Research in human genetics is increasingly enacted with advanced digital technologies, with genomics methods now being introduced into international research on the biological underpinnings of educational outcomes. Molecular genomics methods using bioinformatics hardware and software are highly data-intensive and depend on complex sociotechnical infrastructures of digital technologies, scientific epistemologies and methodologies, social relations and practices, which play an active role in shaping genomic data into knowledge (Stevens 2016). In data-intensive genetic sciences, knowledge about human lives, bodies, and behaviours is produced by large-scale initiatives using biosensor equipment, laboratory computer networks, databases, automation and analytics algorithms, generating novel renderings and understandings of internally-embodied states and processes (Vermeulen 2016). In the genetic sciences, human bodies, lives, and actions are understood in ‘bioinformational’ formats, as codes, networks and programming, which are made legible with data scientific methods of searching and pattern detection (Koopman 2020).

In the last 15 years, genomic methods have been repurposed for analysing educationally-relevant biological processes and structures, particularly the ‘genetic associations’ and ‘genetic architectures’ claimed to underpin learning and achievement, and the environmental factors that mediate them (Malanchini et al 2020), by international networks of scientists in two distinctive fields. Behaviour genetics has a long history in education, including controversial involvement in intelligence testing (Panofsky 2015), and has begun using molecular genomics methods to 'discover' the biological substrates of learning, cognition and school achievement (Kovas et al 2016). Since around 2010, the new interdisciplinary synthesis known as social science genomics, or sociogenomics, has combined expertise in social statistics and bioinformatics to study the genetic bases of social and economic outcomes (Bliss 2018). Utilizing vast ‘biobanks’ of genomic bioinformation and data scientific methodologies, sociogenomics aims to ‘finally open the black box of the genome’ in order to ‘delve into the biological mechanisms and come up with a better understanding of the pathways from cells to society’ (Conley and Fletcher, 2017, 35).

A key research target of sociogenomics to date is the biological associations, architectures and mechanisms underpinning educational outcomes and the environmental factors that mediate them (Martschenko et al 2019). Scientists have generated dozens of studies and scientific articles by utilizing molecular genomics methods to analyse bioinformational samples in the millions and producing new knowledge claims about the genetic bases of educationally-relevant traits, behaviours and outcomes (Plomin 2018; Harden 2021). As such, recent research in behaviour genetics and sociogenomics signifies the emergence of ‘educational genomics’ as a domain of international research, knowledge production, and potential policy influence (Visscher 2022).

As the materialization of a ‘new biological rationality in education’ with its own distinctive methodologies and truth claims (Gulson and Baker 2018), educational genomics is an emerging  science with potentially profound consequences for educational research, policy and practice internationally. This paper presents an analysis of the formation of the ‘knowledge infrastructure’ of educational genomics, drawing on an ‘infrastructure studies’ approach to datafied knowledge production (Bonde Thylstrup et al 2019). A knowledge infrastructure is a relational and sociotechnical system consisting of people and organizations, epistemologies and practices, and technologies and methods that underpin knowledge production (Edwards et al 2013). Theoretically, the paper is informed by science and technology studies conceptualizations of ‘data-centric biology’ (Leonelli, 2016) and a ‘postgenomic condition’ characterized by datafied, molecular explanations of human life (Reardon 2017). Such studies illuminate how the specific software, hardware, algorithms and data structures of bioinformatics analysis, in association with the conceptual schema of scientific communities, are reshaping how biological science is enacted and the kinds of knowledge it produces, with significant consequences in terms of biomedical explanation, public understanding, and political intervention (Chow-White and García-Sancho, 2012; Stevens 2016; Rajagopalan and Fujimura 2018).

Substantively, the paper provides an examination of the social, epistemic and technical constitution of the emerging knowledge infrastructure of educational genomics. Conceptualized as an infrastructure of data-intensive biological knowledge production, educational genomics is constituted by social associations, conceptual architectures, and technical algorithms: (1) educational genomics is performed by large-scale associations or consortia, representing a ‘big biology’ mode of highly-funded, cross-sector and interdisciplinary networked knowledge production in education; (2) educational genomics enacts a specific conceptual schema, or an epistemic architecture for understanding the biological determinants of educationally-relevant outcomes and behaviours; and (3) educational genomics mobilizes methodological apparatuses powered by algorithms for data-intensive analysis of digital bioinformation and the production of new biological knowledge claims related to education. The analysis draws on three main methods. First social network graphing software was used to identify the social relations and organizational associations that constitute educational genomics as a domain of scientific inquiry. Second, documentary analysis of a large corpus of scientific publications on educational genomics was undertaken to trace its main discourses, conceptual schema and knowledge claims. Finally, technographic descriptions of key technologies used in educational genomics were developed, focusing on key algorithmic methods including bioinformational data mining and the production of predictive ‘polygenic scores’. In sum, the methods enabled an analysis of the complex ways that an emerging infrastructure consisting of algorithmic apparatuses that are used to retrieve, process and order bioinformation, in concert with the epistemic architecture of scientists working in interdisciplinary and cross-sector associations, co-produce particular ways of understanding the biological substrates of educationally-relevant behaviours and outcomes. The results of the infrastructural formation of educational genomics are already materializing in proposals for educational policy and practice, including the use of polygenic scores to sort and categorize students by their predicted outcomes, and even to utilize genetic data in personalized forms of ‘precision education’ that would be modelled on ‘precision medicine’ in the biomedical domain.

Findings of the study indicate that educational genomics generates significant implications for educational research, practice and policy. First, educational genomics signifies the emergence of ‘big biology’ as a mode of educational knowledge production, and the formation of powerful new scientific associations in education-relevant research whose large-scale quantitative research may potentially displace other infrastructures of knowledge production. It represents the entry of data-centric, international, consortium-based modes of biomedical investigation into education, directly challenging the authority of 'non-genetic' social sciences to contribute to education policy and practice. Second, new bioinformatized conceptions of educational outcomes are produced by educational genomics. It advances the molecularization of socially-structured phenomena, assuming educational achievements are structured to a statistically significant degree by biological mechanisms that link genetic differences to brain development. Critically, and contrary to its claims of unbiased, data-scientific objectivity, educational genomics may operate politically to superimpose molecular explanations on social problems in education, privileging biological understandings while obscuring the social forces that shape educational outcomes. Finally, by producing new bioinformatized explanations, educational genomics constructs the subjectivity of a searchable student who is quantified and known at the molecular scale through bioinformatics analysis. The searchable student is anatomized using bioinformational data mining methods as statistically significant associations between thousands of interacting genetic differences, and rendered as a single bioinformational number – a polygenic score – that predicts their educational achievement. Moreover, educational genomics significantly reconceives educational outcomes in computational terms, as the result of genetic ‘codes’, ‘programming’ and ‘information’ that are only legible through the deployment of data mining tools capable of searching for patterns in vast masses of bioinformation. As such, educational genomics produces a searchable student whose educational trajectory is said to be explainable and predictable through data mining bioinformation, and who may therefore become the subject of genetically-informed educational policy and practice interventions.

Bliss. C. (2018). Social by Nature: The promise and peril of sociogenomics. Stanford: Stanford University Press. Bonde Thylstrup, N., Flyverbom, M. and Helles, R. 2019 Datafied knowledge production: Introduction to the special theme. Big Data and Society, July–December, 1–5. https://doi.org/10.1177/2053951719875985. Chow-White, P.A. and García-Sancho, M. 2012. Bidirectional Shaping and Spaces of Convergence: Interactions between Biology and Computing from the First DNA Sequencers to Global Genome Databases. Science, Technology, & Human Values, 37(1), 124–164. Conley, D. and Fletcher, J. (2017). The Genome Factor: What the social genomics revolution reveals about ourselves, our history and the future. Oxford: Princeton University Press. Edwards, P. N. et al. 2013. Knowledge Infrastructures: Intellectual Frameworks and Research Challenges. Ann Arbor: Deep Blue. Gulson, K.N. and Baker, B. 2018. New biological rationalities in education. Discourse: Studies in the cultural politics of education, 39(2), 159-168. Harden, K.P. (2021). The Genetic Lottery: Why DNA matters for social equality. Oxford: Princeton University Press. Koopman, C. (2020). Coding the Self: The Infopolitics and Biopolitics of Genetic Sciences. Hastings Report, 50(3), 6-14. Kovas, Y. et al. (2016). How genetics can help education. In Y. Kovas, S. Malykh, and D. Gaysina (eds.) Behavioural genetics for education, 1–23. London: Palgrave Macmillan. Leonelli, S. (2016). Data-Centric Biology: A philosophical study. London: University of Chicago Press. Malanchini, M. et al. 2020. Cognitive ability and education: How behavioural genetic research has advanced our knowledge and understanding of their association. Neuroscience and Biobehavioral Reviews 111: 229–245. Martschenko, D., Trejo, S. and Domingue, B.W. (2019). Genetics and education: Recent developments in the context of an ugly history and an uncertain future. AERA Open, 5(1): 1-15. Panofsky, A. 2015. What does behavioral genetics offer for improving education? Hastings Center Report, 45(5), S43–S49. Plomin, R. 2018. Blueprint: How DNA makes us who we are. London: Allan Lane. Rajagopalan, R.M. and Fujimura, J.H. 2018. Variations on a Chip: Technologies of Difference in Human Genetics Research. Journal of the History of Biology, 51, 841–873. Reardon, J. (2017). The Postgenomic Condition: Ethics, justice, and knowledge after the genome. Chicago: University of Chicago Press. Stevens, H. (2016). Hadooping the genome: The impact of big data tools on biology. BioSocieties, 11, 352–371. Vermeulen, N. (2016) Big Biology. N.T.M. 24, 195–223. Visscher, P. (2022). Genetics of cognitive performance, education and learning: from research to policy? npj Science of Learning, 7, https://doi.org/10.1038/s41539-022-00124-z