Abstract C093: An interactive resource to probe ancestry in cancer cell lines

Abstract

Abstract Recent work has pointed to a lack of diversity in genomics studies from genome-wide association studies to somatic (tumor) genome analyses. Yet, population-specific variation has been shown to contribute to health disparities in cancer risk and prognosis. Immortalized cancer cell lines have been widely used for various aspects of cancer research from mechanistic studies to drug screening. More recently, larger collections of cancer cell lines have been developed with the objective of better representing the genomic heterogeneity found in primary tumors. The genetic ancestral origin of cancer cell line models is rarely acknowledged and often unknown. Genome-wide genotyping data (Affymetrix SNP6.0 array) was obtained for 1,393 cancer cell lines from the Catalogue of Somatic Mutations in Cancer (COSMIC) and Cancer Cell Line Encyclopedia (CCLE) collections. Clustering of the cell lines in relation to reference populations of the 1000 Genomes project was visualized using t-Distributed Stochastic Neighbor Embedding (t-SNE) plots. Ancestry proportions constituting each cell line were estimated using the program Admixture. For over 46% of the cell lines analyzed, ethnicity has not been reported in public databases. The ancestral origins of the cell lines, as measured by genetic markers, were distributed as follow: European 58.1%, East Asian 28.8%, African 1.4%, South Asian <1%, mixed ancestry 11.4%. We also noticed several inaccuracies when comparing reported ethnicity to the ancestry observed from genomic data. Of the 44 cell lines reported as African or Black, 27.0% showed diverse degrees of admixture, with a European component reaching up to 37.5%. Inversely, a proportion of the cell lines reported as Caucasian or white did include admixed cell lines with strong African genetic contribution (up to 86.8%), while others showed patterns of admixture similar to Hispanic populations. Furthermore, we present an interactive online tool, Cell Line Estimated Ancestry Resource (CLEAR), where ancestry can be visualized in relation to reference populations of the 1000 Genomes project. Our data indicate that the collection of cell lines commonly used in cancer research is not representative of the diverse ancestry and admixture characterizing human populations. In addition, we demonstrate that the reported information on ethnicity contains inaccuracies and rarely acknowledges the mixed origins of their genetic background. These observations are likely to impact the reproducibility of studies relying on those models and the relevance of the findings to diverse populations. Citation Format: Zhihua Chen, Alvaro N. Monteiro, Jamie K. Teer, Steven Eschrich, Julie Dutil. An interactive resource to probe ancestry in cancer cell lines [abstract]. In: Proceedings of the Eleventh AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2018 Nov 2-5; New Orleans, LA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(6 Suppl):Abstract nr C093.