Abstract PR12: Dynamic clustering of cancer genomics datasets beyond pre-defined human categories

Abstract

Abstract Background: The history of the race concept sheds the light on the lack of scientific basis of racial categories. Early attempts of using race in scientific practice (ca. 17th century) took place during times of colonial expansions and the transatlantic slave trade. The lenses through which scientists approached race categories changed over the next centuries, but legacies of using race categories in genetic studies are still embedded in some of today’s practices. Consequently, this engenders the need for novel approaches that transcend categories of or related to race in genomics to avoid unintended consequences and improve the study of disease. Methods: We introduce an approach that dynamically generates trait-specific cohorts based on genomic clustering at a subset of loci known to contribute to the disease under study. Reflecting the broad diversity of traits, the number of clusters and cluster memberships vary depending on the trait and its implicated genomic regions. Our method hence directs the lens of studying human biodiversity with respect to disease to the trait level rather than “population level.” Results: We test our approach on cancer and control whole genome sequencing datasets to partition each based on predisposition to 8 cancer types. Focusing on COSMIC genes known to have a germline contribution to cancer, we demonstrate that resulting clusterings vary across cancer types and transcend race, ethnicity—and ancestry categories, supporting our call to generate dynamic categorization sets at the trait level. Conclusion: Our suggested method highlights the importance of historical perspectives to co-shape today’s scientific practice. Experimental results resonate with previous joint calls from scholars in genetics and the humanities and social sciences to focus on the study racial and ethnic disparities in health as a result of rampant societal and environmental disparities—rather than genomic differences partly based on historical attempts to impose fixity on fluid social constructs. Citation Format: Hussein Mohsen, Kim Blenman, Lajos Pusztai. Dynamic clustering of cancer genomics datasets beyond pre-defined human categories [abstract]. In: Proceedings of the AACR Virtual Conference: Thirteenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2020 Oct 2-4. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(12 Suppl):Abstract nr PR12.