Big Biology, Infrastructures, Algorithms, and Race: How Genomics Became Imbricated in Representations of Race

In contrast to 19^th century European anthropologists and biologists who classified humans into what they assumed to be biologically different racial categories, historians, sociologists, anthropologists, and some biologists writing during the last half of the 20^th century, have argued that U.S. designated racial categories do not constitute biologically or genetically distinct populations. Especially since World War II, scholars have emphasized that race categories are social, political, and historical constructions that differ over time and by locale. Yet, in the 21^st century, biomedical research on genomic risks for disease appeared to reinvigorate late 19^th and early 20^th century notions of races as genetic groups. To examine this claim, we conducted an archaeology of the data infrastructures used in genomic research to unearth the assumptions built into its practices and tools. We examine several layers of infrastructures, including algorithms. In other words, we show that infrastructures and algorithms are social organizations. These tools of science incorporate social assumptions and institutional arrangements. In order to understand which understandings and institutional arrangements become part of our scientific knowledge, for example, we deconstructed algorithmic technologies and software to examine how they were built and with which assumptions and definitions; we do not take them for granted. We show how choices made about algorithmic technologies and the data to which they are applied are value choices that have effects on population genetic data and their outcomes. That is, we show how socio-political and institutional concepts of race in the U.S. have been woven into the fabric of contemporary practices of knowledge production in new biomedical genomic and population genetic research.