Abstract A029: Highly-resolved integrated single cell atlas of human breast cancer enables hypothesis-driven analysis of tumor heterogeneity

Abstract Single cell transcriptomics (scRNAseq) remains the method of choice to quantify and annotate cellular heterogeneity in the tumor and its microenvironment (TME), yet most studies are under-powered to associate changes in tumor heterogeneity with phenotypes such as tumor subtype, grade, and patient age, among others. In this study, we created an integrated atlas of human breast cancer (BC), the largest resource of its kind, totaling > 700,000 cells across 145 patients, and optimized computational methods to benchmark integration performance, and robustly perform hierarchical cell type annotation. By combining single profiles, a higher-resolution annotation of immune, stromal, and epithelial cell types was achieved. Further, using this integrated atlas, differential abundance analyses identified significant changes in immune and stromal cell type composition associated with tumor grade. These changes in TME heterogeneity were not discernible or had effects in the opposite direction when the analysis was limited to individual BC studies, highlighting the need for atlas-based mega-analysis approaches. This highly-resolved integrated scRNAseq BC atlas will be a valuable resource for hypothesis-driven analyses of tumor heterogeneity including our own ongoing analysis of comorbidities in BC. Citation Format: Andrew D Chen, Lina Kroehling, Christina Ennis, Gerald Denis, Stefano Monti. Highly-resolved integrated single cell atlas of human breast cancer enables hypothesis-driven analysis of tumor heterogeneity [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Functional and Genomic Precision Medicine in Cancer: Different Perspectives, Common Goals; 2025 Mar 11-13; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2025;85(5 Suppl):Abstract nr A029.