Because of their central population registries, Scandinavian countries have often been referred to as 'laboratories' for epidemiological research. Denmark, for instance, introduced a central identification number for each resident in the context of a tax reform in 1968 and has maintained a large number of disease registries. The possibility of record linkage, i.e. the traceability of individuals across different registries, has co-shaped the research endeavors and routines of social scientists, epidemiologists, and biomedical researchers alike. By following practices of record linkage, I will show how the feature that system developers imagined as 'catalyst of mechanization' has become an infrastructure for biomedical research. The ubiquitous usage of the ID-number in Denmark has generated a continually growing data archive of the interactions between citizens and public services, including the healthcare system. In this talk, I will focus in particular on the temporalities that emerge when this archive is mined through the calculative techniques of epidemiology: Biomedical researchers rediscover older collections of samples and data to speak to 21st century research questions, from environmental health to epigenomics, nutritional epidemiology and health services research. Public health science and evidence-based medicine mobilize aggregate versions of the population's past to predict individual and collective futures that inform decision-making in the present. Calculative devices such as risk scores - anticipatory infrastructures - enroll citizens into preemptive adjustments and optimization. By exploring the data assemblages of epidemiological research, this talk will reflect on the relations of collective and individual as well as of science and the state in public health infrastructures.