end of the chromosome. This pattern of sharing one chromosome with either SC3m or SC4f at every location of the genome is characteristic of comparisons between a grandchild and his/her two grandparents and is also observed in the other autosomal chromosomes.

Extended Data Figure 4.. Alternative family tree fitting all the genetic evidence except the IBD breakpoints co-localization analysis

(Supplementary Section 2.4, Extended Data Figure 5). Individuals are coloured according to the female sub-lineage they belong to (NC1m and NC5m do not belong to any of the four major sub-lineages and are thus given a different color).

Extended Data Figure 5.. Using co-localization of IBD breakpoints to disambiguate between family tree in Fig. 1c and family tree in Extended Data Fig. 4.

a, We show mismatch rate values across sliding windows of 20 Mb on chromosome 3, moving by 1 Mb each step, for comparisons between SC3m and his four second-degree relatives. Recombination events on chromosome 3 needed to explain the observed mismatch rate patterns under b, the scenario of tree in Fig 1c. where 4 recombination events are required or c, the scenario of the tree in Extended Data Fig.4 where 10 recombination events are required including the extremely implausible occurrence of two recombination events at the same genomic locations in four different gametes.

Extended Data Figure 6.. Testing the validity of the family pedigree in Fig. 1c using X-chromosome relatedness and number of shared IBD segments.

a, Relatedness coefficients in the X-chromosome for first and second degree relationships with more than 300 overlapping SNPs. For each comparison, expected values according to the type of relation in the family tree in Fig. 1c are shown in grey boxes. Bars represent 95% confidence intervals. b, Number of shared IBD segments for first- and second-degree relationships. Pairs are grouped according to their type of relation in the family tree in Fig. 1c.

Extended Data Figure 7.. Testing the consistency of the kinship results using NgsRelate.

a, Correlation between the relatedness coefficient r and the Theta coefficient computed with NgsRelate, restricting to comparisons with more than 15,000 overlapping SNPs. b, Cotterman coefficients k0 and k2 for first and second degree relationships, as computed with NgsRelate.

Extended Data Figure 8.. Comparing autosomal relatedness between reproductive partners, different male reproductive partners of a female and different female reproductive partners of a male.

To estimate relatedness coefficients between unsampled and sampled male reproductive partners of a female, we doubled the relatedness coefficient obtained between the son of the unsampled male and the sampled male, to account for the fact that a son is one degree of relationship further away from their father’s relatives as compared to his father. Bars represent 95% confidence intervals.

Extended Data Figure 9.. Principal Component Analysis and inbreeding analysis.

a, Principal component analysis of Hazleton North individuals and other ancient individuals from Britain and Ireland. Ancient individuals were projected onto the principal components computed on a set of present-day West Eurasians genotyped on the Human Origins Array (not shown in the figure). Individuals with fewer than 15,000 SNPs on the Human Origins dataset were excluded for this analysis. b, Runs of homozygosity (ROH) in different length categories for the Hazleton North individuals with higher than 400,000 SNPs covered. ROH were computed using hapROH. On the right, we plot the expected ROH length distribution for the offspring of closely related parents in outbred populations and for individuals from populations with small effective population size.

Fig. 1.. The Hazleton North pedigree in the context of the physical structure of the tomb.

a, Plan of Hazleton North showing the north and south chambered areas in the cairn (grey). Adapted from. b, Burial locations for individuals, with squares for males and circles for females. Individuals are coloured according to the female sub-lineage they belong to. The relative position of each individual within each compartment does not reflect the exact location in which the corpse or remains were placed. c, Reconstruction of the pedigree, using the same colour scheme and indicating the locations of individuals in the tomb, osteological information including age estimates, and different mitochondrial DNA haplogroups as small circles with different colours. Individuals with a dotted outline are unsampled (U) and their existence is inferred. Pink, blue and orange dotted lines indicate likely second-, third- and fourth-degree relationships, respectively. Marks at the top corners of individuals indicate how many genealogical connections linking individuals in the third through fifth generations to male NC1m traverse through that individual (in blue, connections through step-fathers).