Introduction
The first complete ancient human genome—from a 4000-year-old Palaeo-Inuit man from Greenland—was sequenced just 15 years ago (Rasmussen et al. Reference Rasmussen, Li and Lindgreen2010). Since that landmark event, the field of palaeogenomics—the study of complete genomes from deceased organisms—has transformed studies of human evolution, history and medicine. To date, genome-wide data for over 9990 ancient and historical individuals have been generated and made publicly available (Mallick et al. Reference Mallick, Micco and Mah2024). Although broad-scale historical and evolutionary questions still dominate palaeogenomics research, there have been increasing calls for more fine-scale, regional studies to infer local histories and examine nuanced questions about ancient demographic, sociocultural and political aspects of past societies (Ávila-Arcos et al. Reference Ávila-Arcos, Raghavan and Schlebusch2023; Racimo et al. Reference Racimo, Sikora, Vander Linden, Schroeder and Lalueza-Fox2020). Simultaneously, there have been calls for palaeogenomics researchers to collaborate more closely with archaeologists (Dillehay Reference Dillehay2021) and descendant communities (Fleskes et al. Reference Fleskes, Bader, Tsosie, Wagner, Claw and Garrison2022; Kowal et al. Reference Kowal, Weyrich and Argüelles2023) in order to develop better research frameworks, conduct research more ethically and better interpret genetic data in the context of prior archaeological, anthropological and Indigenous knowledge.
One particular area of promise in this regard is the use of ancient genomic data to explore biological relatedness in archaeological contexts. Ancient DNA (aDNA) researchers have long seen the potential of genetic data in inferring kinship in mortuary contexts (Kaestle & Horsburgh Reference Kaestle and Horsburgh2002). However, many of the early studies were limited by a reliance on uniparental markers, such as mitochondrial DNA and Y-chromosome DNA (Vai et al. Reference Vai, Amorim, Lari and Caramelli2020). Improvements to methods for retrieving and analysing genomic data have made genome-wide data far more accessible, even for smaller research groups, and there has been a concurrent increase in interest in applying high-resolution genomic approaches to identifying relatedness among ancient individuals (Amorim et al. Reference Amorim, Vai and Posth2018; Pilli et al. Reference Pilli, Vai and Moses2024; Quilter et al. Reference Quilter, Harkins and Fanco Jordan2025; Rivollat et al. Reference Rivollat, Rohrlach and Ringbauer2023; Sümer et al. Reference Sümer, Rougier and Villalba-Mouco2024; Vai et al. Reference Vai, Amorim, Lari and Caramelli2020; Yaka et al. Reference Yaka, Mapelli and Kaptan2021).
However, the proliferation and visibility of palaeogenomics approaches applied to anthropology have engendered numerous critiques of ethics, study design and interpretation (Strand et al. Reference Strand, Kallen and Mulcare2024). One repeated critique is the lack of familiarity or engagement of geneticists with extensive literature in related fields, and/or failure to collaborate with subject matter experts from other fields, such as archaeology or descendant communities, to frame research questions appropriately in the context of what is already known (Dillehay Reference Dillehay2021). This can have unintended consequences, such as the conflation of biological groupings and social identity, an uncritical reliance on binary categories (such as nature versus culture, male versus female, elite versus non-elite), and framing DNA as a ‘neutral arbiter of past identity’ without context (Crellin & Harris Reference Crellin and Harris2020). The privileging of genetic data over other kinds of data from the social sciences has been termed ‘molecular chauvinism’ (Horsburgh Reference Horsburgh, Strand, Kallen and Mulcare2024), and it can be particularly harmful when research is conducted without permission of or input from descendant communities (Claw et al. Reference Claw, Lippert and Bardill2017; Fleskes et al. Reference Fleskes, Bader, Tsosie, Wagner, Claw and Garrison2022; Kowal et al. Reference Kowal, Weyrich and Argüelles2023).
We acknowledge and agree with these critiques of palaeogenomics research done carelessly or unethically, and also add that a major hindrance to truly interdisciplinary approaches to kinship studies can be mutual unintelligibility between different fields’ extensive bodies of theory and research terminology. Nevertheless, we contend that identifying biological relatedness is an essential component of kinship studies of past peoples because it provides a direct and measurable way to identify or rule out biological connections between individuals, helping us avoid assumptions or inferences based on less direct evidence. In this way, palaeogenomics data nicely complements other lines of evidence from archaeology and the historical record when available. In cases where it is feasible and ethical to employ, we argue that palaeogenomics is the most appropriate and powerful tool available for this purpose.
We offer this article as a step towards bridging disciplinary boundaries. To improve understanding of palaeogenomics approaches by non-geneticists, we briefly review existing methods for inferring biological relatedness from genomic data and define terminology commonly used by geneticists. To illustrate the strength of research that has involved close collaborations between geneticists, archaeologists, anthropologists and descendant communities on addressing different aspects of kinship, we provide a brief overview of four palaeogenomic studies centred on kinship inference. Throughout, to help our colleagues avoid common pitfalls, we offer guidelines for researchers wishing to work in this area. It is our hope that this article will be a resource for scholars in multiple fields wishing to apply genetics to the study of past kinship systems.
Methods for inferring relatedness in palaeogenomics
In recent years, palaeogenomics has provided powerful tools to enable the study of biological relatedness between ancient individuals with unprecedented precision. In genetics, we have our own set of terminology, theory, and methods, which we briefly summarize in this section (see also Box 1).
Key genetic terminology for understanding kinship and relatedness.
Allele: The different versions of a DNA sequence at a given genomic position. These variations arise through mutations, which are processes that are inherent in living cells.
Background relatedness: The baseline level of genetic relatedness among individuals due to shared ancestry from a common ancestor. In genetic studies, background relatedness can confound the identification of recent biological kinship.
Coancestry coefficient: The probability that two alleles randomly drawn from two individuals are identical by descent (see ‘identity-by-descent’ definition below). It is also called coefficient of parentage or coefficient of consanguinity.
Coefficient of relatedness: The proportion of the genome shared in identity-by-descent between two individuals. Also known as ϕ. This measure quantifies the degree of genetic relatedness.
Depth of coverage: The number of times a particular base in the DNA sequence of an individual is read during a sequencing experiment. It is typically averaged across the genome and reported in every palaeogenomic study.
Exogenous DNA: DNA from sources other than the organism being studied, such as soil or contaminants. Its counterpart is the ‘endogenous DNA’.
Genetic markers: Specific sequences in the genome that can be used to study genetic variation, evolution, and biological relatedness. Examples include SNPs and STRs (see definitions below).
Genome: The entire set of DNA sequences found in a cell. In humans, the genome consists of 23 pairs of chromosomes in the cell nucleus plus the mitochondrial DNA.
Identity-by-descent (IBD): The shared inheritance of an identical portion of DNA between two individuals from a recent common ancestor. IBD is a key concept for inferring biological relatedness.
Inbreeding coefficient: A statistic in genetics that is proportional to the probability that an individual has two alleles in IBD at a given gene.
Short Tandem Repeats (STRs): Repeated short DNA sequences found throughout the genome in varying copy numbers. STRs are highly variable across individuals and are thus widely used in forensics and paternity testing.
Single-Nucleotide Polymorphism (SNP): The variation in a single base at a specific position in the DNA. SNPs are the prime option for genetic markers in palaeogenomics studies due to their abundance across the genome and how accurately they can be analysed in degraded DNA samples.
Transitions: A type of mutation where the DNA base A is replaced with G, or C with T, or vice versa. Transitions are caused by innate processes in living cells, such as DNA replication error and spontaneous damage, and may continue to accumulate in DNA molecules even after the death of the organism.
Uniparental markers: Refers to the mitochondrial DNA (inherited maternally) or the Y-chromosome (inherited paternally). They are used to trace maternal or paternal lineage in evolutionary genetics studies.
Central to genetic investigations of kinship is the concept of identity-by-descent (IBD). IBD refers to the phenomenon where two individuals inherit the same segment of DNA from a common ancestor. Available methods to infer biological relatedness in palaeogenomics leverage information across thousands of genetic markers widespread in the genome of two (or more) individuals in order to calculate the probabilities of IBD and infer the degree of relatedness between them. By calculating the proportion of loci in the genome where two individuals share zero, one or two alleles (defined as the different variants of a genetic marker), it is possible to compute the probabilities of IBD across the genome and assess in detail the degree of biological relatedness between these individuals, such as if they are parent–child, siblings, or or other more distant genetic relationships (Table 1). For example, full siblings share one allele in IBD at approximately 50 per cent of the genome and two alleles in IBD at ∼25 per cent, while parent–child pairs share a single allele in IBD across 100 per cent of their genome. These probabilities of IBD inform key statistics used to quantify biological relatedness such as the inbreeding coefficient (the probability that an individual carries two alleles in IBD), the coancestry coefficient (the probability that two alleles taken at random from two individuals are identical by descent), and ϕ or the coefficient of relatedness (the proportion of the genome that two individuals share by IBD).
Identity-by-descent probabilities for two individuals sharing zero (k 0), one (k 1) and two (k 2) alleles according to different degrees of relationships. (Modified from Weir et al. Reference Weir, Anderson and Hepler2006.)
To infer IBD, two types of genetic markers are commonly used: short tandom repeats (STRs) and single-nucleotide polymorphisms (SNPs) (see Box 1 for definitions). While STRs are preferred in forensic contexts and paternity tests due to being highly variable and thus offering great resolution to exclude potentially unrelated individuals (even when using a small number of markers), SNPs are the main choice in palaeogenomics studies (Mallick et al. Reference Mallick, Micco and Mah2024; Vai et al. Reference Vai, Amorim, Lari and Caramelli2020). One of the main reasons is that the fragmented and damaged nature of aDNA means that STRs are variably preserved. Because of that, the number of STRs detected at a given position in the genome may not accurately reflect the number present when the individual was alive. In contrast, because SNPs are single DNA base changes, they are more likely to be accurately identified. SNP sets such as the so-called ‘1240K SNP set’, commonly used in human population genetics studies (Fu et al. Reference Fu, Hajdinjak and Moldovan2015), provide genome-wide coverage for inferring relatedness. Genome-wide data allow for a more comprehensive picture of the IBD patterns across the genome and reduce noise and bias that may arise when using smaller datasets restricted to specific genes.
Among the main methods to infer relatedness and calculate IBD probabilities from genome-wide SNP data, such as the 1240K SNP set, are KING (Manichaikul et al. Reference Manichaikul, Mychaleckyj, Rich, Daly, Sale and Chen2010), NgsRelate (Korneliussen & Moltke Reference Korneliussen and Moltke2015), KIN (Popli et al. Reference Popli, Peyrégne and Peter2023), ancIBD (Ringbauer et al. Reference Ringbauer, Huang and Akbari2024) and READ (Alaçamlı et al. Reference Alaçamlı, Naidoo and Güler2024). As an illustration of how their workflows function, we will focus on READ. This method uses low-coverage aDNA data to calculate the pairwise mismatch rate (P0), i.e. the fraction of sites at which two individuals have differing alleles. If two individuals are closely related, they will have relatively lower P0 due to the fact they share many alleles in IBD. The method then normalizes the calculated values by the expected P0 based on a pair of unrelated individuals from the same population sample. From these normalized P0 values, READ infers patterns of allele sharing across the genome and uses these to estimate ϕ, the proportion of the genome in IBD between a given pair of individuals. This value is critical for identifying specific degrees of relatedness, such as distinguishing between unrelated, first-, second- and third-degree relatives. Note that READ in its current version infers kinship up to third-degree relatedness and has introduced several enhancements over the previous version. These include the ability to distinguish between parent–offspring and full siblings relationships, as well as the support for using of X-chromosome data (Alaçamlı et al. Reference Alaçamlı, Naidoo and Güler2024).
Interpreting relatedness data in palaeogenomics
While READ and other methods are increasingly becoming more user-friendly, making it possible for a wide range of researchers to apply them to their research questions, inferring biological relatedness using aDNA presents several challenges. Here we discuss common technical and conceptual challenges that researchers must contend with.
DNA degradation
After an individual’s death, DNA molecules degrade due to environmental factors such as temperature, moisture and oxygen exposure. As a result, aDNA recovered from human remains is often fragmented into short stretches and exists in very low concentrations. Additionally, these samples typically contain large amounts of exogenous DNA (i.e. DNA from microbes, soil, or other sources unrelated to the organism being studied). Due to these factors, aDNA data is usually sequenced at a low depth of coverage, meaning each position in the genome is sequenced only a small number of times, often averaging 1× or less. This low coverage makes it challenging, if not impossible, to determine accurately whether two individuals share zero, one, or two alleles in IBD at a specific position, a critical component of inferring genetic relatedness. To address this limitation, some of the methods mentioned above were designed to account for the unique characteristics of aDNA, in particular its low coverage, enabling researchers to infer relatedness despite existing challenges (Alaçamlı et al. Reference Alaçamlı, Naidoo and Güler2024; Popli et al. Reference Popli, Peyrégne and Peter2023; Ringbauer et al. Reference Ringbauer, Huang and Akbari2024).
Post-mortem damage
In addition to fragmenting, DNA also accumulates spontaneous chemical damage after an individual’s death (Briggs et al. Reference Briggs, Stenzel and Johnson2007; Dabney et al. Reference Dabney, Meyer and Pääbo2013). This damage often results in specific types of mutations (known as ‘transitions’), particularly near the ends of the short DNA fragments that are sequenced. Proper filtering of aDNA data, with the removal of transitions from the dataset using bioinformatics tools, is therefore crucial for accurate kinship inference. Without appropriate filtering, these mutations can bias results by artificially increasing the number of mismatches in pairwise comparisons.
Background relatedness
Another significant challenge is background relatedness, which refers to the genetic similarity that exists among individuals in a finite population due to shared ancestry at some point in the past (Weir et al. Reference Weir, Anderson and Hepler2006). It is worth noting that ‘genetic ancestry’ refers to the genetic links between individuals and populations, reflecting shared evolutionary and historical relationships. This concept is not synonymous with biological kinship, which implies the sharing of genetic material inherited from a recent common ancestor. A key challenge with genetic ancestry lies in its interpretation—ancestry proportions are statistical estimates that depend on the reference populations used, which may oversimplify human genetic diversity and evolution. For a more detailed discussion on this topic, see Coop (Reference Coop2022).
While human populations generally have low levels of background relatedness, this phenomenon can still confound biological kinship inference, inflating IBD statistics and resulting in false positives—in other words, wrongly categorizing two individuals as related. For instance, close kinship inferred between two individuals living thousands of years apart would be better explained by background relatedness rather than by a direct biological relationship. Addressing challenges such as this demands careful data curation, appropriate methodological choices, cross-validation of the observations with mtDNA analysis and disease markers in bones when available, and a nuanced interpretation of the results, incorporating insights from the observed funerary practices and radiocarbon dates.
Biological relatedness, kinship and ancestry
Population history research in palaeogenomics—and discussions in the press related to research—have been critiqued for over-interpretation of results (such as the conflation of biological ancestry and identity), lack of hypothesis testing and the imprecise use of terminology that has precise meanings in related disciplines (see, for example, Frieman & Hofmann Reference Frieman and Hofmann2019). All of these are also potential hazards in studying kinship systems of past societies as well.
As geneticists, we need to be mindful that social identity is far more complex than a list of biological attributes. But simultaneously we must avoid the pitfall of a false dichotomy between the biological and the social (Crellin & Harris Reference Crellin and Harris2020). We are creatures of both biology and culture, as were our ancestors, and if we are attempting to understand humans in the past, we must bring this complexity and nuance to the ways in which we conduct our research and the ways in which we interpret the results of our research. One approach which may help, suggested by Oras et al. (Reference Oras, de Groot and Björkstén2025), is the development of the ‘biomolecular humanities’—the integration of methodologies and theoretical frameworks from biomolecular sciences (such as molecular biology and genetics) and the humanities/social sciences. This integration calls for steps beyond collaboration or discussion between specialists, such as the co-training of students in multiple paradigms (akin to the four-field approach in many American anthropology departments) and the co-production of knowledge by scholars working together on research. We believe this can help researchers overcome some of the challenges we discuss here. In the next section, we discuss examples of recent studies that have done this effectively.
Insights into kinship from palaeogenomics
The relationship between mortuary practices and kinship
In a 2025 study, Quilter et al. (Reference Quilter, Harkins and Fanco Jordan2025) demonstrated how palaeogenomics can be integrated with archaeology to better understand mortuary practices at one site (perhaps with implications for others). Archaeologists have hypothesized that biological relatedness formed the basis for the mortuary treatment of an elite Moche burial group within the Huaca Cao Viejo temple at the El Brujo site in the Chicama Valley on the north coast of Peru. The research team characterized biological relationships between seven individuals buried within tombs constructed in the floor of an enclosure in the temple, finding that four generations of biological kin were represented in the burials. In combination with radiocarbon dating, stable isotopes, morphological analyses of the skeletons and archaeological analyses of the mortuary context, they were able to generate an extremely detailed model for the burial sequence. The site contained the simultaneous burial of, according to the original study, a very high-ranking female (Señora de Cao), her biological brother, another likely brother (whose DNA preservation made it difficult to distinguish between brother or uncle relationship, but whose age at death made it more likely he was a brother to the other two siblings), their grandfather, and two juveniles (the son of one sibling and the niece of the others). Archaeological analyses revealed further details. Both juveniles were seemingly sacrificed and included in the tombs of the high-ranking female and her brother (the biological father of the male child). The grandfather was most likely a secondary burial, transported from his original burial place to be included with the other elites. The grave goods of the high-ranking female were more numerous and richer than any other individual; the other adult males were buried with far fewer, and their proximity to her burial seemed to be an important factor. Isotopic profiles suggested that all individuals were local to the region except for the female child, who most likely grew up in the Andean highlands.
This study gives an example of how multidisciplinary teams can design a robust study to address specific questions about kinship without being biologically deterministic. The value of such integrative approaches has also been demonstrated in other studies—such as Amorim et al. (Reference Amorim, Vai and Posth2018) and Yaka et al. (Reference Yaka, Mapelli and Kaptan2021)—which successfully combined genome-wide data, uniparental markers, anthropological information and archaeological context to reconstruct kinship structures and illuminate broader patterns of social organization. Another strength of the Quilter et al. (Reference Quilter, Harkins and Fanco Jordan2025) study is its incorporation of community perspectives; the authors conducted outreach with descendant communities and other stakeholders, sharing and receiving feedback on aspects of the project before, during and after the research was conducted.
Social organization and residence patterns in the European Neolithic
The Neolithic site Gurgy ‘les Noisats’ in the Auxerrois within the Paris Basin is a large cemetery containing 128 individuals dating to between approximately 6970–5970 years before present (∼5000–4000 bce). Mortuary practices at Gurgy ‘les Noisats’ are different from those seen at contemporaneous sites throughout the Paris Basin. Individuals were buried in a variety of ways (including pit burials and burials in niches), with the majority placed in a flexed position on their left side within a container (or coffin). The scarcity, nature and random distribution of grave goods, along with the large size of the cemetery, suggests to archaeologists that Gurgy ‘les Noisats’ was a location for the burials of non-elite individuals and that mortuary practices may reflect influences from multiple Neolithic cultures (Rottier et al. Reference Rottier, Mordant, Chambon and Thevenet2005). These factors also made it difficult for archaeologists to link Gurgy ‘les Noisats’ clearly to the broader Neolithic Cerny culture, or address other questions regarding the basis for site organization and use.
Rivollat et al. (Reference Rivollat, Rohrlach and Ringbauer2023) conducted a multidisciplinary analysis of the burials at Gurgy in order to better understand the history of the site and the social organization of the people who used it. From biological relatedness estimates, as well as previously collected mitochondrial and Y-chromosome data and age-at-death estimates, they were able to construct pedigrees for over 100 of the Gurgy individuals. The reconstructed pedigrees showed a linkage of individuals through the male lineage; the absence of half-siblings suggests that reproductive partnering was monogamous. The spatial organization of burials within the cemetery was closely linked to paternal relatedness, although grave goods, body positioning and grave types were not. Adult females were under-represented at the site, and the majority of those who were present did not have parents or ancestors buried in the cemetery (although many had children). Strontium isotopic analyses showed that these female individuals were likely of non-local origin. Mitochondrial lineages carried by these women did not get transmitted past one generation. Together, these data strongly support a scenario of female exogamy and patrilocal residence in the genetic sense; that is, female marriage outside of the biological kin groupings identified in these pedigrees. However, as Cveček and Gingrich note (this issue), exogamy has a different meaning in cultural anthropology: it is marriage out of a socially defined group, which must be specified when using the term as it may or may not use biological relatedness as part of this definition. It is extremely important for researchers in both disciplines to be specific as to their terminology or we run the risk of talking past each other.
Bayesian modelling of radiocarbon dates, strontium isotope profiles, and the reconstructed pedigrees lends support to a scenario in which a multigenerational group moved into the local area from elsewhere, used the site to bury their dead for three to four generations, and then moved on to a different place. Whether the practices and social organization inferred from Gurgy ‘les Noisats’ is true more generally for Middle Neolithic societies is not yet known. Rivollat et al.’s multidisciplinary study provides an excellent example of an approach that can be applied to other sites in order to test this hypothesis. For palaeogenomics researchers, it illustrates the importance of wide sampling and the use of multiple lines of evidence: genome-wide markers as well as uniparental markers (mtDNA and Y-chromosome), age-at-death estimates and archaeological data. More limited sampling—and reliance on only a single data type—would have missed subtle details that were important in discerning residence and marriage patterns.
Social stratification, kinship, and marriage patterns
Cemeteries can give insights into the biological origins and social organization of the people buried within them and test long-standing archaeological and historical hypotheses regarding the significance of material culture as a marker for social status. This was one of the aims of Amorim et al. (Reference Amorim, Vai and Posth2018), who conducted genomic and isotopic analyses of Longobard individuals buried within two cemeteries dating to the Migration Period. They were able to construct highly detailed pedigrees using genomic data from 39 individuals from Szólád (∼604 ce) in present-day Hungary and 24 individuals from Collegno (∼580–630 ce) in present-day Italy, finding genetic similarity between populations at both cemeteries despite their geographic distance from each other.
Amorim et al. found evidence for four biological kin groups (referred to as ‘kindreds’ in the original study) at Szólád, the largest of which contained 10 individuals spanning three generations. Strikingly, mortuary treatment of the individuals within Szólád reflected membership in these kindreds, including the quantity and quality of grave goods, the spatial orientation of the graves, and inferred diet. Members of the largest kindred had the most elaborate burials, as well as diets enriched for animal protein relative to other groups within the cemetery. The kindreds maintained genetic boundaries with each other for several generations.
The study identified three kindreds at Collegno; like Szólád, these biological groups were associated with distinctive mortuary treatments. The majority of individuals belonging to the inferred biological kindreds in both cemeteries had ancestry associated with present-day populations inhabiting central and northern Europe, although southern European genetic ancestry was significantly present in both communities. Along with isotopic evidence, this finding is consistent with historical accounts of migration of the Longobards from the Roman province of Pannonia (present-day Hungary) towards Rome and its vicinities. However, the authors also found extensive genetic heterogeneity at the cemeteries, a pattern somewhat contrary to the highly structured genetic variation seen across populations in Europe today (Novembre et al. Reference Novembre, Johnson and Bryc2008).
These findings highlight the power of palaeogenomics to illuminate the relationship between kinship, biological relatedness, social practices and mortuary treatment. Amorim et al. (Reference Amorim, Vai and Posth2018) was one of the earliest—possibly the earliest—palaeogenomics studies to sample so densely (n=39) at a single archaeological site in an effort to test questions regarding mortuary practices and biological kinship. This large sample size was required in order to identify the extended pedigrees and the genetic heterogeneity observed in both populations. Moreover, the close collaboration between scholars of different disciplines provides a model for how to test archaeological and historical hypotheses robustly.
Challenging prevailing interpretations of family relationships: the Pompeii casts
In 79 ce, the eruption of the Somma-Vesuvius completely destroyed Pompeii, a Roman city in Campania (present-day Italy). The city was covered by pyroclastic deposits, preserving its buildings and streets. The shapes of the bodies of some of the victims, as well as some of their remains, were preserved into casts, made by pouring liquid plaster into the voids left by the decayed bodies. Its unique preservation and importance within the Roman Empire have resulted in Pompeii being designated as a UNESCO World Heritage Site and one of the most famous archaeological sites in the world. Interpretations of the identities of the victims and their family relationships were crafted by the archaeologists who first studied Pompeii and the restorers who built the famous Pompeii casts and were widely spread through museum exhibitions and educational materials. Subsequent palaeogenomics research demonstrated how these interpretations unconsciously reflected the prevailing cultural and social values of the scholars rather than the inhabitants of Pompeii themselves.
From the skeletons embedded in these casts, Pilli et al. (Reference Pilli, Vai and Moses2024) generated genome-wide aDNA data for six individuals to characterize genetic relatedness, sex and ancestry among the victims of the Somma-Vesuvius eruption. Palaeogenomics data showed that the inferred sex and biological kinship relationships do not match the prevailing interpretations about these individuals. One significant case is the one from the ‘House of the Golden Bracelet’. In 1974, four victims were discovered in close vicinity to one another and were thought to be a genetically related family (Guzzo Reference Guzzo2003). Among these, the two adults were interpreted as the parents of the two infants (approximately 4–6 years old). One of the adults was wearing an intricate golden bracelet and apparently holding one of the infants and was thus interpreted as the mother of the family. Genetic analyses indicated that all individuals were genetic males (sex chromosome complement XY). Furthermore, no evidence of biological relatedness of at least third-degree (e.g., first cousins or grandchild and grandparent) was found based on either the mitochondrial DNA or whole-genome data.
The extent to which biological relatedness factors into kinship, and how the two are intertwined with the behaviour of a person in life and the treatment of an individual in death in different societies is, to put it mildly, a complex subject. By challenging the prevailing narrative that these four individuals are members of a genetically related family, these findings by Pilli et al. (Reference Pilli, Vai and Moses2024) show us more broadly the influence of positionality on the interpretation of data. This glimpse into the non-biological relationships between people present at the site invites us to reflect on more complex scenarios, and reminds us of the possibility that kinship, gender and households might have been constructed differently than expected by the original researchers (and by the public interested in the site). Discussions of the complexity of kinship across different societies more broadly can be enhanced by the employment of palaeogenomics as a tool alongside archaeological, osteological, cultural and historical evidence.
Towards a more nuanced study of kinship in the past
An interdisciplinary, ethical and collaborative approach to research is essential for advancing kinship studies in palaeogenomics. Integrating theoretical frameworks from sociocultural anthropology, archaeology, Indigenous studies and other fields can help scholars avoid pitfalls when designing, conducting and interpreting palaeogenomics research in kinship studies. Such an approach promotes a more comprehensive understanding of kinship in the past, ensuring that the insights gained from palaeogenomics are placed within cultural and historical contexts.
The studies presented as examples above highlight the effectiveness of such nuanced approaches. For example, Rivollat et al. (Reference Rivollat, Rohrlach and Ringbauer2023) explicitly acknowledge that biological relatedness is only one form of kinship, stating that ‘Throughout this text we use the terms mother/father, son/daughter and siblings, as well as the binary sex terms male and female, in the genetic sense. We acknowledge that these are western kinship terms, but they are not meant to imply kinship terminologies or identities here. We cannot know if they were understood in this way by the Gurgy community’. We encourage our geneticist colleagues to follow this example of careful consideration of the limits of our data, as well as its potential. As the Quilter et al. study demonstrates, human identity is multifaceted and aspects of our identity change over time; thus, great care must be taken when describing individuals (e.g. ‘farmers’, ‘elites’) so as to avoid oversimplifying or essentializing people. As Amorim et al. and Pilli et al. show, kinship systems, like other forms of social connections, are deeply embedded in cultural contexts (see Sahlins Reference Sahlins2013), and we must acknowledge this complexity in the design, analysis and interpretation of genetic data, as well as in our publications.
We encourage our colleagues of all disciplines to be mindful of the terms and assumptions inherent in our respective fields, and strive for as much clarity and simplicity of language as possible to facilitate understanding. As a way of fostering essential conversations between scholars of different fields, as well as between scholars and the general public, we echo recommendations to publish palaeogenomics research in more diverse journals and to conduct discussions of research results in the press with great care (Booth Reference Booth2019; Sykes et al. Reference Sykes, Spriggs and Evin2019).
We close with our optimistic perspective: Our fields are moving in the right direction. Through dialogue and collaboration, our disparate fields are increasingly intertwining their methods and perspectives. May we grow together towards a more nuanced and accurate understanding of the many ways in which people lived kinship in the past.
Acknowledgements
We thank Dr Lauren Norman for her input on an early version of this manuscript, two anonymous reviewers for their insightful suggestions and Drs Sabina Cveček, Maanasa Raghavan and Penny Bickle for organizing this issue.
Open access
